United States
                   Environmental Protection
                   Agency	
Environmental Monitoring
Systems Laboratory
Las Vegas NV 89193-3478
                   Research and Development
EPA/600/S4-89/036 June
&EPA         Project  Summary
                   Direct/Delayed  Response
                   Project:  Intel-laboratory
                   Differences  in  the  Northeastern
                   Soil Survey  Data

                   D. D. Schmoyer, R. S. Turner, and D. A. Wolf
                    Soil samples collected during the
                  Direct/Delayed Response Project
                  (DDRP) northeastern soil survey were
                  analyzed  using  the procedures
                  described in  the analytical methods
                  manual for the DORP by four different
                  laboratories. Inconsistencies  in
                  results were reported by the different
                  laboratories  regardless of  use  of
                  standardized procedures.  Audit
                  samples  Oa, A,  Bs,  Bw,  and C
                  Horizons were  included in this
                  survey. The audit  sample  data were
                  used for  an  interlaboratory
                  comparison to determine if there are
                  significant differences in  the  audit
                  sample  values reported by the
                  laboratories This analysis  compares
                  the laboratories to each other rather
                  than to  a true audit value.  Such
                  analysis has  been completed for
                  each chemistry variable  in the DORP
                  northeast soil chemistry  database.
                  Various graphics display the results.
                  Numerous significant  interlaboratory
                  differences  were  identified. The
                  impact of significant  interlaboratory
                  value differences on data analyses is
                  difficult to assess. If the samples
                  from each sampling  class, region,
                  and state were uniformly distributed
                  across all  the laboratories, the
                  interlaboratory value  differences
                  would likely have little effect on
                  analyses conclusions. However, this
                  is not the case. The significant value
                  differences among laboratories are
                  not consistent  across  all audit
                  horizons, which makes it even more
                  difficult to assess their effect on data
analyses.  The averaging that takes
place during data  aggregation will
likely  negate  the  effect  of
interlaboratory value differences.
However, it is not possible to test this
negation on a general scale.
  This  Project   Summary  was
developed by EPA's Environmental
Monitoring Systems Laboratory, Lag
Vegas, NV, to announce key findings
of the research project that is fully
documented in a separate report of
the same  title (see Project Report
ordering information at back).

Introduction
  Soil samples collected during the
Direct/Delayed Response  Project (DDRP)
northeastern soil survey  were analyzed
by four different  laboratories.  The
laboratories were required to use the
procedures  described in the analytical
methods manual for the DDRP soil
survey. Despite the  use of standardized
procedures, it  is possible for laboratory
biases to cause consistent differences in
results  reported  by the different
laboratories.
  Audit samples from Oa, A, Bs, Bw and
C horizons were included in the DDRP
northeastern  soil survey. The  audit
sample data were  used for an
interlaboratory  comparison to determine
if there are significant differences in the
audit sample  values reported by the
laboratories. For each variable, the audit
sample results are presented graphically,
the significant  interlaboratory differences
are displayed in tabular form,  and
summary statistics of the DDRP routine

-------
sample  results  are  provided  to aid the
reader  in  determining  the  practical
importance of the observed differences.
Procedure
  The method of analysis of variance was
used to determine significant differences
among the audit sample values reported
by the laboratories. The Scheffe multiple-
comparison procedure was  used  to
identify  the   specific  laboratory
differences  that  were  significant.  All
significance tests were performed at the
0.05 significance level.
  The  statistically  significant differences
among  values  reported   by the
laboratories may or may not  be  of
practical  importance to the data  user.
These  significant differences  were
determined  relative to  the  within-
laboratory variation observed in the audit
samples.  If the variation  in the  audit
samples  was smaller  than  the  variation
observed  in   routine  samples, the
interlaboratory differences may  have little
impact on the conclusions  derived from
    the soils data. Therefore,  the  observed
    interlaboratory  differences  were  also
    compared with the within-sampling-class
    standard deviation  estimates.  When the
    interlaboratory differences were  large,
    relative to this routine data  variation, they
    could affect inferences based on the
    routine data and should be considered
    when interpreting  the results  of  data
    analyses.
    Results and Discussion
      Numerous statistically  significant
    interlaboratory value  differences were
    found,  however,  only  a few  of  these
    differences  are large with respect to the
    variation observed in  the  routine data.
    Interiaboratory  value  differences  that
    could cause problems  in the analyses of
    routine sample data occurred for iron in
    0.002  M CaCI2,  aluminum in citrate-
    dithionite, aluminum in 0.002  M CaCI2,
    aluminum  in  KCI,  calcium  in  0.002 M
    CaCI2,  cation-exchange  capacity  in
    NH4CI, and  very fine sand.
Conclusions and
Recommendations
  The  impact of the  interlabora
differences on  data analyses is diffici
assess. The soil survey design used
concept of sampling classes which gi
soil series that, in theory,  should be
similar.  If  the  samples  from  e
sampling  class,  region,  and  state  v
uniformly  distributed across all
laboratories, the interlaboratory  v;
differences would likely have  little el
on analyses  conclusions. However, th
not the case.
  Many of the proposed levels of <
evaluation require data aggregated to
sampling  class level.  This aggrega
involves some type of averaging wi
sampling class and horizon combinatii
This averaging  will  likely negate
effect  of the interlaboratory  VE
differences;  however, this is very diffi
to test  on a  general   scale.
researchers  using the DDRP data she
be advised  to include the appropr
caveats concerning interlaboratory v;
differences  in  the conclusions of  tl
studies.
 D. D. Schmoyer and D. A. Wolf are with Martin Marietta Energy Systems, Inc., Oak
     Ridge, TN 37831-6285 and R. S. Turner is with Environmental Sciences
     Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6285.
 L J. Blume is the EPA Project Officer (see below).
 The complete report, entitled "Direct/Delayed Response Project: Interlaboratory
     Differences in the Northeastern Soil Survey Data," (Order No. PB90-219
     4031 AS; Cost: $17.00, subject to change) will be available only from:
        National Technical Information Service
        5285 Port Royal Road
        Springfield, VA22161
        Telephone: 703-487-4650
 The EPA Project Officer can be contacted at:
        Environmental Monitoring Systems Laboratory
        U.S. Environmental Protection Agency
        Las Vegas, NV 89193-3478
      United States
      Environmental Protection
      Agency
Center for Environmental Research
Information
Cincinnati OH 45268
      Official Business
      Penalty for Private Use $300

      EPA/600/S4-89/036

-------