United States
            Environmental Protection
            Agency
             Office O1 Water
             (WH-585)
EPA822/R-92-001
February 1992
vvEPA
Consumption Surveys For
Fish And Shellfish

A Review And Analysis
Of Survey Methods
                                    X,.'.". Pnnted on Recycled Paper

-------
Consumption Surveys For
     Fish And Shellfish

  A Review And Analysis
     Of Survey Methods
            February 1992
             Prepared For:


     U.S. Environmental Protection Agency
       Office of Science and Technology
           401 M Street, SW
          Washington, DC 20460


        EPA Contract No. 68-C9-0013
  For additional copies of this
  publication write:
  Fish Contamination  Workgroup
  401 M Street, SW  (WH-585)
  Washington, DC  20460
  or phone; 202-260-70*'

-------
                                                               CONTENTS
                                                                            Page
Summary   	v

Acknowledgements	vii

I.   Introduction	1

II.  Survey Approaches	7

         Recall - Telephone Survey	8
         Recall - Mail Survey   	10
         Recall - Personal Interview	14
         Diary  	17
         Creel Census  	18

III.  Important Method Considerations	23

         Survey Design	25
         Selection of Respondents	27
         Information Sought	28
         Quality Assurance  	31
         Statistical Analyses	32

IV.  Discussion	35

V.  Bibliography	37

Appendix - Summary of Survey Methods Information   	49

-------
IV

-------
                                                                        SUMMARY
Although several studies have demonstrated that fish and shellfish consumption rates differ both
regionally and within specific subpopulations, most States do not have available sufficient data to
calculate local consumption rates or to identify special populations at risk. Examples of these special
populations are recreational and subsistence anglers and members of their households-in particular,
women of child-bearing age, children, and the elderly-who frequently consume fish obtained from
contaminated sites. This report was designed as a critical assessment of fish tissue consumption
rate survey approaches and methods and their applicability for estimating consumption rates in
recreational and subsistence fishing populations.  Additional information is provided to assist
Federal and State agencies in developing appropriate surveys to answer questions and resolve issues
related to the fish consumption rates of special populations.

Five approaches to obtaining fish consumption data were reviewed: (1) recalled information collected by
telephone; (2) recalled information collected by in-person (face-to-face) interviews; (3) recalled
information requested on self-administered mailed questionnaires; (4) diaries maintained by anglers;
and (5) on-site creel censuses. The effectiveness of the approach used to obtain adequate information for
fish consumption rate calculations varied with the objective(s) of the survey.  For example, creel censuses
usually failed to collect data on consumption. Many surveys combined two or more approaches in order to
maximize the number of respondents or validate the information obtained. Several studies addressed actual
contaminant exposures through physical examinations and measurement of blood serum levels of contami-
nants, while others investigated risk perception and compliance with fish consumption advisories and bans by
the targeted anglers.

Five elements common to all surveys have been identified, and specific methodological details are
provided to help solve problems that may be encountered when undertaking a fish consumption
survey. (1) Survey design must address the purpose for which the survey is to be conducted, the
resources available for  carrying it out, including time and funding available, and the approach to be
used. (2) Survey participants should be identified from a pool of subsistence or recreational anglers,
and the method by which the sample is selected may vary depending on the approach that will be
used to collect the data and how the data  will be analyzed.  (3) The information to be collected
should examine  sociodemographic factors that may influence fish consumption rates, as well as
those factors that are  needed to calculate fish  consumption rates, minimizing the number of
assumptions that could compromise results. The survey length and complexity should be carefully
considered in order  to elicit maximum cooperation from respondents.  (4) Appropriate quality
assurance procedures need to be developed before beginning the survey, and quality control must
be carefully monitored  during the survey to ensure the validity of the data before statistical analyses
are conducted.  (5) Data processing procedures and statistical  analyses should be performed to
provide the desired information and correlations.

-------
VI

-------
                                               ACKNOWLEDGMENTS
This document was compiled by Esther Peters, under the direction of Clyde Houseknecht, U.S.
Environmental Protection Agency (USEPA), with  assistance from Andrew Zacherle, Paula
Stouder, and Mary Beth Corrigan of Tetra Tech, Inc.  Alison Greene, Richard Hoffmann,
Elizabeth Southerland, and Quentin Stober of the USEPA reviewed drafts of the manuscript.
Additional comments were received from Chuck Cox, Ministry of the Environment, Toronto,
Ontario, Canada; Ellen Ebert, ChemRisk, Portland, ME; Beth Jones Fiore, Wisconsin Depart-
ment of Health  and Social Services, Madison, WI; Patricia Guenther, U. S. Department of
Agriculture, Hyattsville, MD; Barbara Knuth, Cornell University, Ithaca, NY; Pamela Shubat,
Minnesota Department of Health, Minneapolis, MN; and Patrick West, University of Michigan,
Ann Arbor, MI.  Representatives of Federal and State agencies contributed to the development
of the issues and information requirements presented in Table 1.
                                                                                VII

-------
Vlll

-------
                                                        I.  INTRODUCTION
Concern over potential human health risks associated with chemically-
contaminated fish and shellfish* has led many States to issue consumption
advisories and bans in an effort to limit exposures to certain organic com-
pounds and metals that may become concentrated in the tissues of these
organisms. However, the processes and procedures by which States issue
fish consumption advisories and bans have varied. In a recent effort to
evaluate the fish advisory  process in the States, the U.S. Environmental
Protection Agency (EPA) provided a grant for the American Fisheries
Society (AFS) to conduct a survey of State fish advisory practices (Cunning-
ham et al., 1990). In the survey, State representatives were asked to describe
their fish advisory process and procedures, to identify State concerns related
to the advisory process, and to recommend actions that could be undertaken
by the Federal government to improve the effectiveness of the advisories.

To follow up on the State recommendations for Federal action, EPA invited
officials from State agencies to attend a Federal-State Forum on August 30,
1990, in Pittsburgh.  Representatives of agencies from 27 States and the
District of Columbia, as well as several Federal agencies, including EPA, the
Food  and Drug Administration  (FDA), the National Oceanic and Atmo-
spheric Administration (NOAA), the U.S. Fish and Wildlife Service (FWS),
the Tennessee Valley Authority (TVA), and  the Agency for Toxic  Sub-
stances and Disease Registry (ATSDR) were present. The agenda for the
Forum contained a list of  the Federal action  items identified in the AFS
survey.  Participants were  asked to  rank proposed Federal action items as
short- or long-term priorities and to recommend  other action items not
previously identified in the survey.  Each participant was also requested to
submit the three action items that were most important to his/her program.
The second most frequently requested short-term action item contributed by
Forum participants was to conduct  surveys/studies to assess the fish con-
sumption rates of various subpopulations in different regions of the country
(Southerland, 1991).

Fish consumption rates differ throughout the country and for specific sub-
populations (e.g., Hu, 1985). The use of an "average" consumption rate for
typical households, recreational anglers, and subsistence anglers may not
""Hereafter, in this document, "fish" and "fishing" will include shellfish and
shellfishing, except where specific surveys are discussed.

-------
and subsistence anglers may not accurately reflect the local consumption rate
in a particular subpopulation and may overestimate or underestimate the risk
associated with the consumption of contaminated fish tissue by different
members of households. Presently, most States do not have available suffi-
cient data to calculate local consumption rates or identify special populations
at risk. As a result, a variety of methods are used for estimating consumption
rates when calculating risk associated with the consumption of chemically-
contaminated fish tissue (USEPA, 1989). To further complicate the issue,
recreational anglers may catch fish from contaminated sites for sport, but not
consume them, while subsistence anglers may be obtaining a large proportion
of their diet from contaminated  resources because they cannot  afford to
purchase other foods.  There are  also commercial-type subsistence fishing
operations, which obtain fish on a larger scale to provide these items to
communities. The amount of time spent in these activities may vary depend-
ing on the weather and the state of the fishery (seasonal  restrictions, for
example).

Human exposures to chemical contaminants (e.g., dioxins or poly chlorinated
biphenyls from  industrial sources, pesticides from  nonpoint sources, or
mercury from natural  sources) through fish consumption are a function of
the quantities of these foods consumed by humans as well as the ability of
different species of fish to bioconcentrate the chemicals of concern. The
EPA, FDA, National Marine Fisheries Service (NMFS/NOAA), and other
organizations are responsible for monitoring the chemicals found in these
organisms. A number of recent studies, however, have pointed out that the
national surveys that are the basis for many human health risk assessments
fail to target some of the potentially most high-risk populations, including
recreational and subsistence anglers and their families.

Early studies of fish consumption provided only limited data (e.g., Nash,
1971; Hu, 1985). Although the number offish meals was tallied, socioeco-
nomic or demographic questions were usually limited and no distinction was
made between recreationally (self-caught) and commercially harvested (in-
cluding processed/canned) fish consumed. Furthermore, in these early stud-
ies there was usually no characterization  of types of fish consumed (an
exception was  the National Marine Fisheries Service Survey 1973-74).
Surveys were either specific to particular regions of the country or national
in scope (nonspecific).  Thus, while these databases did provide important
information on consumers and frequency of consumption of various fish
products, there was no effort to identify subsistence and recreational anglers
obtaining their catches from polluted waters. Regionally specific data could
not be used in consumption rate calculations for other areas of the country,
(Rupp et al., 1980). Another review by SRI International (1980) found that
the most reliable source of data on human fish consumption was the National
Purchase Diary Fish Consumption Survey, a national questionnaire survey
conducted during 1973-74 by NPD Research Inc. SRI performed additional

-------
tabulations of the corrected data to obtain mean consumption rates for various
demographic categories and fish species.

Wagstaff et al. (1986) examined three types of national studies: (1) commer-
cial production data on landings, imports, and exports of food fish; (2) surveys
of food (including fish) intake; and (3) surveys specific for fish intake. Commercial
fish production data failed to include commercial freshwater fish, recreationally
caught fish, or marine fish sold at roadside stands (see  also report by SRI
International, 198O, Kleiman, 1985). General food intake surveys were limited in
scope, season, or demographic and socioeconomic data.  Specific fish intake
surveys, based on weekly diaries for periods up to a year, included all fish meals,
whether caught or bought, but recreational catch information  was sparse to
nonexistent Although estimates of per capita fish consumption based on these
surveys  were similar, Wagstaff et al. (1986) concluded that improved survey
design and implementation of quality control  in conducting, documenting, and
reporting the results of such a survey were needed.  Fisher (1988) reviewed nine
early surveys and found that none of them provided the data needed to estimate
usual or heavy fish intake or to examine recreationally-caught and consumed fish
species. Hence, these surveys had only limited value in determining diet/health
relationships or performing risk assessments associated with fish consumption.
He noted, however, that because studies suggested an upward trend in per capita
consumption, more recent and more detailed information was needed "either by
expansion of currently planned nutrition surveys or by focused efforts to obtain
such data from surveys on fish consumption."

Despite the numerous limitations of these early studies, calculations of fish
consumption rates suggested that certain subpopulations, based on race,
ethnic origin, age, sex, income, and residence, did consume more fish than
other groups.   More recent surveys of Michigan sport anglers and their
families by West et al. (1989a,e) revealed that minorities from cities, rural
Native Americans,  and the elderly also caught and consumed more fish.
Some recent surveys have attempted to link fish  consumption rates to
epidemiologic studies of health status (USEPA, 1984b) or body burden levels
of contaminants (Fiore et al., 1989).  These studies and observations of
fishing  activity at known chemically-contaminated sites (e.g., Puffer et al.,
1982a,b; Belton et al., 1985; Smith and Enger, 1988; Smith and Thompson,
1989) indicated that more detailed surveys targeting subsistence and recrea-
tional anglers were warranted to improve calculations of fish consumption
rates and risk assessments for specific subgroups (Table 1).  Such informa-
tion is important for determining the success of advisories and bans issued
to reduce health  risks from eating contaminated fish and/or of changing
waterbody management policies to reduce or el iminate toxic chemical inputs.

EPA recognizes that studies of fish consumption patterns need to be con-
ducted to update  current information and to focus on potentially high-risk
geographical or cultural populations.  To address this  need,  EPA has

-------
         TABLE 1. ISSUES AND INFORMATION REQUIREMENTS
                  FOR FISH CONSUMPTION SURVEYS

Sociodemographic Characteristics of Angler:
      •  Age
      •  Occupation/employment status
      •  Income level
      •  Education level attained
      •  Number of household members
      •  Race/ethnic group, sex, age, height, and weight of the fisherman and each
        household member
      •  Pregnancy/lactation status of women in the household
      •  Language spoken at home
      •  City of residence
Fishing Activities:
      •  Location(s) of fishing activities (specific sites, type of waterbody)
      •  Distance(s) of fishing activities from principal residence
      •  Seasonal and temporal distribution of fishing activities (total number of days
        per season, which months of the year,  for each location)
      •  Fishing effort (hours/outing, hours/day, outings/month, days/month)
      •  Purpose for fishing (consumption, sport only: catch and return, etc.)
      •  Mode of fishing (nets, traps, hook and line, etc.; pier, shore, private boat,
        charter boat,  SCUBA)
      •  Type of fish captured (general category such as bottomfish, flatfish; or iden-
        tified to species or group of species)
      •  Numbers of fish captured per outing by species
      •  Size ranges of fish captured (minimum and maximum weights and lengths
        by species)
      •  How the fish were disposed of (released, consumed  by household, sold ,
        given away)
      •  How long involved  in fishing activities and consuming self-caught fish (new
        to sport or years

-------
                          TABLE 1. Continued
Preparation and Consumption Patterns:
      •  Portions of fish consumed (may vary with the species)
      •  How the fish were prepared for eating (skinned, fillet, steak, shucked, etc.)
      •  How fish were cooked (baked, fried, steamed, etc.)
      •  Amounts (weight) of wild-caught fish eaten per meal/day/week/month for
        each person in household
      •  Special cultural/ethnic practices in fish consumption and preservation
      •  Consumption offish purchased in supermarkets, fish markets, or roadside
        stands; purchased at the dock; or obtained by bartering (amounts, fre-
        quency)
      •  Consumption of other aquatic organisms, waterfowl, or wildlife that may
        have consumed fish from same sites (amounts, frequency)
      •  Fish frozen or preserved and eaten throughout the year or eaten only when
        fresh
      •  Participation in food assistance program
      •  Source of home water supply
      •  Voluntary risk patterns (smoking, drinking)
Fish Consumption Advisory Awareness and Understanding:
      •  Has the angler heard, from announced fishing bans or posted notices, of
        the possible contamination offish by chemical or biologic agents in areas
        where presently fishing or where planning to fish?
      •  If the answer to the previous  question is yes, has it affected his/her fishing
        activities, fish preparation methods, or consumption patterns?
      •  What,  if anything, would stop the angler from eating the fish that he/she has
        caught?
      •  Did the angler ever get sick from eating self-caught/self-prepared fish or
        shellfish?
      -  Did the angler ever observe any abnormalities, internal or external, in cap-
        tured fish? If so, were the fish consumed, thrown out, or given away?

