United States Toxic Substances EPA-560/13-80-21
Environmental Protection Washington DC 20460 August, 1980
Agency
Toxic Substances
&EPA Asbestos Analytical Programs
Bulk Sample Analysis:
New York City and Maryland
-------
EPA 560/13-80-21
ASBESTOS ANALYTICAL PROGRAMS BULK SAMPLE ANALYSIS:
NEW YORK CITY AND MARYLAND
by
A. V. Rao
Corette Parker
Deborah Whitehurst
Don Lentzen
Ty Hartwell
Research Triangle Institute
Research Triangle Park
North Carolina 27709
and
Joseph J. Breen
Office of Pesticides and Toxic Substances
Exposure Evaluation Division
U.S. Environmental Protection Agency
EPA Contract Number 68-01-5848
EPA Task Manager: Cindy Stroup
EPA Project Officer: J. Carra
Design and Development Branch
Exposure Evaluation Division
Office of Pesticides and Toxic Substances
Washington, D. C.
September 1980
-------
PREFACE
This document describes the statistical analyses of
bulk sample data taken from school buildings in two loca-
tions: New York City and Maryland. The bulk samples were
analyzed for asbestos by various laboratories and the re-
sults given to the Research Triangle Institute for analysis.
In particular, Section I of this report describes the analy-
sis of the New York City data which reported the presence or
absence of asbestos (Amosite or Chrysotile) in 474 bulk
samples taken from various types of building material in New
York City schools. Section II describes the analysis of the
Maryland data which contains analysis of 37 split-asbestos
bulk samples taken from Maryland public schools. The split-
samples in Maryland were sent to two or more laboratories
which determined presence or absence of asbestos (Amphibole
and Chrysotile) as well as actual levels of asbestos found.
ii
-------
TABLE OF CONTENTS
Page
I. NEW YORK CITY
1.1 Introduction 1
1.2 Distribution of Asbestos by Type of
Material 1
1.3 Distribution of Asbestos in Friable and
Non-Friable Samples 1
1.4 Summary 2
II. MARYLAND
II. 1 Introduction 26
II.2 Data 26
II. 3 Statistical Methodology 27
II. 3.1 Introduction... 27
II. 3. 2 Model... 27
II. 3. 3 Analysis of Variance (ANOVA) 28
II. 3. 4 Nonparametric Methods 29
II. 4 Results .. 29
II.4.1 Comparisons of the Data on
Presence/Absence of Asbestos
Among Laboratories '.. . 29
II. 4.2 Summary Statistics 30
II.4.3 Comparisons of Laboratory
Determinations Among Laboratories.. 30
II. 4.4 Analysis of Variance 31
II. 5 Summary and Conclusions 31
REFERENCES. 32
ill
-------
LIST OF TABLES
Table Title Page
I.I.A Distribution of Asbestos Present According
to Type of Material 4
I.l.B Distribution of Asbestos Present According
to Material Type 5
I.2.A Summary Statistics for Friable and Non-
friable Samples With Asbestos 6
I.2.B Summary Statistics for Non-fraible Samples
With Asbestos 6
I.2.C Summary Statistics for Friable Samples With
Asbestos 6
II. 1 Data Listing 33
II.2 Presence of Asbestos by Laboratories 35
II.3 Summary Statistics on Levels of Asbestos
by Laboratory 36
II.4 Test Statistic Based on Friedman Procedure
for Comparing Levels of Asbestos for the
Three Laboratories 37
II.5 Test Statistic Based on Wilcoxon Signed
Rank Sum Procedure for Comparing Levels of
Asbestos for Paris of Laboratories 37
II. 6 Results Based on ANOVA 38
IV
-------
LIST OF FIGURES
Figure Title
I.I".A Distribution of the Percent Total Asbestos
in the 235 Friable and Non-Friable Samples
Containing Some Asbestos 17
I.l.B Distribution of the Percent Total Asbestos
in the 91 Non-Friable Samples Containing
Some Asbestos 18
I.l.C Distribution of the Percent Total Asbestos
in the 144 Friable Samples Containing Some
Asbestos 19
I.2.A Distribution of the Percent Amosite in the
44 Friable and Non-Friable Sample Contain-
ing Amosite 20
I.2.B Distribution of the Percent Amosite in the
4 Non-Friable Samples Containing Amosite.. 21
I.2.C Distribution of the Percent Amosite in the
40 Friable Samples Containing Amosite 22
I.3.A Distribution of the Percent Chrysotile in