-------
implemented a three-phased approach for assisting the States in estimating
fish tissue consumption rates in potentially high-risk populations.  This
approach includes the following steps:

-   Review and critically  evaluate existing fish tissue  consumption  rate
   survey methods and determine their applicability for estimating consump-
   tion rates in recreational and subsistence fishing populations.
•   Conduct a workshop for the States presenting the results of the review and
   critical evaluation offish tissue consumption survey methods.
•   Provide direct support to the States in conducting fish tissue consumption
   surveys, targeting recreational and subsistence anglers.

This document was prepared to meet the first step in this process. Existing
literature concerning fish tissue consumption was reviewed, and selected
surveys were evaluated to identify approaches (recall vs. diary vs. creel
census) and methods for survey design and analysis.  The purpose of this
report is to assess the attributes and shortcomings of these approaches and to
explore the underlying methods involved in designing and conducting fish
consumption surveys. The report also discusses the types of questions that
need  to be answered if we are to understand fish consumption patterns in
high-risk populations (Table 1). It does not, however, recommend a specific
protocol for  use  by the States.  The methods, approaches,  and  questions
chosen will depend on the goals, objectives, and situations of the  particular
State and may also vary for the high-risk subpopulation  to be investigated.
Therefore, a  variety of options and guidelines for designing  and  executing
surveys is presented. This document is intended to assist Federal and State
agencies in revising surveys so that the types of information needed for valid
statistical analyses to adequately address human health risks in subsistence
and recreational anglers and their families can be collected efficiently and
cost-effectively.  Survey professionals from government, academia, and/or
private industry should also be consulted to ensure a successful survey.

-------
                                            II.  SURVEY APPROACHES
A variety of approaches has been used in attempts to develop appropriate
estimates of fish consumption rates. Fisher (1988) examined techniques for
obtaining data for a "national" sample of individuals and for samples of
subpopulations that might be more frequent fish consumers. He noted that
the complex problem of estimating fish consumption for possible assessment
of diet/health relationships and risks associated with the use of fish as food
required consideration of the following:

•   Sociodemographic characteristics of consumers;
•   Geographic and seasonal variations in consumption;
•   Species of fish and geographic origin of species consumed;
•   Parts of fish consumed; and
•   Quantities consumed.

The approaches to collecting data on fish consumption were categorized as
follows:

•   Indirect  - data collected on food disappearance into marketing channels
   or households (the unit of observation) and
•   Direct -  data collected on actual food use or food consumed by a variety
   of methods (i.e., the household or individual intake is the variable mea-
   sured).

Indirect techniques were usually deemed unsuitable for small-scale studies
and did not allow for correction of waste  or individual intakes.  Direct
techniques, such as food diaries or records, weighed intake, dietary recall,
food frequency, and duplicate portion studies, provided individual consump-
tion data but were more labor-intensive in both data gathering  and analysis.
More information on the attributes and limitations of direct approaches to
quantification of daily consumption by individuals was provided in Ander-
son (1986).

Five different approaches to conducting surveys of subsistence and recrea-
tional anglers were identified during this review of recent fish  consumption
surveys. In this section, each approach will be described and the advantages
and disadvantages presented. Four categories of information needs similar
to those listed above by Fisher (1988) have been compiled and are presented
in Table 1,   These categories include questions that need to be answered or
issues that need to be resolved in order to be able to calculate fish consump-
tion rates for special populations. The questions were derived from  recent

-------
                    fishing/shellfishing surveys and comments from representatives of Federal
                    and State agencies and other organizations. Although the types of data on
                    sociodemographic characteristics, fishing activities, preparation and con-
                    sumption patterns, and fish advisory awareness and understanding will be
                    discussed in more detail below, the ability of each approach to adequately
                    address these information needs will be examined in this section, in particu-
                    lar:

                    •  Can the approach assess region-specific (rather than national) consump-
                       tion rates?
                    •  Can the approach target and identify specific subpopulations of concern
                       (i.e., subsistence and recreational anglers)?

                    The use of any particular approach will depend on the specific objectives of
                    the study and the questions asked, as well as other factors. These will be
                    discussed further in Chapter in, Important Method Considerations.
Recall - Telephone Survey
                    The telephone survey recall approach consists of making contact with
                    respondents by telephone  and asking questions to elicit information on
                    current or recent fishing trips and fish consumption.  The answers are
                    recorded directly on preprinted questionnaires, usually by interviewers work-
                    ing  from one  central  location under the supervision of an experienced
                    researcher.  Although this approach  "is rapidly  becoming the  principal
                    method of collecting survey  data in research situations where probing or
                    in-depth exploration of the issues is not required" (USEPA, 1984a), interest-
                    ingly, none of the surveys reviewed relied exclusively on data collected by
                    telephone interviews. Instead, these surveys combined this technique with
                    either on-site personal  interviews or mailed questionnaires. A comparison
                    study of different protocols  examined  by  USEPA (1984a) found thai a
                    telephone interview based  on a written questionnaire previously mailed to
                    the respondent was the most effective approach because the respondent had
                    time to review the questions and survey information. Often, the telephone
                    was used to gather information on  non-response bias or to confirm, adjust,
                    or add  to data collected in the field  (see West, 1989b; National Marine
                    Fisheries Service, 1991). Telephone surveys may minimize recall bias and
                    achieve a better overall response than mail surveys because the personal
                    contact involved may encourage  the respondent's participation and jog
                    his/her memory.

                    Telephone surveys may be appropriate for collecting certain types of infor-
                    mation where long-term recall or familiarity with certain facts is not required
                    (such as species names offish caught/consumed). SRI International (1980)
                    found that a 7-day recall period could be quite inaccurate; however, West et

-------
al. (1989b) observed that a 7-day recall period was as accurate as 1-day recall
(see discussion under "Recall - Mail Survey").  Fisher (1988) found that
single-day, 24-hr recalls could be used to estimate mean intakes of population
groups if the days were distributed throughout the year and if the survey
population were large enough. Although large numbers of respondents could
often be reached at a cost savings over personal interviews (e.g., the National
Marine Fisheries  Service studies), the types of information that could be
reliably collected by this approach were limited. For example, anglers may
not  divulge their fishing sites or give  accurate answers to certain
sociodemographic questions.  To  maintain cooperation, each interview
lasted no longer  than  10 minutes and therefore the questions were few
(although other successful telephone surveys have used longer interviews).
Hence, the surveys examined  in this  review usually relied on other ap-
proaches as their primary means of gathering data.

A number of problems were found in the use of telephone surveys, including
difficulty in scheduling to make contact with selected respondents, absence
of respondents at time of calling, unlisted numbers, and lack of a phone. The
last  could be a problem when trying to include  low-income, suspected
subsistence, anglers in the sample. Wendt (1986) observed that low-income
anglers consumed more freshwater fish than those with higher incomes.
List-assisted dialing, in which  respondents were  identified from lists  of
licensed  anglers or participants in fishing tournaments, or other such pre-
selected  lists, was considered to be better than random-digit dialing tech-
niques (see  Brown, 1981, for  an example of the  latter)  because specific
populations  of anglers could be more easily identified.  Computer-assisted
survey techniques (e.g., Computer Assisted Telephone Interviewing  or
CATI, USEPA, 1984a) may be more efficient and less prone to errors made
when transferring written data to computers since printed questionnaires are
not used and the information is directly entered into the computer during the
interview.  If extensive narrative questioning is the basis of the survey,
however, the use  of CAT! could compromise data  collection.  Verification
of the information given is important and could require much additional
work.

Advantages:

•  The telephone  survey can assess region-specific consumption rates, de-
   pending on how the respondents are selected, i.e., determined by residence
   or proximity to a particular waterbody.
•  This approach  can target and identify specific subpopulations of concern
   when these populations can be preselected on some basis or when specific
   limiting questions are included on the surveys.
•  This approach is generally less expensive (by approximately one-half)
   and less time-consuming than personal interviews (since less training of

-------
                       interviewers is required and travel costs are not necessary),  so  large
                       numbers of respondents may be contacted (see USEPA, 1984a).
                    •   A high rate of success for completing interviews is likely, although the
                       sucess rate is 5 percent lower than that for personal interviews (USEPA,
                       1984a) because of lack of personal contact.
                    •   Sensitive information may  be obtained more easily than with  other
                       approaches.
                    •   This approach provides immediate responses to questions, so  analyses
                       may be completed more quickly.

                    Disadvantages:

                    •   Interviewers cannot reach people who do not have phones or those with
                       unlisted numbers.
                    •   Interviews may need to be limited in scope and length, so the number of
                       questions must be carefully chosen.
                    •   It is difficult to verify information given.

                    To Solve These Problems:

                    •   Use telephone interviews only as a follow-up to collecting information by
                       other approaches.
                    •   Use other approaches to contact low-income people.
                    •   Use random-digit dialing to reach those with unlisted numbers (USEPA,
                       1984a),  although considerably more effort may be required  to  reach
                       members of the target population.
                    •   Have one very specific objective for the survey, such as fishing  activities
                       or fish consumption patterns, to limit the number of questions.
                    •   Carefully design the survey to  examine  specific subpopulations and
                       carefully prepare the  questions to be asked to obtain optimal responses
                       and to serve as self-checks on information given.
                    •   Use combined mail/telephone techniques to provide questions and visual
                       aids or other information prior to contact.
Recall - Mail Survey
                    A number of surveys used self-administered mailed questionnaires to obtain
                    information from recreational anglers. As noted by USEPA (1984a), these
                    mail surveys are best for collecting detailed technical data, especially if the
                    respondents need to think about the questions or consult their records. The
                    types of information ranged from simple creel census harvest/angler use data
                    (Swanson and  Stephenson,  1982) to more detailed data on fish meals
                    consumed by the household and methods of cooking. The Wisconsin survey
10

-------
(Fiore et al., 1989) additionally obtained blood samples for chemical analyses
from some of the respondents who agreed to participate in a follow-up study.
A mail survey by Diana (1989)  investigated  behavioral groupings that
indicated compliance with fish consumption advisories and respondents'
knowledge of the fish contaminant situation. The available sample popula-
tion for mail fish  consumption surveys was most often  identified from
records of anglers holding State fishing licenses and was sometimes geo-
graphically stratified to target those anglers nearest waterbodies of concern
(e.g.,  the Great Lakes,  coastal  counties, specific rivers).  The actual costs
associated with this  method will vary with the length of the survey and
number of questionnaires sent, the number  of reminders,  and the type of
follow-up performed.

Success rates for the return of completed questionnaires varied widely. SRI
International (1980) considered an 80 percent response rate to be acceptable,
but many surveys fell far short of this goal.  Cox, et al.,  (1987) distributed
three sets of questionnaires with the Guide to Eating Ontario Sport Fish.  In
1978, questionnaires were sent randomly to people who had requested a
guide in response to newspaper advertising and 876 (44%) responses were
received.  In 1983 and  1986, the questionnaire was included in the back of
the guide, and 807 and 1483 responses were received, respectively. The most
recent mailing in 1989 included 100,000 questionnaires placed in the book,
but only 913 responses  were received (Cox et al., 1990).  This response rate
indicated that other methods, including providing incentives or contacting
nonrespondents, would be necessary to improve the sample  size (C. Cox,
Ministry of  the Environment, Toronto, Canada, personal communication).
However, direct mail questionnaires were much more effective than the insert
questionnaires in the guide.  Questionnaires were mailed to Great Lakes
salmon anglers using randomly-selected names from fishing derby entry
forms, together with an informative covering letter and postage-paid return
envelope.  The return rates for  three mail-outs (600-800 names each) were
65.6 percent, 67.3 percent,  and 71.8 percent (Cox and Johnson, 1990).
Questionnaires sent to  a  10 percent random sample of Arizona's  resident
1980  Class A and F license holders with a postage-prepaid return envelope
resulted in only a 35 percent return rate (Swanson and Stephenson, 1982).
Most  of the other surveys reviewed for this report had higher return rates,
but they required relatively  more money and time and included  advance
letters, stamped return  envelopes, reminder letters or postcards, a second
mailing of the survey to  nonresponders and, finally, follow-up telephone calls
to check on  non-response bias.  A variation of this technique, known as the
"Dillman Method" (Dillman, 1978), in which advance notices and several
reminders are also mailed, increased the response rates up to 47-64 percent
(Fiore et al., 1989; Connelly et  al., 1990; West, 1989a; Chem-Risk, 1991a).
Babbie (1973) reported greater success when the survey questionnaire was
either personally delivered to the respondent or picked up later.
                                                                                      11

-------
                    The information collected in the mail survey approach is typically based on
                    recall periods of days to months, up to 1 year (Fiore et al., 1989). Thus, these
                    surveys are all subject to problems of longer term recall accuracy. West et
                    al. (198%) examined several possible modes for conducting their survey and
                    concluded that of the the 1-day recall, 7-day recall, or 7-day diary record, the
                    7-day recall would be best The 7-day recall proved to be as accurate as the
                    1-day recall when determining group means and was more representative of
                    fluctuations over time. They noted, however, that this time period could be
                    subject to "telescoping," in which respondents tend to include events from a
                    longer  time frame than is called for. West et al. (1989b) used the "bounded
                    recall" technique to minimize telescoping by first having respondents mark
                    out in  a  one-week  calendar the meals at which fish  were eaten before
                    providing detailed consumption information. The ChemRisk (199la) study
                    noted that in addition to the length of the recall period (up to 1 year in this
                    survey), the self-reporting nature of the mailed questionnaire survey, social
                    desirability of the  sport (prestige bias), importance of fishing to the individ-
                    ual,  and  frequency  of fishing trips also contributed to overestimates of
                    consumption.  Avid anglers were more successful and therefore consumed
                    more; consequently, 10 percent of the anglers consumed 90 percent of the
                    fish in  that study.