in the 206 Friable and Non-Friable Samples
Containing Chrysotile 23
I.3.B Distribution of the Percent Chrysotile in
the 88 Non-Friable Samples Containing
Chrysotile 24
I.3.C Distribution of the Percent Chrysotile in
the 118 Friable Samples Containing
Chrysotile 25
-------
I. NEW YORK CITY
Data Supplied by:
Roger A. Chi1jean
Anthony R. Smith
New York City
Board of Education
Division of School Buildings
-------
I.I Introduction
This data comes from the polarized light microscopic
(PLM) analysis of building material bulk samples taken
during 1977-1979 from New York City schools. The samples
were analyzed for asbestos by Walter C. McCrone Associates
for the New York City Board of Education, Division of
School Buildings. For each sample analyzed, the following
data items were available:
1. School identification,
2. District,
3. Type of material (acoustic plaster, spray fire-
proofing, soft acoustic spray, or thermal insu-
lation, and miscellaneous pipe covering, tile,
etc.),
4. Whether asbestos is present or not,
5. Type of asbestos present (amosite, chrysotile),
and
6. Percent of asbestos present.
Other data items available on the New York City raw data
sheets such as dates of collection, mailing, and analysis of
samples were not included on the computer file.
There were 474 samples in this data set. A listing of
the data with an explanation of the codes used is given in
Appendix A.
I.2 Distribution of Asbestos by Type of Material
The data on presence/absence of asbestos are crosstabu-
lated with the type of material in Table I.I.A. The percent
of samples with asbestos varied from 31.6% among acoustic
plaster samples to 79.3% among miscellaneous samples. How-
ever, it is important to note here for friable materials
(codes 2, 3, and 4) that the percent of samples with asbestos
present was 77.4% as shown in Table I.l.B. Compared to the
other three types of material, a statistically significant
smaller percentage of acoustic plaster (.which is non-friable)
samples Cp<.0001) had asbestos.
1.3 Distribution of Asbestos in Friable and
Non-Friable Samples
The distribution of total asbestos in the 235 friable
and non-friable samples with asbestos is given in Figure
I.I.A. Figure I.2.A gives the distributions of amosite and
Figure I.3.A gives the distribution of chrysotile. Similar
distributions are given for non-friable samples in Figures
I.l.B, I.2.B, and I.3.B while distributions for friable
samples are given in Figures I.l.C, I.2.C, and I.3.C.
-------
The related summary statistics for all friable/non-
friable samples are given in Table I.2.A. Summary statis-
tics for non-friable samples containing asbestos are given
in Table I.2.B with those for friable samples containing
asbestos in Table I.2.C. While 86.8% of the non-friable
samples with asbestos had less than 10% asbestos, only 8.3%
of the friable samples with asbestos had less than 10%
asbestos. All 4 of the non-friable samples containing
amosite had less than 10% amosite with 22.5% of the friable
samples with amosite containing less than 10% amosite. For
samples containing chrysotile, 86.4% of the non-friable
samples contained less than 10% chrysotile with just 9.3% of
the friable samples containing less than 10% chrysotile.
The percentages of non-friable samples with 1% asbestos were
50.0% for amosite, 14.8% for chrysotile and 14.3% for total
asbestos. For friable samples, none of the samples contain-
ed just 1% asbestos. The range of percent asbestos present
in non-friable samples was 1 to 80 for chrysotile and total
asbestos and 1 to 3 for amosite. Ranges of percent asbestos
present in friable samples were 5 to 90 for amosite, 3 to 90
for chrysotile, and 3 to 95 for total asbestos. As indicated,
although asbestos can be present in both friable and non-
friable samples, much higher percentages of asbestos tend to
be present in those friable samples containing asbestos.
I.4 Summary
Four hundred seventy-four samples of building materials
gathered from New York City schools during 1977-1979 were
analyzed by Walter C. McCrone Associates for asbestos con-
tent. For all of the samples, 49.6% had asbestos. Compared
to samples of other types of material, a lower percentage of
acoustic plaster (non-friable) samples had asbestos (Table
I.l.B) and, for those acoustic plaster samples with asbestos,
the mean level was 8.98% with a standard deviation of 15.68
(Table I.2.B). -For friable materials, the percentage of
samples with asbestos was between 70 and 80% (Table I.I.A)
and, of those with asbestos, the mean level was 46.61% with
a standard deviation of 27.86 (Table I.2.C).
The mean percent of asbestos in both friable and non-
friable samples by type of asbestos was 26.5% for amosite,
30.7% for chrysotile, and 32.0% for all asbestos (Table
1.2.A). For non-friable samples alone, this mean percent of
asbestos by type was 1.75% for amosite and 9.2% for chryso-
tile (Table I.2.B). The mean percent asbestos in friable
samples by type was 29.0% for amosite and 46.8% for chryso-
tile (Table I.2.C). Samples with less than 10% total as-
bestos constituted about 39% of all postive bulk samples,
over 85% of the positive non-friable bulk samples, and under
25% of the positive friable bulk samples. The range of per-
-------
cent total asbestos present was 1 to 95 in the 474 total
samples, 1 to 80 in the acoustical plaster samples, and 3 to
95 in the friable samples.
-------
Table I.I.A. DISTRIBUTION OF ASBESTOS PRESENT ACCORDING
TO TYPE OF MATERIAL
Presence
of
Asbestos
No
Yes
Total
Type of Material *
1
197i/
82. 4l/
68. 4^/
91
38.7
31.6
288
60.8
2
14
5.
21.
52
22.
78.
66
13.
9
2
1
8
9
3
11
4.
28.
27
11.
71.
38
8.
6
9
5
1
0
4
17
7.
20.
65
27.
79.
82
17.
1
7
7
3
3
All
239
50.
235
49.
474
4
6
Chi-square = 95.7 with 3 degrees of freedom; p < .0001
* Type of Material codes are:
1 - Acoustic plaster
2 - Spray fireproofing
3 - Soft acoustic spray/thermal insulation
4 - Miscellaneous - pipe covering, etc.
Note, in general, materials 2, 3, and 4 are friable while
material 1 (acoustic plaster) is not friable.
I/ frequency
2_/ row percent
3/ column percent
-------
Table I.l.B. DISTRIBUTION OF ASBESTOS PRESENT ACCORDING
TO MATERIAL TYPE
Presence
of
Asbestos
No
Yes
Total
Material
Non-Friable
197i/
82. 4-/
68. 4s-/
91
38.7
31.6
288
60.8
Type
Friable
42
17.6
22.6
144
61.3
77.4
186
39.2
All
239
50.4
235
49.6
474
Chi-square = 94.922 with 1 degree of freedom; p < .0001
I/ frequency
2/ row percent
3/ column percent
-------
Table I.2.A. SUMMARY STATISTICS FOR FRIABLE AND NON-FRIABLE
SAMPLES WITH ASBESTOS
Type of
Asbestos
All
Asbestos
Amosite
Chryso-
tile
** percent
Table I.2.B
Type of
Asbestos
All
Asbestos
Amosite
Chryso-
tile
** percent
Table I.2.C
Type of
Asbestos
All
Asbestos
Amosite
Chryso-
tile
Standard
Mean Deviation
32.04 30.11
26.52 26.66
30.74 29.81
with 1%
Range of
Values
1 to 95
1 to 90
1 to 90
Percent
Samples
With Less
Than 10%
Asbestos
38.7
(5.5)**
29.5
(4.5)
42.2
(6.3)
No. of
Samples
235
44
206
. SUMMARY STATISTICS FOR NON-FRIABLE SAMPLES
WITH ASBESTOS
Percent
Samples
With Less
Standard Range of Than 10% No. of
Mean Deviation Values Asbestos Samples
8.98 15.68
1.75 0.96
9.20 15.90
with 1%
1 to 80
1 to 3
1 to 80
86.8
(14-.3)**
100.0
(50.0)
86.4
(14.3)
. SUMMARY STATISTICS FOR FRIABLE SAMPLES
ASBESTOS
Percent
Samples
With Less
Standard Range of Than 10%
Mean Deviation Values Asbestos
46.61 27.86
29.00 26.73
46.80 27.58
3 to 95
5 to 90
3 to 90
8.3
(0.0)**
22.5
(0.0)
9.3
(0.0)
91
4
88
WITH
No. of
Samples
144
40
118
** percent with 1%
-------
Number
of
Samples
Frequency
70 *
1
60 1
t
50 *
i
40 1
!