                    The mailed questionnaire surveys did target recreational anglers but usually
                    did not specifically examine the occurrence of subsistence fishing (except to
                    include questions  from which analysts might infer subsistence fishing, such
                    as income levels).  Only a few studies linked specific waterbodies to the
                    consumption offish from those waterbodies (e.g., ChemRisk, 1991a; see also
                    Connelly et al., 1990, which targeted Lake Ontario fish consumption; other
                    Great Lakes  examined by Fiore et al.,  1989; West et al.,  1989b).   The
                    information collected to calculate fish consumption rates varied in complex-
                    ity and ease of analysis, but certain assumptions needed to be made to cover
                    recall bias identified from follow-up surveys.

                    Advantages:

                    •  Mail surveys can assess region-specific consumption rates, depending on
                       how the respondents are selected (obtaining addresses from license appli-
                       cations, fishing tournament entries, etc.).
                    •  This approach  can target and identify specific subpopulations of concern
                       when  these populations can be preselected on some basis or when specific
                       limiting questions are included on the surveys.
                    •  This approach is the least costly since no interviewers are required except
                       for obtaining follow-up information.  Large numbers of respondents may
                       be contacted over a broad area (see USEPA, 1984a).
                    •  Respondents are most likely to provide honest answers  and fewer "so-
                       cially-desirable"  responses (USEPA, 1984a).
12

-------
•   Complex technical data may be obtained because the respondent can take
   time to consider the questions asked and consult other sources if neces-
   sary.
•   The survey may cover more types of questions, so more than one objective
   may be evaluated.

Disadvantages:

•   Mail surveys cannot reach people who lack mailing addresses, such as
   migrant workers. If addresses are obtained from specific sources, such as
   licensed  anglers,  the survey will miss unlicensed anglers and others
   possibly at high risk from fish consumption.
•   Questions must be carefully designed to compensate for the lack of social
   interaction provided by telephone or personal interviews and must provide
   adequate instructions to elicit satisfactory  responses  and motivate the
   respondents to cooperate (USEPA, 1984a).
•   Questions need to be limited in scope and complexity, preferably requir-
   ing only short answers or checking off multiple choices, to maintain
   cooperation by the respondent.
•   Voluntary mail surveys require substantial follow-up efforts or incentives
   to achieve reasonable response rates (either by conducting telephone
   interviews or by offering  the respondents the choice of phoning in their
   answers).
•   A mail survey is  likely to produce a  higher number of inaccurate and
   incomplete responses because it lacks the personal contact provided by
   other approaches to instruct and motivate (USEPA, 1984a).
•   This type of survey may miss respondents  who are illiterate, who have
   difficulty understanding the questions, or who  cannot read the language
   in which the questions are written.

To Solve These Problems:

'   Use  mail surveys in conjunction with telephone interviews or other
   approaches to check on non-response and recall biases.
•   Increase the return success rate by sending out several waves of follow-up
   reminders, conducting follow-up telephone interviews, offering respon-
   dents the choice of phoning in their answers, personally picking up the
   completed questionnaires, or using incentives.
•   Use other approaches, such as personal interviews, to contact low-income
   people or subsistence anglers, or those who cannot fill out the question-
   naires because of literacy  problems or language differences.
                                                                                       13

-------
                       Carefully design the survey to examine specific subpopulations, and
                       carefully prepare the questions to be asked to obtain optimal responses
                       and to serve as self-checks on information given (see USEPA, 1984a).
                       Carefully plan and pretest the questions to be answered to minimize the
                       length of recall time required.
                       See Babbie (1973) for more information on how to conduct self-admin-
                       istered questionnaire surveys.
Recall - Personal Interview
                    Personal interviews were conducted in a variety of surveys to obtain infor-
                    mation ranging from angler use to fish consumption patterns. The interviews
                    occurred at known fishing sites (which personnel had to cover up to 18 hours
                    per day to contact early morning and late evening anglers) or at home. Home
                    interviewees were selected from samples of licensed anglers identified by
                    State fish and game departments, or households located near fishing locations
                    (Wolfe and Walker, 1987). Subsistence anglers were also specifically iden-
                    tified  because they were participating in special programs, such as that
                    conducted by the Expanded Food and Nutrition  Education Program  of
                    Cornell University  Cooperative Extension Service (Wendt, 1986). In all
                    cases, the respondents were asked a fixed set of questions and the answers
                    were recorded on the questionnaires.

                    Although the questions in most on-site interviews were limited to those of a
                    creel census nature (see "Creel Census" below), a few interviews collected
                    data for fish consumption rates. Three surveys were conducted in the Puget
                    Sound area (McCallum, 1985; Landolt et al., 1985, 1987). The Landolt et
                    al. studies targeted shoreside anglers and boating anglers as they returned to
                    boat ramps. Over 4,000 shoreside angler interviews were conducted during
                    the first year, but only 437 boating anglers were interviewed the second year.
                    Landolt et al. noted that the latter interviews produced fewer cooperative
                    respondents (only 83 percent), and the anglers either refused to give the exact
                    sites  of their fishing activities  or only vaguely  identified  them.  The
                    McCallum study interviewed all anglers, crabbers, and clammers from one
                    end of the beach or pier to the other end at specific sites throughout the year.
                    The survey was advertised at local marinas and bait shops  to aid in eliciting
                    cooperation from respondents.

                    Smith and Enger (1988) conducted 703 interviews at fishing sites along the
                    Tittabawassee River  in Michigan.  Although fishing bans had been  an-
                    nounced because of contamination with dioxins, the survey found that fishing
                    effort had increased as the result of the successful restocking of the river with
                    walleye.  Only 2.7 percent  of the anglers interviewed said that they were
                    fishing for food, but  the authors suspected  that  this was not an accurate
                    percentage. This particular survey relied on long-term (up to 1 year) recall
                    of the anglers' fish consumption habits and did not target the actual catch of
14

-------
the day. Despite recall bias and the fact that weather conditions prohibited
fishing in some parts of the river later in the summer (suggesting that this
was not the best representation of the normal fishing effort and catch there),
the study did target a specific local population where chemical  exposure
through fish consumption was of concern.

Other types of personal interviews were conducted at home. The Nationwide
Food Consumption Surveys (1977-78, 1987-88), conducted by the U.S.
Department of Agriculture (USDA), used a list of foods to assist the house-
hold respondent in recalling the kind, form, quantity, and cost (if purchased)
of foods used at home during the previous  7  days. The interviewer also
obtained information on those characteristics that might be related to food
consumption (demographic and socioeconomic data). The interviewer then
recorded the preceding day's food intake for each eligible household member
present and instructed each individual to record his or her intake for the day
of the interview and the next day. This procedure provided three consecutive
days of dietary information. The interviewer returned to pick up the records,
and each household received $1 for each record returned  (up to $10 per
household). Data were collected throughout the week. Respondents were
contacted in advance to participate in the survey. The 1977-78  survey
included 15,000 households and 38,000 individuals. The Continuing Survey
of Food Intakes by Individuals, also conducted by the USD A, asked individ-
uals to provide from 1 day to 6 days of dietary  data at intervals of 2 months
over a  1-year period (see, for example, USDA, 1985b, 1986b).  The first
day's data were  collected by personal  interview, with  subsequent data
collected by telephone interview. Unfortunately, these studies did  not target
consumers of recreationally-caught fish, and  the consumption of potentially
contaminated fish could not be determined from the questions administered.

The  U.S. Department of Health and Human  Services (USDHHS,  1989)
survey of fishing patterns and contaminant exposure in Lake Coeur d'Alene,
Idaho,  included personal interviews with 299 households, using recall peri-
ods of up to 1 month.  In another study, Wendt (1986) targeted low-income
freshwater fish consumption by carrying out personal interviews with re-
spondents identified as  participants in the Expanded Food and  Nutrition
Education Program at Cornell University. She found that these interviews
required the presence of aides to serve as go-betweens or to keep children
entertained during the 30-minute interview. They did, however, yield inter-
esting  information on the fish consumption patterns of the participants.
While the on-site interviews could collect information on the species offish
caught and consumed, household interviews suffered from the inability of
the respondents to identify species of fish and to assess recall  bias. Both
on-site and household interviews encountered  literacy (understanding) and
language barriers.
                                                                                      15

-------
                    Advantages:

                    •   Personal  interviews can assess region-specific consumption  rates by
                       targeting the waterbody or residence of the respondent.
                    •   This approach can also identify specific subpopulations of concern by
                       obtaining data from known contaminated fishing/shellfishing sites or by
                       using specific programs to identify potential respondents.
                    •   Personal  interviews can provide first-hand observations of the respon-
                       dents and the interview sites.
                    •   Literacy and language barriers may be more easily overcome using this
                       approach.
                    •   Recall bias can be minimized by providing appropriate visual aids (for
                       portion or meal size) or basing the survey on the fish caught at the time
                       of the interview.
                    •   This approach has a  high  rate of success for completing interviews
                       because of personal contact.  Interviewers also can clarify confusing
                       questions or neutrally probe for answers.
                    •   Verification of information may be easy, especially if data collected are
                       based on the actual catch of the day. It is also relatively easy to use special
                       techniques such as visual aids and probing.

                    Disadvantages:

                    •   The number and complexity of survey questions may need to be limited
                       so that surveys can be performed quickly, depending on the respondents'
                       availability and interest.
                    •   Personal interviews are the most costly approach, requiring the coordina-
                       tion, hiring, training, and close supervision of interviewers and field staff
                       at more than one location, as well as additional paperwork to control the
                       fieldwork and processing operations (USEPA, 1984a).

                    To Solve These Problems:

                    -   Conduct  the survey in different languages (or use bilingual interviewers
                       or translation assistance from other family members or associates) and
                       provide visual aids such as fish models to assist in obtaining information
                       from the  respondents.  McCallum (1985) cited problems with questions
                       on the parts of the catch eaten, fishing frequency, and how the fish were
                       prepared for eating.
                    •   Use the "clustering" technique to limit the  number of sites or group the
                       residences where  interviews will be held, thus reducing costs.
16

-------
   Carefully prepare the survey questions to minimize the length of the
   survey yet provide the precise information needed to achieve the objec-
   tives.
   Provide adequate training (including practice interviews) and supervision
   of interviewers throughout the survey.
   See Babbie (1973) or other survey methodology texts for more informa-
   tion on how to conduct interviews for surveys.
While complete food consumption diaries havebeen used in general nutrition
surveys (e.g., USDA, 1983a,b), none of the fish consumption surveys
examined for this report employed this approach for obtaining data.  Block
(1982) found that diary methods were subject to selective forgetting or lapses
in diary keeping even after only a few days, and it was  difficult to get
respondents committed to the project, especially if no personal contact was
involved.  However, Fisher (1988) noted that such records, kept at home for
periods of days to months, can provide reliable data on patterns of food
intake.  This  approach does require respondent literacy,  and the  act of
keeping records itself may affect dietary practices, so there is a need to
analyze for changes that may occur in motivation of the subject or changes
in food records  (Fisher, 1988). West et al. (1989b) observed  that  earlier
studies have shown that the most valid and accurate studies offish consump-
tion have been diary studies  involving repeated personal contact with the
study subjects (e.g., Humphrey, 1976,1983).  Such contact probably maxi-
mized motivation and minimized alterations  in diet and recording by the
respondents.  Diary records may provide sound information for examining
fish consumption patterns if the survey is carefully designed and monitored.

Advantages:

•  The diary approach can assess region-specific consumption rates if re-
   spondents are selected appropriately.
•  Diaries can provide data over long  periods of time for particular  sub-
   populations of concern if such subpopulations have been appropriately
   preselected.
•  This approach is less expensive than personal interviews.
•  The diary approach can be used with persons inaccessible by telephone.
•  Large numbers of respondents may be included.
-  This approach results in minimal recall bias, although other potential
   sources of error or alterations in record-keeping may occur.
                                                                                   Diary
                                                                                       17

-------
Creel Census
                    Disadvantages:

                    •  Interviewers must be trained to teach the respondents how to complete
                       the diary.
                    •  Using the diary approach requires respondent literacy, a high degree of
                       motivation, and constant monitoring to maintain consistency in the data
                       collected.
                    •  The act of keeping records can affect dietary practices.

                    To Solve These Problems:

                    •  Combine the diary approach with other approaches (such  as personal
                       interviews) to provide additional, in-depth, or longer term information on
                       fish consumption patterns.
                    The creel census approach is used by fishery managers to obtain harvest data
                    collected on-site, from single anglers (hook and line, castnet, clam rake, etc.)
                    or from larger scale commercial-type operations (trawl, gill nets, etc.) that
                    obtain fish for a specific community. This information is then used to make
                    management decisions for optimal utilization of the resource. For example,
                    a number of creel censuses have been done in Georgia, such as Scott (1981),
                    Hottell et al. (1983), Schmitt and Hornsby (1985), Fowler and Holder (1987),
                    and Spencer (1987).  These on-site interviews examined the species fished
                    for, species caught, weight caught, method, bait, origin, and type of fishing
                    (boat, bank, dock, bridge), but did not include questions on fish consumption
                    or sociodemographic data.  These surveys also did not  distinguish whether
                    the fish caught were going to be consumed, given away, sold, or released.
                    Other surveys of this type include Mullis (1989), who obtained data on angler
                    effort associated with striped bass fishing on the Roanoke River, and
                    Ranthum (1975), who recorded lengths and weights of species offish caught.
                    ChemRisk (1991a) found that creel censuses were often used to estimate
                    angler use and fish harvest from specific waterbodies, but noted that because
                    individual anglers may fish in more than one location, such a survey might
                    not completely characterize the total freshwater fish harvest or consumption
                    for anglers and others sharing their catch.

                    Diverse time periods have been selected for creel  census interviews.
                    Ranthum sampled two consecutive census days, with a varied schedule (7 to
                    11  am, 11 am to 3 pm, or 3 to 7 pm) for on-site interviews during the 2-month
                    study period. The Wisconsin series of creel censuses (e.g., Thuemler, 1981;
                    Heizer 1986,1988; Schumacher, 1987) used a stratified random sampling
                    schedule for 2-hour periods. Counts of anglers present on the lake were made
                    at 2-hour intervals from 7:00 a.m. to 7:00 p.m., with a final count at 8:00
18

-------
p.m. Between counts, anglers were interviewed to determine the number,
length, and species offish caught and the angler's residence. More censuses
were scheduled for weekend days and holidays and for the entire opening
weekend of the fishing season.   Fifty percent of the remaining weekend
periods and 30 percent of the weekday periods were sampled, with an equal
amount of effort given to each month and each hourly time period. Brown
(1981) looked at recreational shrimping along the Gulf coast by allocating
most interviews to the opening days of the seasons for brown and white
shrimp, with the remainder of the interviews taking place on weekend days.
Chandler and Brown (1978) examined  potential  problems that might be
encountered while collecting marine recreational fishing  and shellfishing
harvest data for the Atlantic and Gulf coasts. They noted that the selection
of fishing sites and times with the highest levels of fishing activity was best
for  obtaining the maximum number of  interviews that  needed  to be con-
ducted.