I
I
30 1
20 *
i
10 :
i
i
i
•
*****
*****
*****
*****
*****
*****
*****
*****
'*****
*****
*****
*****
* * * * *
*****
*****
*****
*****
*****
*****
*****
*****
*****
Percent Total
Asbestos
1 -
-
19 -
31 -
43 -
55 -
67 -
79 -
91 -
Tota
*****
*****
*****
*****
*****
***** *****
***** *****
***** *****
***** *****
1-6 7-18 19-30 31-42
6
18
30
42
54
66
78
90
95
1
Number of
Samples
72
45
25
14
5
26
20
27
1
235
*****
*****
Cumulative
30.6 30.6
19.2 49.8
10.6 60.4
6.0 66.4
2.1 68.5
11.1 79.6
8.5 88.1
11.5 99.6
.4 100.0
100.0
* * * * ft
43-54 55-66 67-78 79-90 91-95
Percent Total Asbestos
Figure I.I.A.
Distribution of the Percent Total Asbestos in the 235 Friable and Non-Friable
Samples Containing Some Asbestos
-------
Number
of
Samples
Frequenc
45 ;
1
1
tfU *
I
!
i
35 . ;
I
I
1
30 »
i
i
i
t
25 *
!
i
t
«
20 *
1
1
S
15 *
1
•
10 !
i
i
f
5 »
1
1
1
y
A * * * *
*****
*****
*****
*****
*****
* * * ft *
*****
*****
*****
*****
*****
*****
*****
*****
* ft * * *
* ft * * *
*****
*****
* * * * ft
* ft ft * ft
*****
*****
*****
ft * * * ft
*****
*****
* * ft * *
*****
*****
*****
*****
*****
*****
* * ft * ft
ft ft * ft *
* ft ft * *
* * ft * *
*****
*****
*****
*****
*****
*****
* * ft * *
*****
*****
*****
* ft * * *
*****
*****
* ft * ft *
*****
ft ft ft * *
4 * ft ft *
* ft * * A
* ft * * *
* ft * * ft
Percent Total Number of Cumulative
Asbestos Samples % %
1-4 38 41.8 41.8
5-14 45 49.4 91.2
15-24 0 0 91.2
25-34 1 1.1 92.3
35-44 1 1.1 93.4
45-54 1 1.1 94.5
55-64 2 2.2 96.7
65-74 2 2.2 98.9
75-84 1 1.1 100.0
Total 91 100.0
*****
***** ***** ***** ***** ***** *****
1-4
5-14 15-24 25-34 35-44 45-54
Percent Total Asbestos
55-64
65-74
75-84
Figure I.l.B.