Five creel census surveys that attempted to obtain direct information on the
fish consumption patterns and habits of recreational and subsistence anglers
were conducted. An early study by Pierce et al. (1981) sampled fishing and
shellfishing effort at four subareas around Commencement Bay in Washing-
ton State that were suspected of having potentially hazardous seafood. The
on-site interviews were followed by telephone surveys to determine whether
the  fish that were caught that day had been eaten and how they had been
prepared.  The study suffered from problems in changes of sampling sites
and the number of survey days during the study, but did provide data on fish
consumption patterns for the area.

Puffer  et al. (1982a) examined the  consumption of potentially hazardous
marine fish and shellfish from 12 sites in the Los Angeles area known to be
both fished and polluted. Teams of two surveyors conducted 1,059 inter-
views with anglers on  different days of the week and at  different times (61
percent of interviews were held during the week, 39 percent on  weekends,
for  a total of 400 site  visits).  Incentives for cooperation  included fishing
maps, copies of regulations, and/or recipes sent following the interviews.
Photographs were frequently taken to ensure the correct taxonomic identifi-
cation of the fish caught, to document site conditions, and to confirm sport
anglers' counts.  Initially, the surveyors recorded the number  of anglers
present at a site and their sex, race, and approximate age.  Then only those
anglers who had actually caught fish were interviewed (if more than 20 had
caught fish, a systematic sampling approach was used), resulting in a bias
toward frequent anglers.  The more successful the fishing, the more fre-
quently the fisherman was inclined to fish.  The information necessary for
accurate fish consumption rate calculations was weak, and there were prob-
lems with having to change sites during the study because  of weather, sewage
overflows, and chemical disposal problems.  However, the study did indicate
                                                                                     19

-------
                    a need to assess health risks to consumers at specific sites and for specific
                    subpopulations.

                    Belton et al. (1985) examined fishing effort at sites known to be chemically
                    contaminated and where specific fishing bans had been instituted along the
                    Raritan River and other sites in New Jersey. An interesting aspect of this
                    study was that the interviewers initially conducted only visual observations
                    and informal interviews at the six sites (cross-cultural anthropological field-
                    work techniques) because they were concerned that this population of anglers
                    might be leery of formal surveys and distrustful  of outsiders.  Later,  a
                    subsample of the fishing population was selected to answer the questionnaire
                    by personal interview at the site. A monetary incentive of $10 was provided
                    for those who agreed to  participate in a  long-term study.  Although the
                    questionnaire was fairly thorough and easy to interpret, the data collected for
                    fish  consumption rate calculations were  incomplete.  The questionnaire
                    included no questions on size of portions consumed per meal or species.
                    Assumptions were made  to allow such calculations,  but data from other
                    studies were used to evaluate the health risks involved.

                    Kleinschmidt Associates (1989) examined the fish consumption patterns of
                    anglers from two areas on the Androscoggin River in New Hampshire. The
                    areas chosen were the relatively pristine section of the river north of Berlin
                    to the Errol Dam near the Maine-New Hampshire border and the river below
                    the James River Corporation's paper mill at Berlin.  The New Hampshire
                    Division of Public Health Service had  issued a fish consumption advisory
                    for that portion of the river downstream from  Berlin.  While  spending
                    approximately equal periods of time in each  area  on three weekends in
                    August, the interviewers were able to find only three anglers fishing in the
                    lower area, resulting in 5.26 hours per initiated interview there, compared to
                    66 interviews in the upper area taking approximately 0.3 hour per interview.
                    Interviewees  in the lower area did not  consume  the fish they caught there
                    because of health/safety concerns. The report  concluded that the fish con-
                    sumption advisory had been effective. The consumption data collected were
                    based on recall of how often the anglers ate fish from the particular  section
                    of the river (per week or month) and an average estimate of the meal  size (in
                    number of 8- to 10-inch  fish).  These data were  easy to obtain, but the
                    ambiguous nature of some of the questions  makes calculations  of fish
                    consumption rates impossible.

                    Finally, a study was conducted to estimate  consumption rates of selected
                    chemicals from contaminated fish caught in San Diego Bay (San Diego
                    County Department of Health Services, 1990).  The survey protocol was
                    based on that of Landolt et al., (1985) and questions covered species, weight,
                    and  length of fish caught and eaten by  Bay  anglers and others in their
                    households, demographics of the angler population, and characterization of
                    fish  consumption rates and patterns.  Interviews were obtained from 369
20

-------
anglers at popular pier and shoreline fishing sites and boat launches over a
one-year period, but only 59 interviews (representing 195 potential consum-
ers) contained all of the data for calculating individual consumption rates.
As in the Landolt et al., (1987) study, samples offish (obtained separately)
were  analyzed for  chemical contaminants and these data were used to
estimate various subpopulation exposures and potential risks of adverse
health effects.

Advantages:

•  The creel census approach, as a personal interview approach, can assess
   region-specific consumption rates by targeting specific waterbodies.
•  This approach can also identify specific subpopulations at high risk by
   obtaining  data  from  actual  anglers  at  known  contaminated  fish-
   ing/shellfishing sites.
•  Creel censuses can provide first-hand observations of the respondents,
   their fishing activities, and the interview sites.
•  Recall bias can be minimized  by providing appropriate visual aids (for
   portion or meal size) and by basing the survey on the fish caught at the
   time of the interview.
•  The rate of success for completing interviews is high because of personal
   contact.
•  Verification of information may be easy, especially if data collected are
   based on the actual catch of the day. It is also relatively easy to obtain
   sensitive information and to use special techniques such as visual aids and
   probing.
•  When the appropriate questions are included, this type of survey can more
   accurately assess fishing behavior by anglers, fish species can be  more
   accurately identified to species, and important information on consump-
   tion rates and characteristics of the anglers can be  easily obtained and
   verified.

Disadvantages:

•  The number and complexity of survey questions must be limited so that
   surveys can be performed quickly.
•  Interviewers may encounter language barriers.
•   Creel censuses are costly because they require the coordination, hiring,
   training,  and close supervision of interviewers and field staff for quality
    control, as well as additional  paperwork  to control the fieldwork and
    processing operations.
                                                                                        21

-------
                    To Solve These Problems:

                    •   As may be done for personal interviews, conduct the creel census in
                       different languages.
                    •   Provide visual aids to increase the response rates while minimizing the
                       level of effort and time needed to conduct the interviews.
                    •   Use the "clustering" technique to hold down costs by limiting the number
                       of sites where interviews will be held. Try to select sites where there will
                       be more respondents over longer periods of time or at different seasons
                       in order to limit the time needed to "search" for respondents at a site.
                    •   Carefully prepare the survey questions to minimize the length of the
                       survey and the time needed to conduct each interview yet still provide the
                       precise information needed to meet the objectives.
22

-------
            III.  IMPORTANT METHOD CONSIDERATIONS
Fisher (1988) noted that, with regard to fish consumption surveys,  two
considerations were important in using available data or in designing ap-
proaches to collecting data: (1) the methodology used to collect data and (2)
the population sampling techniques. He explained that the former was more
complex because the dietary data collection method selected would depend
on both  the population surveyed and the purpose for deriving the estimate.
Population sampling techniques are important in surveying population sub-
sets that preferentially consume fish because representativeness becomes
important in statistical analyses. Fisher (1988) further stated that since fish,
as a food item, has unique attributes, any of the approaches taken to collect
these  data will have inherent  advantages and disadvantages in regard to
determining preferential fish consumption.

Many of the recent surveys examined for this report used more than one
approach to obtain information. For example, a simple on-site creel census
might be combined with a telephone interview to include additional charac-
teristics  of the fishing population and to determine whether the catch  was
consumed. Or, a questionnaire mail survey  might ask for general  fish
consumption information and then include a table to be filled in asking for
more specific fish consumption data over a period of several days. Other
surveys  have combined mailed surveys with follow-up telephone surveys to
check on  non-response bias  or to  obtain more complete information.
Springer (1990) used several different types of  questionnaires and either
mailed surveys or conducted personal interviews, depending on the target
audience (the former for recreational anglers, fisheries and health care
experts,  and the latter for migrant workers and low-income individuals) to
investigate risk communication theories and the effectiveness of fish advi-
sories and bans. More than one approach may need  to  be used to make the
survey as informative and useful as possible for the desired objective.  The
choice of approach  will also  depend on  the characteristics of the target
population, data requirements,  obligation to reply, target response rate, time
available,  and funds available (USEPA, 1984a).  Highlights of the  five
approaches examined in Chapter II are presented  in Table 2.

All approaches, however, share a number of common elements including
design, selection of respondents, information to be sought, quality assurance
procedures, and statistical analyses. For the purposes of this document, the
most important  criteria for  preparing a fish consumption survey are the
following:

•   Thoroughness;
•   Applicability to subpopulations of concern;
                                                                                    23

-------
       Table 2.  Comparison of different approaches to conducting fish consumption surveys.

Can assess region-specific
consumption rates
Can target and identify specific
subpopulations of concern
Allows first-hand observations of
respondents and fishing locations
Provides immediate answers
to questions
Easy to verify information
given
Can be used where illiteracy may
be encountered
Success rate for completed,
accurate interviews
Relative cost per interview
Telephone
Survey
yes"
yes"
no
yes
no
yes
high
$$$$
Mall
Survey
yes"
yesb
no
no
no
no
low
$
Personal
Interview
yes
yes
yes
yes
yes
yes
high
-^£4A*fc
1"(**t"lrN'
Diary
Record
yes-
yes"
no
no
yes
no
moderate
$$$
Creel
Census
yes
yes
yes
yes
yes
yes
high
$$
'Depends on how respondents are selected.
"Depends on how these subpopulations are Identified.
                                                  24

-------
•   Scientific/analytical validity;
•   Ease in interpreting results;
•   Reasonableness of assumptions made; and
•   Sufficient data to evaluate potential risks.

This section addresses various elements of survey design and analysis that
must be considered prior to undertaking a survey to estimate fish consump-
tion rates.  Additional information and references can be found in Babbie
(1973), Dillman (1978), and USEPA (1983,1984a).  These resources also
recommend discussing the survey plans with qualified, experienced survey
research consultants and statisticians who can provide answers to questions
that may arise while planning the survey. This step should ensure that the
design will meet the particular survey objectives and adequately sample the
populations of concern.

Survey Design
Fisher (1988) noted that the primary factors controlling the selection of a
strategy to obtain the desired data were (1) the purposes for which an estimate
of usual or preferentially high fish intake is sought and (2) the resources
available to obtain such an estimate. Because of these factors, no one strategy
may fulfill all possible needs for such data, yet each survey may be quite
complex in its own right. He emphasized that the purpose of the survey must
be narrowly defined and tailored to  address the identified needs, such as
whether the survey was to be used for diet/health concerns or risk assessment.

Since differences in preferred species, availability, access, length of fishing
season, and cultural heritage greatly influence freshwater fish consumption
in a particular region (ChemRisk, 1991a), these factors must also be exam-
ined when planning a survey. The period of data collection is important.
Will most of the subsistence and recreational anglers be active during the
summer months only or during the entire open season for a particular fish?
If information is required for an entire year, can recall be depended upon to
provide the answers or should a year-long continuing survey be conducted?
The level of literacy of potential respondents should also be estimated.  Will
the respondent be able to understand  the written or spoken questions?  The
types of questions should be prepared and tested with respect to simplicity
and clarity, as well as their ability to elicit the desired information.  Language
barriers may also exist, further limiting the amount of data that may be
gathered from  important constituencies.  Thus, surveys  may need to be
conducted using more than one approach or in more than one language to
adequately cover all of the subsistence and recreational anglers in a particular
area.
                                                                                       25

-------
                    Another general consideration for survey design is the level of detail re-
                    quired. Although extensive information may be desired for some programs,
                    longer surveys will require more time and resources for the conduct of the
                    survey and analysis of the data. Furthermore, increased length and complex-
                    ity of the survey  design  may  limit the cooperativeness of respondents,
                    resulting in inadequate or incomplete data that are ultimately useless. All of
                    the surveys reviewed for this report experienced problems in getting partic-
                    ipants to cooperate and complete the questionnaire materials.  To improve
                    participation rates, some of the surveys provided incentives, such as lapel
                    pins, maps, additional information in brochures, copies of the survey results,
                    and/or cash. (The relative effectiveness of the incentives was not reported.)
                    Other surveys limited the number of questions but consequently lost valuable
                    information that would have been helpful for fish consumption rate calcula-
                    tions.

                    The approach taken to collect the data will affect the cost of the survey. As
                    a general guideline, personal interviews cost at least twice as much and take
                    twice as long as telephone interviews. Both of these methods are more costly
                    than a mail survey (USEPA, 1983).  However, if personal interviews are
                    clustered at specific locations to maximize the number of respondents to
                    obtain the fish consumption rates  of recreational/subsistence anglers, less
                    time and effort will be involved than if the same number of respondents who
                    fished at those specific sites had to be culled from extensive telephone or
                    mail surveys of the general population or licensed anglers. Some flexibility
                    in the survey design may be required to accommodate any problems that may
                    be encountered.

                    Few of the surveys reviewed for this report include information on level of
                    effort, length of time to conduct the survey, and/or costs.  Limited data were
                    obtained by contacting the individual or agency responsible for the surveys
                    (see Appendix). Costs and levels of effort ranged widely depending on the
                    type of survey and its  extent. Some surveys were conducted by volunteers
                    or graduate students or were conducted as part of normal in-house responsi-
                    bilities.  Hence, it would  be difficult to determine a meaningful "average"
                    cost per survey.  As  noted by  West (1989a), inadequate resources may
                    prevent the researcher from obtaining statistically valid sample sizes. There-
                    fore, it is important to have a clear understanding of the amount of funding
                    and  number of personnel available in order to achieve the best balance of
                    resources for the desired survey. This analysis should determine work that
                    can be done in-house versus work that must be done by outside consultants
                    or volunteers, as well as other pertinent factors that may affect the costs.

                    The survey design should specify the following (see USEPA, 1983 for more
                    information):
26

-------
   The objectives of the survey, clearly  stated in terms  of the kinds of
   information to be collected, the problems to be solved, the hypotheses to
   be tested,  and the  key survey variables.   For example, what are the
   minimum questions that need to be answered? What hypotheses will be
   tested-consumption rates related to income, ethnic group, or frequency
   of fishing, etc.? Each of the key survey variables and the specific data to
   be collected to meet the stated objectives must then be defined.
   The population to be surveyed and the extent of coverage (e.g., regional,
   site-specific, recreational anglers, subsistence fishermen, minority an-
   glers).
   Identification of respondents by probability sampling.  This means that
   every unit (e.g., person/household) has a known, non-zero chance of being
   included in the sample, thus allowing  for statistically  valid inferences
   about the entire population the sample is designed to represent.
   The required level of precision, specified in terms of sampling error-that
   is, the difference between the values and statistics that would have been
   obtained had all the members of the target population been surveyed and
   the values and statistics that were obtained from the sample population.
   This level will depend on the survey purpose, intended use of the data,
   level of effort, and available funds. Nonsampling errors, such as random,
   deliberate, wrong or unintentional replies and systematic one-sided errors
   or biases, must also be considered.
   The target response rate, defined as the ratio of the number of completed
   interviews to the total number of eligible units in the sample.  As noted
   above, a  75 percent response rate is  acceptable  for  an in-person or
   telephone  survey, but a mail survey may receive less than a 40 percent
   response unless telephone calls or personal  visits can raise the initial
   response rate
  Recommendation:

  Follow-up activities should be included with whatever approach  is
  selected to ensure an appropriate response rate, check for non-response
  bias, and confirm data.
                                                                Selection of Respondents
Various methods have been used to select the anglers to be interviewed, (the
sampling frame) depending on the approach to be taken to obtain the data.
The sample may consist only of anglers or may include members of their
households. (USEPA, 1983).  Some national surveys have used random-
digit dialing to obtain their samples.  However, recreational or subsistence
                                                                                       27

-------
                    anglers might not be captured by this technique.  On the regional or local
                    level, lists of sport fishing license holders may be used to obtain stratified
                    samples based on a particular type of license or geographic reference, such
                    as counties located close to the waterbody in question. Intercept or on-site
                    interview approaches may attempt to question everyone, interview only those
                    who have caught fish at the time (non-uniform), or randomly select anglers
                    to be questioned. The Georgia creel censuses used non-uniform probability
                    sampling, interviewing anglers on a predetermined basis up to a certain
                    number required for statistical validity. The number of interviews needed
                    was based on preliminary surveys, and these numbers were readjusted every
                    6 months to reflect possible changes in the fishing population.  Depending
                    on the objectives of the survey, other strategies may be required to obtain
                    samples of recreational and subsistence anglers.

                    Sample selection must be carefully planned to achieve the numbers necessary
                    for statistical validity. Final sample size will depend on the level of precision
                    required for the estimates. The Bureau of Census may be consulted to obtain
                    information about total population and/or subpopulation numbers present in
                    a particular area.  The Bureau can provide breakouts by  age, sex, and/or
                    ethnicity, for a cost  Then probability tables can be generated to determine
                    the minimum numbers of respondents required. This technique can also be
                    used to select subsamples of licensees or other designated groups.  Sample
                    stratification in the Bureau of Census design also  allows  for sampling
                    procedures that are self-weighted.  Additional respondents may need to be
                    added to the pool to provide sufficient numbers if there are problems in
                    obtaining the interviews (e.g., loss of questionnaires in the  mail, non-re-
                    sponders, language barriers, etc.). Babbie (1973) contains detailed informa-
                    tion on the selection of a survey population, probability sampling theory, and
                    sampling distributions.
                     Recommendation:

                     It is essential to work with a statistician prior to initiation of the study to
                     ensure that appropriate and representative sample sizes of the population
                     to be examined are obtained by the technique that is selected.
Information Sought
                    The type of data to be collected will depend on the purposes of the study, as
                    well as the complexity and length of the survey (single or multiple seasons,
                    multi-year) to be conducted. Some of the surveys reviewed for this document
                    were designed to collect information for purposes other than measuring fish
                    consumption and therefore did not ask the "right" questions (e.g., creel
                    censuses). Many assumptions must be made to obtain estimates  of fish
28

-------
consumption in these cases, and errors of overestimation or underestimation
may be  large.  Questions on sociodemographic characteristics, fish-
ing/shellfishing activities, preparation and consumption patterns, and aware-
ness of fish advisories and bans can all contribute data that may be used to
analyze fish consumption rates.  Examples of questions that  need  to be
answered or issues that  need to be resolved for fish consumption rate
determinations are presented in Table 1.

Sociodemographic variables such as age, community type, educational level
of head of household or respondent, ethnic origin or race, family size and
composition, geographic region, income, occupation of head of household,
and religion may influence patterns of intake.  Current employment  status
may affect the  amount of time spent in fishing/shellfishing activities and the
amounts of these foods consumed. However, this type of information may
be difficult to obtain or controversial (for example, income level, race, ethnic
group,  language  spoken  at home, religion).  Thus,  including
sociodemographic questions may decrease the number of successful inter-
views or completed questionnaires.

Although the  residence of anglers may indicate that they are fishing at
potentially contaminated  waterbodies, it is  more important  to determine
whether there is a possible contaminant exposure problem at  the particular
fishing site(s).  Other information on fishing activities, such as distance from
residence and mode of fishing, may be irrelevant for some surveys but may
provide important data for others, depending on the objectives of the study.

Accurate identification of the type and amount offish caught and consumed
is important. Levels of contamination vary with the preferred habitat  of the
organism, its trophic level (for example, bottom-feeding versus planktivore),
and its lipid content, as well as the waterbody. Pictures of the fish may need
to  be taken for verification if identification in the field is impossible. The
level of detail required for fishing activities may  also be difficult to obtain,
especially since anglers often keep  their fishing spots secret   Providing
visual aids such  as maps of fishing  areas may produce more accurate
information from respondents when recalling the  location(s) of fishing
activities. Persons involved in illegal sales of fish from contaminated sites
would probably refrain from answering these questions.

Actual amounts of fish consumed need to be determined as  accurately as
possible. Silverman (1990) noted that "data on average meal size is glaringly
absent..." from most studies.  Pictures or models of portion  sizes may be
provided to aid in  estimating the size offish portions consumed (although it
should be noted that this may  vary with the species involved).  West et al.
(1989b) included pictures of an 8-ounce steak and fillet portion on a normal
size plate for comparison, then estimated "more" as 10 ounces and "less" as
5 ounces for their survey.  This 8-ounce portion size was derived from
                                                                                      29

-------
                    restaurant surveys and the USDA (1983a,b) study. Cox et al. (1990) found
                    that 8 ounces was the portion most commonly reported.  These and other
                    product identifiers can be used to provide the level of detail required in the
                    answers.  The survey should also specify whether the amounts eaten are to
                    be determined for the angler only or for all household members.

                    Other information that may be useful in examining consumption patterns and
                    levels of exposure to potentially contaminated fish includes the following:

                    •   Whether a part of or the whole animal is eaten: for fish, muscle only, skin,
                       head, entrails, broth; for shellfish, muscle, hepatopancreas, entrails.
                    •   Method(s) of preparation (raw, dried, canned, smoked, steamed, boiled,
                       baked, fried, stewed, marinated, barbecued; or whole, fillet, skin removed,
                       etc.); accompaniments used  in preparation  (butter, lemon/lime juice,
                       tomato sauce, garlic, etc.).
                    •   Other types of aquatic organisms consumed from the same site(s) such as
                       snapping turtles, frogs, sea cucumbers, sea urchins, squid, algae or other
                       vegetation, etc. (quantities, frequency consumed).
                    •   Whether fish are also  consumed from other sources (market, restaurant,
                       or gifts).

                    Finally, the objectives of the survey may require an assessment of the
                    awareness of health advisories and an understanding of contamination issues.
                    Questions may  address knowledge of fish consumption advisories or  bans
                    present in a particular fishing area and behavioral modifications resulting
                    from these concerns.  The impact of advisories may be reflected in changes
                    in fishing locations and in the species, sizes, and parts  of fish kept and
                    consumed.  If these issues are not addressed, biases may be introduced into
                    the survey.  Several surveys have also investigated links between diet and
                    health. As noted by Fisher (1988), such studies may include questions on
                    medical history, a physical examination for health status and clinical signs
                    of deficiency or toxicity,  food and nutrient intakes, body measurements, and
                    hematological and biochemical tests (USEPA, 1984b; Fiore  et  al., 1989).

                    For any approach, requiring long-term recall may produce substantial bias
                    in estimates of fish consumption and other variables. Although difficult to
                    document, recall bias can be affected not only by anglers' attitudes toward
                    their sport, their skill, and their investment in time and equipment, but also
                    hy the value of the fish to the family diet (P. Shubat, Minnesota Department
                    of Health, Minneapolis,  personal communication).  Thompson and Hubert
                    (1990) reported that anglers inflated the amount of time spent fishing. Other
                    self-report surveys have documented overestimates of fishing statistics (see
                    Westat, Inc., 1989, ChemRisk, 1991a).
30

-------
Information requirements for the survey must be carefully planned based on
the survey  approach to be used.  For example, food frequency recall ap-
proaches may need to include some probing questions to jog memories for
consumption of fish meals over extended periods. Personal interviews and
telephone surveys may also need such "neutral probing" to obtain complete,
clear, relevant, and specific answers. Survey questions must be worded for
understandability.  The  questions must then be pretested and revised as
necessary before beginning the full survey. Actual phrasing of the questions
is critical to obtaining usable data. For example, "How many fish did you
eat?" will require assumptions on portion size, frequency of consumption,
and preparation for cooking (whole, steak, fillet). Note that "preparation" of
the fish may mean  different things if the purpose is for cooking (scaled,
filleted), for eating (pan fried, broiled), or for preservation (salted, smoked);
this information, however, may provide acceptable data on portion  sizes
consumed or whether the preparation minimized exposure to contaminants.

It is equally important to ensure that the wording of the question will provide
the correct data for evaluation,  rather than give a vague or potentially
uninterpretable response that becomes useless when calculations are  to be
made. The information to be collected may need to be modified depending
on whether the survey will be answered directly by the respondent or through
proxy.  Conducting interviews in more than one language  may also be
necessary to reach the population of concern (for example, see Puffer et al.,
1982a, and the National Marine Fisheries Service studies).  Babbie (1973)
discusses the construction and sequencing  of questions in order to avoid
many common pitfalls, as well as providing sufficiently clear instructions so
the survey can be completed. He also suggests methods for conducting and
evaluating pretests and pilot studies (see also Sudman and Bradburn, 1982).
  Recommendation:

  The selection and phasing of questions to meet the survey objectives is
  critical. Questions must ultimately be used only for the purposes intended
  and not stretched to try to fit other unrealistic purposes, thus introducing
  serious biases.  The importance of consulting experts in nutrition and
  survey design and analysis cannot be overstressed.
Appropriate quality assurance procedures must be incorporated into both the
planning and execution of the survey. The types of quality controls proposed
will ultimately depend on the approach to be taken, but should include the
following (USEPA, 1983):
                                                                       Quality Assurance
                                                                                      31

-------
Statistical Analyses
                    •  Validation of at least 10 percent of the interviews to verify that the
                       interviews did take place and that information was accurately obtained
                       and recorded.
                    •  Manual checking of questionnaires for completeness and proper entry of
                       answers.
                    •  Checks on the manual coding operations and comparisons of results and
                       error rates in interviews conducted by different interviewers.
                    •  Verification of correct data entry; for example, by having all of the data
                       entered twice and then compared.
                    •  Computer edits to detect inadmissable and out-of-range values.

                    Other quality assurance considerations include the qualifications and training
                    of interviewers. (Can they conduct interviews pleasantly and correctly? Can
                    they identify the fish to species?) Close supervision should  be provided
                    throughout the survey to make sure that all data are entered and recorded on
                    the forms correctly, all interviewers are performing similarly,  and each
                    interview session is conducted as the previous ones were conducted. Another
                    technique is to use responses given in one category to check those in another,
                    such as fishing history (catch rates, locations) vs. fish consumption (amounts,
                    species).  For mailed questionnaires, each questionnaire should be assigned
                    an identifying number both at the time of delivery to the respondent and on
                    completion or receipt. This procedure will allow monitoring of the number
                    of questionnaires returned each day, as well as the cumulative total returned
                    (Babbie, 1973), to help plan follow-up mailings and reminders.
                    After the data have been collected, answers must be compiled and numeri-
                    cally analyzed.  As emphasized above, it is essential to work with experi-
                    enced statisticians during the design of the survey to ensure an adequate
                    representation of the survey population.  It is also important  to conduct
                    pretests  and to  ensure  that test conditions, including the  questions and
                    instructions, are adequate for the purposes of the survey.  Appropriate
                    correlations cannot be made if the data are weak or missing.

                    Additional problems may be encountered when attempting to look at special
                    subpopulations,  such as those who eat fish frequently.  The design  of
                    population surveys and sampling techniques for events and populations that
                    are nonuniform or infrequent presents  statistical issues that result in  an
                    additional series of trade-offs (Kalton and Anderson, 1986;  Sudman et al.,
                    1988).  Fisher (1988) examined the case of looking for preferential fish
                    consumers and noted that a larger total sample may be required.  The recent
                    ChemRisk (1991a) survey based the size of the population sample on the
                    most constraining piece of data among the questions to be answered and
32

-------
calculated the sample size required to ensure that the minimum number of
replies needed for statistically valid results would be received. In this case,
the inverse of the participation rates (percent anglers seeking to catch perch)
for perch harvest from warm water riverine fisheries was multiplied by the
desired number of consumption observations for perch (believed to represent
the rarest subpopulation to be encountered) and concluded that a minimum
of 1,363 completed surveys would be required. ChemRisk (1991a) then
estimated the percentage of undeliverable mailed surveys  (10 percent),
the percentage of those not answered due to changes in fishing status (10
percent), and the  potential response rate  of anglers who  received  the
surveys (75 percent) and calculated  a  minimum sample size of  2,244
using the following equation:

r.-       Tr
where:

Ts =   Total number of surveys sent,
Tr =   Total required for a statistically valid sample size (1,363),
Pdi =  Fraction of surveys deliverable as addressed (0.90),
Pd2 =  Fraction of 1989 licensed anglers who also purchased
       1990
       licenses or had changes in fishing status (0.90), and
R =   Expected rate of response to delivered surveys (0.75).

ChemRisk mailed surveys to 2,500 anglers, selected from a pool of 2,953
names drawn randomly from the fishing license files to represent different
resident categories and other special types.