Distribution of the Percent Total Asbestos in the 91 Non-Friable ;Samples Containing
Some Asbestos
-------
Percent
Total
Asbestos
1 -
7 -
19 -
31 -
43 -
55 -
67 -
79 -
91 -
6
18
30
42
54
66
78
90
100
Number of
Samples
8
26
24
13
4
24
18
26
1
r/
/o
5.6
18.0
16.7
9.0
2.3
16.7
12.5
18.0
0.7
Cumulative
%
5.6
23.6
40.3
49.3
52.1
68.8
81.3
99.3
100.0
Total
144
100.0
Frequency
i . *****
25 * *****
i ***** *****
i ***** *****
i ***** *****
i ***** . *****
20 * ***** *****
i ««*«« *****
i ***** *****
*****
• ***** *****
1C + ***** *****
Number ? ***** *****
Q£ I ***** ***** *
i ***** ****** *
Samples , .
JLO + ***** ***** *
• ***** ***** *
i *«* ***** ***** •
i *** ***** ***** *
t * * * ***** ***** *
C 4 *ft* *»*** ***** *
I *** ***** ***** *
1 *** ***** ***** 4
1 ft** ***** ***** *
1 * ft * ***** ***** *
*
ft
*
*
*
ft
*
*
ft
ft
ft
4 *
*
*
*
*
ft
*
*
*
* ft * ft * *
ft ft ft 4 4 *
« ft ft 4 ft *
ft ft ft ft ft *
* *
ft ft
ft ft
* *
* ft
ft ft
ft ft
ft ft
• ft
* ft
ft ft
* *
* *
* ft
ft ft
ft 4
ft *
ft ft
ft 4
4 4
4 ft
ft ft
ft ft
4 *
4
I
1
1
4 4
4
4
4
ft
4
4
4
4
4
4
4
4 4
44
4 4
4 4
4 4
ft 4
ft
ft
ft
4
ft
ft
ft
4
ft
4
4
4
4 4
4 4
4 *
ft 4
4 4
4 4
ft ft 4
ft 4 4
ft ft 4
ft ft 4
ft ft 4
* ft 4
« ft 4
ft ft 4
4 ft ft
ft 4 ft
* * *
ft ft 4
ft ft ft
* ft ft
* * 4
* 4 4
* 4 4
ft ft ft
444
ft * 4
ft ft 4
444
ft ft 4
ft ft 4
ft 4 4
444 444*4
1-6
7-18 19-30 31-42 43-54 55-66
Percent Total Asbestos
67-78
79-90
91-100
Figure I.l.C.
Distribution of the Percent Total Asbestos in the 144 Friable Samples Containing
Some Asbestos
-------
Number
of
Samples
Frequency
20 •„
19=5
18
17 '
16
15
14
13 '
12
11 •
10 v
9 '
8
7
6 it
•
5$
4*
3
2
1
+
i
•*
i
+
i
+
i
4
!
»
i
4
1
+
1
•»
|
*
1
+
f
+
(
*
1
*,
4
1
4 >
1
4
4
1-6
* * *
* * *
4 * *
* * *
* * *
* * *
* * *
* * *
* * *
* * *
* ft *
* ft ft
* ft *
* ft *
* * *
* * *
* ft *
ft ft ft
ft * *
* * *
* * A
* * *
* ft ft
* * *
* ft ft
* * *
* * *
* ft *
* * *
* * *
* * *
* * *
ft ft *
ft
*
*
*
*
*
*
*
*
*
ft
*
*
ft
*
*
*
*
*
w
*
ft
*
*
*
*
*
*
ft
*
*
*
*
*
*
*
*
*
4
*
*
*
*
*
*
* *
*
*
«
*
* ,
*
*
*
*
*
*
*
*
*
*
*
*
*
*
ft
7-18 19-30
Percent
Amosite
1 -
7 -
19 -
31 -
43 -
55 -
67 -
79 -
4 ft * *
* ft ft ft
* ft * ft
6
18
30
42
54
66
78
90
*
ft
£
31-42
Number of
Samples
6
20
6
2
0
3
4
3
43-54
%
13
45
13
4
0
6
9
6
.6
.5
.6
.6
.0
.8
.1
.8
*
*
*
ft
55-66
Cumulative
%
13.6
59.1
72.7
77.3
77.3
84.1
93.2
100.0
67-78
*
*
ft
*
ft
79-90
Percent JSmosite
Figure 1.2.A.