Other statistical considerations include the accuracy of the responses that can
be expected depending on the approach used to collect the data. For example,
Carline (1972) found that harvest rates on the number offish caught per day
or per year were much higher than the catch rates determined by personal
interviews of anglers.  Swanson and Stephenson  (1982) observed that the
numbers of fish reported caught were often rounded off to  5, 10, 20, etc.,
indicating that biases and sources of error were greater for recall of angler
harvests than of angler effort.  The inaccuracy of respondents' memories is
troublesome for recall, in addition to the inability of respondents to accurately
identify fish species, confusion over the questionnaires, and frank exagger-
ation.

Additional checks may need to be made to examine non-response bias. West
et al. (1989a) found that those who did not return surveys ate less fish, thereby
resulting in a skewed calculation of consumption rates if the results were
assumed to  be accurate  for the entire subpopulation originally sent the
                                                                                      33

-------
                    questionnaires (see West et al., 1989c for calculation of nonresponse bias
                    adjustment factor). Weighting techniques, based on demographic character-
                    istics or other factors, may need to be applied to allow more  accurate
                    determinations of consumption rates for various subpopulations (e.g., Pao et
                    al., 1982).

                    The final statistical calculations should be  carefully  planned and based
                    solidly on the data collected in order to minimize assumptions that could
                    compromise the results. For example, were the questions correctly phrased
                    to elicit the number of fish meals per angler or per household?  If the latter,
                    were the members of the household enumerated, or must an average size
                    household be assumed to determine the individual fish consumption rates?
                    Can the data be used to calculate fish consumption by race or ethnic group,
                    income, education, sex, or other factors that the survey may wish to test? A
                    number of multivariate analyses may be used to compare  differences in
                    consumption rates for many  factors.  Detailed discussions on statistical
                    analyses that may be used with survey  data are presented in Babbie (1973).
                    Whenever possible, these should be investigated and the appropriate infor-
                    mation and numbers of responses should be planned prior to the collection
                    of data.
                     Recommendations:

                     It is important to consider using data management protocols that will
                     allow the data to  be readily accessible.  These include using standard
                     formats, such as dBASE™ or Lotus®, standard statistical packages, and
                     simple coding systems for ease in interpretation of the survey data.

                     A system for archiving the data (both paper and computer records) and
                     ensuring future availability should also be determined prior to comple-
                     tion of the study, and this information should be included in the final
                     report to aid other researchers on fish consumption.
34

-------
                                                             IV.  DISCUSSION
Wagstaff et al. (1986) noted that fish constitute the only class of foods subject to total governmental
prohibition in large geographic areas of the United States for substantial time periods because of
exposure to potentially hazardous environmental pollutants. Therefore, nutritionists, the medical
community, marketing specialists, fishery resource managers, and ecosystem administrators would
benefit from fish consumption databases that are "well-defined, validated, and accessible."

A variety of methods and approaches have been used in the fish consumption surveys presented in
this report, but it appears that a thoroughly satisfactory survey remains to be conducted. Although
the surveys may have  been satisfactory for the objectives of the designers at the time, the goal of
obtaining valid fish consumption rate data for high-risk subsistence anglers remains elusive. On-site
interviews are more likely to reach subsistence/recreational anglers, who may not be licensed, but
more detailed data may be obtained by diaries and written questionnaires. Many questions remain.
Over what period of time must a survey be conducted (for one day, seasonally, or for one year)?
What is the best approach for reaching the most anglers in a particular region?  How have fish
consumption bans and advisories affected the utilization of these resources and changed the public's
perception of risks (see Reinert et al., 1991)? Silverman (1990) reviewed recent national and Great
Lakes regional studies and noted the absence of detailed information about the public's consumption
of sport fish.  She also  found that fish consumption had been partitioned into commercial vs.
recreational portions that were not adequately covered in the surveys. Important information was
missing on the sport fish  consumption habits of the nonfishing public. Silverman (1990) described
the West et al. (1989b) survey as one of the best of its type because fish consumption was reported
based on demographic variables. However, other variables, such as fishing frequency, may prove
necessary to our understanding of fish consumption rates.

Clearly, additional efforts will be required to improve the survey methods and identify the best
approaches) to obtaining the desired data for fish consumption rates of subsistence and recreational
anglers and special subpopulations of concern.  For example, because most State fishery managers
conduct creel census  surveys routinely to assess resource use, additional public health-related
questions could be included in the creel census to provide information needed for health officials.
Other interagency and interdisciplinary cooperative ventures should be encouraged to reduce costs
and effort. It is hoped that the information in this document will assist fisheries managers and health
officials in designing and conducting surveys that will more accurately assess the fish consumption
rates by various subpopulations in different regions of the country.
                                                                                      35

-------
36

-------
                                                           BIBLIOGRAPHY
Ahmed, F.E. (ed.).  1991.  Seafood safety. Institute of Medicine, Committee on Evaluation of the
   Safety of Fishery Products. Washington, DC: National Academy Press. 432 pp.

Anderson, S.A. (ed.).  1986.  Guidelines for use of dietary intake data. Report prepared for the
   Center for Food Safety and Applied Nutrition, Food and Drug Administration, Washington,
   DC.  Life  Sciences Research Office, Federation  of American Societies for Experimental
   Biology, Bethesda, MD.   89 pp.  National Technical Information Service, PB87-210886,
   Springfield, VA.

Anderson, S.A. (ed.).  1988.  Estimation of exposure  to substances in the  food supply. Report
   prepared for the Center for Food Safety and Applied Nutrition, Food and  Drug Administration,
   Washington, DC  Life Sciences Research Office, Federation of American Societies for Exper-
   imental Biology,  Bethesda, MD.  National Technical Information Service, PB89-205199,
   Springfield, VA

Babbie, E.R. 1973. Survey research methods. Belmont, CA: Wadsworth Publishing Co. 384 pp.

Beak Consultants, Inc. 1989. Human consumption rates of fish from the Columbia River, Project
   #73296. Prepared for Northwest Pulp and Paper Association, Bellevue, WA, December 1989.
   27 pp. + appendix.

Belton,  T.,  B.  Ruppel, K. Lockwood,  S. Shiboski, G. Bukowski, R. Roundy, N. Weinstein, D.
   Wilson, and H. Whelan.  1985.  A study of toxic hazards to  urban recreational anglers and
   crabbers. Office of Science and Research, New Jersey Department of Environmental Protection,
   Trenton, NJ. 68 pp.

Belton,  T.,  R.  Roundy, and N. Weinstein. 1986. Urban fisherman: Managing the risk of toxic
   exposures.  Environment 28:19-20, 30-37.

Block, G.  1982.  A review of validations of dietary assessment methods.  Amer. J. Epidemiol.
   115(4):492-505.

Brown,  G.L. 1981. A survey of recreational shrimping in the bay and sound systems of the Gulf
   coast for 1980. Human Sciences Research, McLean, VA, Document No. HSR-RR-81/2-Pon,
   April 1981. Prepared for Gulf states Marine Fisheries Commission, Ocean Springs, MS. 140pp.

Brown,  H.S., R. Gable, and L. Telelbaum. 1988. Methodology for assessing hazards of contami-
   nants in seafood.  Reg. Toxicol. Pharmacol. 8:76-101.

Capuzzo, J.M., A. McElroy,  and G. Wallace. 1987.  Fish and shellfish contamination in New
   England waters: An evaluation and review of available data on the distribution of chemical
   contaminants. Coast Alliance, Washington, DC. 59 pp. + appendices.
                                                                                   37

-------
Carline, R.F. 1972. Biased harvested estimates from a postal survey of a sport fishery. Trans. Am.
   Fish. Soc. 2:262-267.

Chandler,  K.A.,  and G.L. Brown.  1978.  A pretest of an approach to collection of marine
   recreational fishing data on the East and Gulf coasts. Human Sciences Research, McLean, VA,
   Document  No. HSR-RR-78/1-C1, January  1978.  Prepared for  National Marine Fisheries
   Service, Data Management and Statistics Division, Contract No.6-35339, Washington, DC. 175
   pp.

Chase, D.R., and M. Harada.  1984.  Response error in self-reported recreation participation rates.
   J. Leisure Res. 15(4):322-329.

ChemRisk. 1991a. Consumption of freshwater fish by Maine anglers. ChemRisk, Portland, Maine,
   July 1991.

ChemRisk.  1991b.   Saco River creel survey.  Preliminary report ChemRisk, Portland, Maine,
   January 1991.

ChemRisk. 1991c. Penobscot River creel survey. Preliminary report. ChemRisk, Portland, Maine,
   January 1991.

Clark,  M.J., L. Fink, and D. DeVault,   1987. A new  approach  for the establishment of fish
   consumption advisories.  Great Lakes Reservoir 13(3):367-374.

Congdon,  J.C  1988. Angler use and harvest on Fox Lake, Wisconsin. Wisconsin Department of
   Natural Resources, Bureau of Fisheries Management, Madison, WI. Fish Management Report
    133. 20pp.

Connelly,  N.A., T.L. Brown, and B.A. Knuth. 1990. New York statewide angler survey 1988. New
   York State/Department of Environmental Conservation, Division of Fish and Wildlife, Albany,
   NY. 158pp.

Cooper, C.B.,  M.E. Doyle, and K. Kipp.  1991. Risks of consumption of contaminated seafood:
    the Quincy Bay case study.  Environ. Health Perspect. 90:133-140.

Cordle, F., P. Cornellussen, C. Jelinik, B. Hackley, R. Lehman, J. McLaughlin, R. Rhoden, and R.
    Shapiro. 1978. Human exposure to polychlorinated biphenyls and polybrominated biphenyls.
   Environ. Health Perspect. 24:157-172.

Cordle, F., R.  Locke, and J.  Springer.  1982. Risk assessment in a federal regulatory agency: An
    assessment of risk associated with the human consumption of some species contaminated with
    polychlorinated biphenyls (PCBs). Environ. Health Perspect 45:171-182.

Cox, C., and A.F. Johnson. 1990.  A study of the consumption patterns of Great Lakes salmon and
    trout anglers. Ontario Ministry of the Environment, Water Resources Branch, Toronto, Ontario.
    199 pp. (ISBN-0-7729-7428-4)

-------
Cox, C, A. Vaillancourt, and A.F. Johnson. 1987. A comparison of the results from the "Guide
   to Eating Ontario Sport Fish"  questionnaire. Ontario Ministry of the Environment, Water
   Resources Branch, Toronto, Ontario.  85 pp. (ISBN 0-7729-2359-0)

Cox, C. A. Vaillancourt, and A.F. Johnson. 1989. A method for determining the intake of various
   contaminants through the consumption of Ontario sport fish. Ontario Ministry of the Environ-
   ment, Water Resources Branch, Toronto, Ontario. 20 pp. (ISBN 0-7729-3498-3)

Cox, C, A. Vaillancourt, and A.F. Johnson. 1990. The results of the 1989 "Guide to Eating Ontario
   Sport Fish" questionnaire.  Ontario Ministry of the Environment, Water Resources Branch,
   Toronto, Ontario. 27 pp. (ISBN 0-7729-7084-X)

Crunkilton, R.L., L.M. Smith, J.D. Petty, and R.D.  Kleopfer.  1987.  Residues of  2,3,7,8-
   tetrachlorodibenzo-p-dioxin in the Spring River, Missouri. Water Air Soil Pollution (Nether-
   lands) 32 (1-2):219-231.

Cunningham, P.A., J.M. McCarthy, and D. Zeitlin. 1990.  Results of the 1989 census of state fish
   consumption advisory programs. U.S. Environmental Protection Agency, Office of Water
   Regulations and Standards, Washington, DC.

DeVault, D.S.  1985. Contaminants in fish from Great Lakes harbors and tributary mouths. Arch.
   Environ. Contam. Toxicol. 14:587-594.

Diana, S.C.  1989. Anglers' perceptions and practices related to contaminants in sport-caught fish
   from Lake Ontario. M. S. thesis, Cornell University, Ithaca, NY. 284 pp.

Dillman, D.A. 1978. Mail and telephone surveys: The total design method. New York, NY: John
   Wiley and Sons.

Dillman, D.A. 1983. Mail and other self-administered questionnaires. In: Handbook of Survey
   Research (P.H. Rossi,  J.D. Wright,  and A.B. Anderson, eds.), pp.  359-376.   New York:
   Academic Press.

Fiore, B.J., H.A. Anderson, L.P. Hanrahan. L.J. Olsen, and W.C.  Sonzogni.  1989. Sport fish
   consumption and body burden levels of chlorinated hydrocarbons:  A study of Wisconsin
   anglers. Arch. Environ. Health 44(2):82-88.

Fisher, K.D. 1988. Approaches to estimating fish consumption in the United States. Prepared for
   Center for Food Safety and Applied Nutrition, U.S. Food and Drug Administration, Order no.
   FDA-003459-01-87 by Life Sciences  Research Office, Federation of American Societies for
   Experimental Biology, Bethesda, MD. 58 pp.

Food Chemical News.  1988. NMFS survey of recreational fishing halted by lack of funding. Food
   Chem. News 30:35-37.
                                                                                    39

-------
Foran, J.A., M. Cox, and D. Croxton.  1989.  Sport fish consumption and projected cancer risks in
   the Great Lakes Basin. Am. J. Health 79:322-325.

Fowler,  D.L., and D.R. Holder.  1987.  A  fisheries survey of the St. Mary's River.  Georgia
   Department of Natural Resources, Game and Fish Division, Atlanta, GA. 81 pp.

Gahler, A.R., R.L. Arp, and J.M. Cummins. 1982. Chemical contaminants in edible non-salmonid
   fish and crabs from Commencement Bay, Washington. U.S. Environmental Protection Agency,
   Environmental Services Division, Seattle, WA. 117 pp. National Technical Information Ser-
   vice, PB83-172163, Springfield, VA.

Haines,  P.S., B.M. Popkin, and  D.K, Guelkey.  1987.  Determinants of fish consumption among
   U.S. women. Paper presented at the American Dietetic Association meeting October 19-24,
   1987.

Heatwole, C.A., and N.C. West.  1983.  Urban shore-based fishing: A health hazard?  Vol.  Ill,
   Coastal Zone '83, Proceedings of the Third Symposium on Coastal and Ocean Management.
   Publ. by American Society of Civil Engineers.

Heizer,  R.E.  1986.  Summer creel survey  of Cosgrove and Elwood Lakes, Florence County,
   Wisconsin, 1979. Wisconsin Department of Natural Resources, Madison, WI. Fish Manage-
   ment Report 130. 23 pp.

Heizer, R.E. 1988.  A creel survey on Bass and Boot Lakes, Oconto County, Wisconsin, December
   1981-September 1982.  Wisconsin Department  of Natural Resources, Bureau  of Fisheries
   Management, Madison, WI. Fish Management Report 137. 24 pp.

Honstead, J.F., T.M. Beetle, and J.K. Soldat.  1971.  A statistical study of the habits of local
   fisherman and its application to evaluation of environmental dose.  Prepared for U.S. Environ-
   mental Protection Agency by Battelle Pacific Northwest Laboratories, Richland, WA.

Hottell,  H.E., D. R. Holder and C. E. Coomer, Jr. 1983. A fisheries survey of the Altamaha River.
   Georgia Department of Natural Resources, Game and Fish Division, Atlanta, GA. 68 pp.

Hu, T-W.  1985. Analysis of seafood consumption in the U.S. 1970, 1974,  1978,1981.  National
   Marine Fisheries Service Grant No. NA82AA-H-00053. Institute for Policy Research and
   Evaluation, The Pennsylvania State University, University Park, PA.

Humphrey, H. 1976. Evaluation of changes of the level of polychlorinated biphenyls (PCBs) in
   human  tissue. Final report  on FDA contract 223-73-2209.  Michigan Department of Public
   Health, Lansing.

Humphrey, H. 1983. Evaluation of humans  exposed to waterborne chemicals in the Great Lakes.
   Final report to the EPA.  Cooperative agreement CR-807192.  Michigan Department of Public
   Health, Lansing.
40

-------
Kalton, G., and D.W. Anderson. 1986. Sampling rare populations. J. R. Stat. Soc. (A) 149:65-82.

Kleiman, CF.  1985.  Fish consumption by recreational fisherman: An example of the Lake
   Ontario/Niagara River region. Prepared for Office of Enforcement and Compliance Monitoring,
   U.S. Environmental Protection Agency, Washington, DC, by Environ Corporation, May 1985.
   16pp.

Kleinschmidt Associates. 1989. Report on the fish consumption survey, Androscoggin River, New
   Hampshire.  Prepared for James River Corporation, Berlin, NH. November 1989.  17 pp. +
   appendices.

Landolt, M.L., FJL Hafer, A. Nevissi, G. van Belle, K. Van Ness, and C. Rockwell.  1985. Potential
   toxicant exposure among consumers of recreationally caught fish from urban embayments of
   Puget Sound.  NOAA Technical Memorandum NOS OMA 23. National Oceanic  and Atmo-
   spheric Administration, Rockville, MD. 104 pp.

Landolt, M., D. Kalman, A. Nevissi, G. van Belle, K. Van Ness, and F. Hafer. 1987. Potential
   toxicant exposure among consumers of recreationally caught fish from urban embayments of
   Puget Sound: final report. NOAA Technical Memorandum NOS OMA 33.  National Oceanic
   and Atmospheric Administration, Rockville, MD. 104 pp.  National Technical Information
   Service, PB91-172833, Springfield, VA.

Lepkowski, J., and R. Groves. 1986. A two-phase probability proportional to size telephone sample
   design. In: Proceedings of the survey methods research section, American Statistical Associa-
   tion, pp. 357-362.

McCallum, M.  1985. Recreational and subsistence catch and consumption of seafood from three
   urban industrial bays of Puget Sound: Port Gardner, Elliott Bay, and Sinclair Inlet. Washington
   Department of Social and Health Services, Division of Health, Olympia, WA. 59 pp.

Mullis, A* W. 1989. Age composition and sport harvest of striped bass from Roanoke River. North
   Carolina Wildlife Resources Commission, Raleigh, North Carolina, October 1989.15 pp.

Nash, D.A. 1971. A survey offish purchasers by socioeconomic characteristics. Data  Report No.
   62. National Marine Fisheries Service, Seattle, WA. National Technical Information Service,
   COM 71-00647, Springfield, VA.

National Marine Fisheries  Service.   1976.  Seafood consumption study 1973-1974.  National
   Marine Fisheries Service, Washington, DC 146 pp. (Tape of raw data also  available.)

National Marine Fisheries Service. 1986a. Marine recreational fishery statistics survey, Atlantic
   and Gulf coasts, 1986. Current Fishery Statistics no. 8392, National Marine Fisheries Service,
   Washington, DC
                                                                                   41

-------
National Marine Fisheries Service.  1986b.  Marine recreational fishery statistics survey, Pacific
   coast, 1985. Current Fishery Statistics no. 8328, National Marine Fisheries Service, Washing-
   ton, DC

National Marine Fisheries Service.  1989.  Fisheries of the United States -1988. Current Fisheries
   Statistics no. 8800. National Marine Fisheries Service, Washington, DC.

National Marine Fisheries Service.  1991. Marine recreational fishery statistics survey, Atlantic
   and Gulf Coasts, 1987-1989. Current Fisheries Statistics no. 8904, National Marine Fisheries
   Service, Washington, DC

NOAA, FDA and EPA. 1986 and 1987. Report on 1984-1986 Federal Survey of PCBs in Atlantic
   Coast bluefish (Data Report 1986; Interpretive Report, 1987). National  Oceanic and Atmo-
   spheric Administration, U.S. Food and Drug Administration, and U.S. Environmental Protection
   Agency. National Technical Information Service, PB86-218070, Springfield, VA.

Nuclear Regulatory Commission. 1983. Georgia fishery study: Implications for dose calculations.
   Memo from M.D. Turcotte to H. Olson.  Savannah River Laboratory,  Technical Division,
   August 5, 1983.

Olson, L. J.,  1988. Fish consumption rate literature survey compilation. Wisconsin Department of
   Health and Social Services, Madison, WI.

Pao, E.M., K.H. Fleming, P.M. Guenther, and S.J. Mickle. 1982. Foods  commonly eaten by
   individuals: Amount per day and per eating occasion. Home Economics Research Report No.
   44. U.S. Department of Agriculture, Hyattsville, MD.

Pierce, R.S., D.T. Noviello, and S.H. Rogers.  1981. Commencement Bay seafood consumption
   report. Preliminary report. Tacoma-Pierce County Health Department, Tacoma, WA.

Puffer, H.W., S.P. Azen, M.J. Duda, and D.R. Young. 1982a. Consumption rates of potentially
   hazardous  marine fish  caught in the metropolitan Los Angeles area.  U.S.  Environmental
   Protection  Agency, Environmental Research Laboratory, Corvallis, OR.  EPA 600/3-82-070.
   National Technical Information Service, PB82-229493, Springfield, VA.

Puffer, H.W., MJ. Duda, and S.P. Azen. 1982b. Potential health hazards offish caught in polluted
   coastal waters of Los Angeles County. N. Am. J. Fish. Management 2:74-79.

Ranthum, R.G. 1975. Five-year summary of Mississippi River special tailwater creel censuses in
   Pool 7, 1969-1973.  Bureau of Fish and Wildlife Management, Wisconsin  Department of
   Natural Resources, Madison, WI.

Reinert, R.E., B.A. Knuth, M.A. Kamrin, and Q.J. Stober. 1991. Risk assessment, risk management
    and fish consumption advisories in the United States. Fisheries  16(6):5-12.
 42

-------
Rupp, E.M.  1990.  Age dependent values of dietary intake for assessing human exposures to
   environmental pollutants.  Health Physics 39:151-163.

Rupp, E.M., F.L. Miller, and LCF. Baes. 1980. Some results of recent surveys offish and shellfish
   consumption by age and region of U.S. residents. Health Physics 39:165-175.

San Diego County Department of Health Services.  1990.  San Diego Bay  health risk study.
   Environmental Health Services, Document No. 25467, San Diego, CA. 322 pp.

Schuman, L. M., et. al.  1982. Evaluation of the methods used to determine potential health risks
   associated with organic contaminants in  the Great Lakes basin.  School  of Public Health,
   University of Minnesota, Minneapolis, MN. EPA-600/3-84-002. National Technical Informa-
   tion Service, PB84-128305, Springfield, VA.

Schmitt,  D.N., and J. H. Hornsby.  1985.  A fisheries survey of the Savannah River. Georgia
   Department of Natural Resources, Game and Fish Division, Atlanta, GA. September 1985.

Schumacher, E.R. 1987.  Creel survey on Pewaukee and Nagawicka Lakes, Waukesha County,
   Summer 1982.  Wisconsin Department of Natural Resources, Bureau of Fish Management,
   Madison, WI. Fish Management Report 131.

Schwartz, P.M., S.W. Jacobson, G.G. Fein, J.L. Jacobson, and H.A, Price.  1983. Lake Michigan
   fish consumption as a source of polychlorinated biphenyIs in human cord serum, maternal serum
   and milk.  Am. J. Public Health 73:293-96.

Scott, T.M., Jr. 1981.  Creel survey on the  Flint River between Lakes Blackshear and Worth.
   Georgia Department of Natural Resources, Game and Fish Division, Atlanta, GA. 23 pp.

Silverman, W.  1990.  Fish consumption in the  Great Lakes basin:  A review of the literature.
   Prepared for U.S. Environmental Protection Agency, Region 5, Water Division, with support
   of the National Network for Environmental Management Studies.

Skea, J.C., H.A. Simonin, EJ. Harris, S. Jackling, JJ. Spagnoli, J. Synula, and J.R.  Colquhoun.
   1979. Reducing levels of Mirex, Aroclor 1254, and DDE by trimming and cooking Lake Ontario
   brown trout and smallmouth bass. J. Great Lakes Res. 5:153-159.

Smith, B.F., and E.E. Enger.  1988. A survey of attitudes and fish consumption of anglers on the
   Lower Tittabawassee River, Michigan. Michigan  Department of Public Health, Center for
   Environmental Health Sciences, Document No. 172, Lansing, MI.

Smith, B.F., and W.N. Thompson. 1989. Environmental sociology: Anglers of the Tittabawassee.
   Environment 26(5):5,43.

Southerland, E. 1991.  Federal assistance plan for state fish consumption advisories.  Presented at
   the Institute of Medicine Symposium on Seafood  Safety Issues, May  14, 1991.  National
   Academy of Sciences, Washington, DC.
                                                                                    43

-------
Spencer, M.D.  1987.  A survey of the fishery resources of the Conasauga and Coosawattee Rivers.
   Georgia Department of Natural Resources, Game and Fish Division, Atlanta, GA.  48 pp.

Springer, C 1990.  Risk perceptions and communications needs in Lake Ontario's chemically
   contaminated sport fishery. M.S. thesis, Cornell University, Ithaca, NY.

SRI International. 1980.  Seafood consumption data analysis. Final report Prepared for U.S.
   Environmental Protection Agency under contract 68-01-3887.

Sudman, S., and N. Bradburn.  1982. Asking questions: A practical guide to questionnaire design.
   San Francisco, CA: Jossey-Bass.

Sudman, S., M.G. Sirken, and C.D. Cowan. 1988. Sampling rare and elusive populations. Science
   240:991-996.

Swanson, E.D., and  R.L.  Stephenson.  1982.  Fishing effort and harvest by Arizona's licensed
   resident anglers.  Arizona Game and Fish Department, Planning Branch, May 1982.

Thompson, T., and W.A.  Hubert.  1990. Influence of survey method on estimates of statewide
   fishing activity. North American Journal of Fisheries Management 10:111-113.

Thuemler, T.F. 1981. A creel census on Lake Noquebay, Marinette County, Wisconsin, 1977.
   Wisconsin Department of Natural Resources, Bureau of Fish Management, Madison, WI. Fish
   Management Report 108. 20 pp.

Tollefson, L., and F. Cordle.  1986. Methylmercury in fish:  A review of residue levels, fish
   consumption and regulatory action in the United States. Environ. Health Perspect. 68:203-208.

USDA. 1978. Foods commonly eaten by individuals. U.S. Department of Agriculture, Hyattsville,
   MD.

USDA. 1983a. Food consumption: Households in the United States, seasons and year 1977-78.
   Nationwide Food Consumption Survey 1977-78. Human Nutrition Information Service, Con-
   sumer Nutrition Division.  Report no. H-6.   National Technical Information Service, PB91-
   105379, Springfield, VA.

USDA. 1983b. Food intakes: Individuals in 48 states, year 1977-78. Nationwide Food Consump-
   tion Survey 1977-78.  Human Nutrition Information Service, Consumer Nutrition Division.
   Report no.  1-1.

USDA.  1985a. Food and nutrient intakes:  Individuals in four regions, 1977-78. Nutrition
   Monitoring Division, Human Nutrition Information Service. Report no. 1-3.

USDA. 1985b. National food consumption survey: Continuing survey of food intakes by individ-
   uals, low-income women 19-50 years, 1 day 1985. NFCS, CSFn Report no. 85-2.  National
   Technical Information Service, PB88-110127, Springfield, VA.
44

-------
USDA.  1985c. Nationwide food consumption survey:  Continuing survey of food intakes by
   individuals, men 19-50 years, 1 day 1985. MFCS, CSFII Report no. 85-3.

USDA.  1985d. Nationwide food consumption survey:  Continuing survey of food intakes by
   individuals, low-income women 19-50 years and their children 1-5 years, 4 days 1985. NFCS,
   CSFII Report no. 85-4. National Technical Information Service, PB88-246202, Springfield,
   VA.

USDA  1986a. Nationwide food consumption survey: Continuing survey of food intakes by
   individuals, women 19-50 years and their children 1-5 years, 1 day 1986. NCFS, CSFII Report
   no. 86-1. U.S. Department of Agriculture, Hyattsville, MD.

USDA.  1986b. Nationwide food consumption survey:  Continuing survey of food intakes by
   individuals, low-income women 19-50 years and their children 1-5 years, 1 day 1986. NFCS,
   CSFII Report No. 86-2. National Technical Information Service, PB88-110119, Springfield,
   VA.

USDA.  1986c.  Nationwide food consumption survey: Continuing survey of food intakes by
   individuals, women 19-50 years and their children 1-5 years, 4 days 1986. NCFS, CSFII Report
   no. 86-3. U.S. Department of Agriculture, Hyattsville, MD.

USDA.  1986d. Nationwide food consumption survey:  Continuing survey of food intakes by
   individuals, low-income women 19-50 years and their children 1-5 years, 4 days 1986. NFCS,
   CSFH Report no. 86-4. National Technical Information Service, PB89-179485, Springfield,
   VA.

USDA,  1987-88. Nationwide food consumption survey, Household. National Technical Informa-
   tion  Service, PB92-50001,  Springfield,  VA.  (Magnetic tape of data.  Documentation also
   available separately.)

USDHHS.  1989.  The relationship of human levels of lead and cadmium to the consumption of
   fish caught in and around Lake Coeur D'Alene, Idaho. Final report. U.S. Department of Health
   and Human Services, Agency for Toxic Substance and Disease Registry, Atlanta, GA.

USEPA. 1983. Survey management handbook. Volume I:  Guidelines for planning and managing
   a statistical survey. U. S. Environmental Protection Agency, Office of Policy, Planning, and
   Evaluation, Washington, DC. EPA-230/12-84-002, November 1983.  105 pp. + appendices.
   National Technical Information Service,  PB85-187672, Springfield, VA.

USEPA. 1984a. Survey management handbook. Volume II: Overseeing the technical progress of
   a survey contract. U. S. Environmental Protection Agency, Office of Policy, Planning, and
   Evaluation, Washington, DC EPA-230/12-84-002, December 1984.  168 pp.  PB85-187680.
                                                                                  45

-------
USEPA. 1984b. Evaluation of the methods used to determine potential health risks associated with
   organic contaminants in the Great Lakes Basin. U. S. Environmental Protection Agency, Office
   of Research and Development, Duluth, MN.  EPA-600/3-84-002, January 1984.  458 pp.
   National Technical Information Service, PB84-128305, Springfield, VA.

USEPA 1989. Assessing human health risks from chemically contaminated fish and shellfish: a
   guidance manual.  U.S. Environmental Protection Agency, Office of Marine and Estuarine
   Protection/Office  of Water  Regulations and Standards. Publication EPA-503/8-89-002.
   National Technical Information Service, PB91-168369. Springfield, VA.

USFDA. 1990. Risk assessment methodology for environmental contaminants in fish and shellfish.
   U.S. Food and Drug Administration. October 1990.

Volland, M.,  K. Gall, D. Lisk, and D. MacNeil.  1990.  The effectiveness of recommended
   fat-trimming procedures on the reduction of PCB and Mirex levels in Lake Ontario brown trout
   (Salmo trutta): A final report. New York Sea Grant Extension Program, Cornell University,
   Ithaca, NY.

Wagstaff, D.J., M. Meaburn, M. Boger, S. Conrath, and B. Hackley.  1986.  Status of data sources
   on fish consumption in the United States. Mar. Fish. Rev. 48:20-23.

Walker, R.J., D.B. Andersen, and L. Brown. 1988. Community profile databases catalog, Volume
   2, Southwest and Western Alaska Regions.  Alaska Department of Fish and Game, Anchorage,
   AK.

Washington Department of Social and Health Services.  1985. Report: Seafood catch and consump-
   tion in urban bays of  Puget Sound. Washington Department of Social and Health Services.
   January 1985.

Wendt, M.E.  1986. Low income families' consumption of freshwater fish caught from New York
   State waters. Master's thesis, Cornell University, Ithaca, NY.

West, P.C, M.J. Fly,  F. Larkin,  and R. Marans.  1989a. Minority anglers and toxic fish consump-
   tion: Evidence from a state-wide survey of Michigan. School of Natural Resources, University
   of Michigan, Ann Arbor.

West, P.C., M.J. Fly, R. Marans, and F. Larkin. 1989b. Michigan sports anglers fish consumption
   survey. A report to the Michigan Toxic Substance Control Commission. Michigan Department
   of Management and Budget Contract No. 87-20141. Natural Resource Sociology Research Lab,
   Technical Report  No. 1.

West, P.C.  M.J. Fly, R. Marans, and F. Larkin. 1989c. Michigan sports anglers fish consumption
   survey, Supplement I, Non-response  bias  and consumption suppression effect adjustments.
   School of Natural Resources, University of Michigan, Ann Arbor. Natural Resource Sociology
   Research Lab, Technical Report No. 2.
46

-------
West, P.C, MJ. Fly, R. Marans, and F. Larkin.  1989d.  Michigan sport anglers fish consumption
   survey, Supplement n, Test for stability of consumption rates over time.  School of Natural
   Resources, University of Michigan, Ann Arbor. Natural Resource Sociology Research Lab,
   Technical Report No. 3.

West, P.C, M.J. Fly, R. Marans, and F. Larkin.  1989e. Toxic fish consumption and the elderly.
   School of Natural Resources, University of Michigan, Ann Arbor.

Westat Inc.  1989.  Investigation of possible recall/reference period bias in national surveys of
   fishing, hunting and wildlife-associated recreation. U.S. Fish and Wildlife Service, Washing-
   ton, DC.

Willett, W.C., L. Sampson, MJ. Stampfer, B. Rosner, C. Bain, J. Witschi, C.H. Hennekens, and
   F.E. Speizer.  1985. Reproducibility and validity of a semiquantitative food frequency ques-
   tionnaire. Am. J. Epidemiol.  122:51-65.

Wisconsin Department of Health and Social Services.   1987.  Study of sport fishing and fish
   consumption habits and body burden levels of PCBs, DDE, and mercury of Wisconsin anglers.
   Final report to participants.  September 1987.

Wolfe, R.J., and R.J. Walker.  1987. Subsistence economies in Alaska: productivity, geography,
   and developmental impacts. Arctic Anthropology 24(2):56-81.
                                                                                     47

-------
48

-------
              APPENDIX
SUMMARY OF SURVEY METHODS INFORMATION
                                              49

-------
50

-------
SUMMARY OF SURVEY METHODS INFORMATION


Title of Survey
A Pretest of an Approach to
Collection of Marine
Recreational Fishing Data
on the East and Guff
Coasts



















1977
Fishing Effort and Harvest
by Arizona's Licensed
Resident Anglers


1980
I Contact
Address
Phone No.
Creel census
and telephone
survey






















Mail survey and
creel census




The following
information was
given in K.A.
Chandler and G.L
Brown, HSR-PR-
78/1 -C1, 25
January 1978,
prepared for NMFS

















Eric Swanson
Arizona Game and
Fish Department
Phoenix, AZ
(602) 942-3000
6x1608
||

Time J| Cost
6,077 telephone
surveys, 1 ,644
fishermen
interviewed at 3
locations to
estimate sample
sizes required
and number of
days
















Sent out 18,000
surveys (10% of
registered
fishermen);
33% response

For a total of
18,800 fish to
provide
estimates of the
proportional
distribution of
fish caught for
an area (not to
determine fish
consumption
rates), estimated
132 days to
interview 3,003
fishermen in
Rhode Island,
120 days for
3,087 interviews
in South
Carolina, 282
days to interview
6,373 in Texas




About 9 months
including set-up,
data gathering
and analysis


Telephone
interviews:
RI$1.50;SC
$1.73;TX
$1 .68; cost for
intercept
interviews not
given but
average
number of
interviews per
hr: Rl 2.59;
SC 2.29; TX
2.26; assumed
10hrsof
interviewing
per day

Cost for
surveys in
these 3 States
estimated to
be $333,236
(1979)

Funded
through
Federal aid





Comments
Noted cost per interview for
surf fishermen may be higher























Ballpark estimates


Have done subsequent
surveys


-------
SUMMARY OF SURVEY METHODS INFORMATION (continued)
II
Type of
Title of Survey Survey
Commencement Bay
Seafood Consumption
Study
1981
Fisheries Surveys:
Altamaha River
St. Mary's River
1982
1986
A Study of Toxic Hazards
to Urban Recreational
Fishermen and Crabbers
1983
Evaluation of Methods
Used to Determine
Potential Health Risks
Associated with Organic
Contaminants in the Great
Lakes Basin
1983
Creel census
Creel census
Personal
interview and
creel census
Telephone and
mail surveys
Contact
AddreM
Phone No.
Doug Pierce
Tacoma- Pierce
County Health
Department,
Tacoma, WA
(206) 591-5543
Dan Holder
Georgia Dept. Nat.
Resources, Game
& Fish Div.
Atlanta, GA
(912) 285-6094
Bruce Ruppel
NJ Dept. Environ.
Protection,
Trenton, NJ
(609) 984-6548
Given in report:
USEPA
Environmental
Research
Laboratory, DiHuth,
MN
Level of Effort | Time
5 months in the
field collecting
data;
7 months writing
report.
10-month creel
survey using
college students,
random samples
87 interviews on-
site
Collected data by
3 different
protocols, 587
respondents
1 year
10-month or 12-
month creel
survey
2 years for entire
study
About 2 years
Co«
$25,000
primarily to
pay contract
staff
$9,077 (based
on$5.50/hr
wage for
surveys)
Estimate:
$50,000,
funded by the
State
About $21 per
participant for
each protocol,
excluding data
analysis
i
Comments
Cost does not include tissue
analysis done by EPA
Ballpark estimates
Have done subsequent
surveys
Also funds from Hudson
River Foundation

                       52

-------
SUMMARY OF SURVEY METHODS INFORMATION (continued)
1 Type of
Survey
Recreational and
Subsistence Catch and
Consumption of Seafood
from Three Urban Industrial
Bays of Puget Sound
1983-1984
Low Income Families'
Consumption of Freshwater
Fish Caught from New York
State Waters.
1985
Potential Toxicant Exposure
Among Consumers of
RecreationaJly Caught Fish
from Urban Embayments of
Puget Sound
1983-1987
Study of Sport Fishing and
Fish Consumption Habits
and Body Burden Levels of
PCBs, DDE, and Mercury
of Wisconsin Anglers
1985
Personal
interview and
creel census
Personal
interview
Personal
interview and
creel census
Mail survey
Contact
Addrees
Phone No.
Mary McCallum
Washington State
Division of Health,
Epidemiology
Section, Seattle,
WA
(206)753-5964
Marie Wendt
KVRHA
122 State Street
Augusta, ME
04330
Dr. Marsha
Landott, School of
Fisheries,
University of
Washington
Seattle, WA
(206) 543-4270
Beth Fiore
Wisconsin Division
of Health
Madison, Wl
(608) 266-6914
Level of Effort
1643 interviews
oo-site
40 personal
interviews over a
2-week time
frame
1st year -4, 181
angler interviews;
2nd year - 437
interviews on-site
at boat ramps
1 600 surveys
mailed
801 returned
Time
Data collection
over a 12-month
period,
2 years total
Data collection
and analysis -
1 year
2 years
About 1 year
Co*t
Grant-
Si 00,000 for
salary of
supervisor
Graduate
student
thesis funded
through Sea
Grant
$207,000
(excluding
indirect costs)
Estimate of
$27,250
Comment*


Significant portion of funds
were for analytical chemistry;
rest for data entry and
analysis, salaries of
interviewers, etc.
Phone follow-up to mail out
50% responded
Cost does not include blood
analyses for contaminants
Would use two-tiered
approach next time
1) Great Lakes
2) general
                       53

-------
SUMMARY OF SURVEY METHODS INFORMATION (continued)


Title of Survey
Marine Recreational
Fishery Statistics Survey.
Atlantic and Gulf Coasts


1986
1987-1989
Relationship of Human
Levels of Lead and
Cadmium to the
Consumption of Fish
Caught on and around
Lake Coeur d' Alene, Idaho




1986-1987
A Survey of Attitudes and
Fish Consumption of
Anglers on the Lower
Tittabawassee River,
Michigan


1987
Angler Use and Harvest on
Fox Lake, Wl



1987

Type of
Survey
Creel census






Personal
interview or
telephone
survey







Creel census







Creel census





Contact
AddreM
Phone No.
Mark Holliday
National Marine
Fisheries Service,
NOAA,
Washington, DC

(301) 427-2328
Mike Qreenwell
Agency for Toxic
Substances and
Disease Registry
Public Health
Service
US Dept of Hearth
& Human Service
Atlanta, GA
(404) 639-0700

John Hesse
Michigan
Department of
Public Health,
Lansing, Ml
(517) 335-8353
(8350)

James C. Congdon
DNR Madison
Wisconsin Bureau
of Fisheries Mgmt
Horicon County
(414) 485-3003


Level of Effort
46,000 intercept
interviews and
74,000 telephone
interviews (1 986)



299 households,
follow-up study
on 33 individuals








5 interviewers
conducted 703
interviews





1/2 FTE doing
creel survey for
entire fishing
season (1 May-
1 5 March,
11 1/2 mos


Time
Data collection
1 year - data
ready fa
distribution within
4 months


About 2 years










4 months for
surveys (1 May
to 31 Aug)





1 1 1/2 months







Cost
Collaboration
with 5 State
agencies -
$2,000,000



Done
in-house









$6,500







Funded with
state funds






Comment*
The 1987 - 1989 survey Is
now available

Have done similar surveys for
the Pacific coast


Done by Division of Health
Studies, Sharon Campofulu









FoHow-up telephone survey
done by Michigan State
University as part of a survey
class










                     54

-------
SUMMARY OF SURVEY METHODS INFORMATION (continued)
Title of Survey
Michigan Sports Anglers
Fish Consumption Survey
1988
New York Statewide Angler
Survey
1988
A Study of the
Consumption Patterns of
Great Lakes Salmon and
Trout Anglers
1989
Consumption of Freshwater
Fish by Marine Anglers
1990
Type of
Survey
Mail survey
and telephone
survey
Mail survey
Mail survey
Mail survey
Contact
Address
Phone No.
Dr. Patrick West
School of Natural
Resources
University of
Michigan
(313)764-7206
(313) 763-2200
Dr. Nancy Connelly
Cornell University
NY State College
of Agriculture and
Life Sciences,
Fernow Hall,
Ithaca, NY
(607)255-2830
Chuck Cox
Water Resources
Branch, Ministry of
the Environment,
Toronto, CANADA
(416) 323-4994
Ellen Elbert
ChemRisk
1685 Congress St.
Portland, ME
(207) 744-0012
Level of Effort
2,600 surveys
mailed out 4
waves of mailings
and follow-up
phone calls for
non-response
bias
17,000 mailed
out
3 follow-up
mailings
200 telephone
follow-ups for
non-response
bias
10,31 4 quest.
returned
21 00 surveys
mailed out, 1 427
returned (68%
response)
2,500 mailed out
1,612 returned
Time
1 year
10 months,
total time
about 18 months
4 months for
data collection
and analysis
9 months
Cost
$30,000
Funded by
Dept. Environ.
Conserv.,
Bureau of
Fisheries,
State of New
York
$1,500 mailing
costs, plus
staff time for
processing
results
Client
confidential
Comments


Very effective with proper
cover letter, stamped return
envelope, and multiple choice
questionnaire. Also provide
space for comments, so
anglers may voice concerns.
Revised draft report available
                   55

-------