Distribution of the Percent Amosite in the 44 Friable and Non-Friable Samples
Containing Amosite
-------
Number
of
Samples
Frequency
2 *
i
i
i
!
!
!
i
•
i
t
i
i
!
!
i
!
i
i
' \
i ;
!
i
!
i
i
i
!
i
i
i
i
•
i
i
i
i
i
!
I
* * * * *
....* Percent Nu
***** Amoslte S
* * * * * "I
***** -
*****
***** 3
*****
..... Total
*****
*****
*****
*****
*****
*****
***** * * *
***** • * * *
***** * * *
***** ft * *
***** ft**
***** * * *
***** ***
***** * * *
***** ***
***** ***
***** * * *
***** ft ft ft
***** * * *
***** ***
***** **•
***** * * *
***** ***
mber of Cumulative
amples % %
2 50.0 50.0
1 25.0 75.0
1 25.0 100.0
4 100.0
• * *
* * *
* * ft
ft * ft
• . - * * ft
ft * ft
ft * *
ft ft ft
•' * ft ft
ft 4 ft
ft ft ft
ft * ft
ft * ft
ft ft ft
ft * ft
ft ft ft
ft ft ft
ft * ft
ft * ft
ft * ft
Percent Amosite
Figure I.2.B. Distribution of the Percent Amosite in the 4 Non-Friable Samples Containing Amosite
-------
Frequency
Number
of
Samples
21 '
20
19
18
17 •
16 '
15 '
14"'
13 '
12 *
11 !
10
9
8
7
6
5
4 '
3
2 '
4
1
I
4
4
1
*
4
1
4
4
1
4
1
4
I
4
1
+
I
4
!
4
1
4
1
4
1
4
1
4
1
4
'
•*
i
« * * * *
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
***** *****
***** *
***** *****
***** *****
***** *****
Percent
Amosite
1
6
18
30
42
54
66
78
90
- 5
- 17
- 29
- 41
- 53
- 65
- 77
- 89
- 100
Total
Number of
Samples
1
21
4
4
0
3
4
2
1
40
%
2.5
52.5
10.0
10.0
0.0
7.5
10.0
5.0
2.5
100.0
Cumulative
%
2.5
55.0
65.0
75.0
75.0
82.5
92.5
97.5
100.0
*****
***** ' ' •
***** . *****
***** *****
1-5
6-17 18-29 30-41 42-53
Percent Amosite
54-65
66--77
78-89
90-100
Figure I.2.C. Distribution of the Percent Amosite in the 40 Friable Samples Containing Amosite
-------
Frequency , Percent
Chrysotile
60
; * * * * *
****** 1
****** _
****** '
'+*•*** 19
I ***** 01
***** 31
! *****
! ***** 43
• *;*;
50 +*•*•• 57
40
Number
of
Samples
30
20
10
* ***** '"
! *****
•» *****
• * * * * *
> *****
? * * * * *
* * * * * *
+ * * * * *
* * * * * *
! *****
* * * * * *
! *****
•* *****
• *****
* * * * * *
• *****
* *****
+ ft ft ft * A
* *****
I * ft ft ft *
* *****
- 6
-18
- 30
/ 1
- 42
- 54
- f\(\
w
- 78
- 90
G
Total
*
*
1-6 7-18 19-30 31-42
***,**
43-54
Number of
Cumulative
Samples % %
68
39
19
1 i
14
2
9ft
£.\J
16
22
206
33.0
18.9
9.3
6O
.8
0.9
177
X£ . /
7.7
10.7
100.0
33.0
51.9
61.2
£. O r\
68.0
68.9
81 f.
ox . o
89.3
100.0
*•
55-66 67-78 79-90
, Percent Chrysotile
Figure I.3.A.
Distribution of the Percent Chrysotile in the 206 Friable and Non-Friable
Samples Containing Chrysotile
-------
Number
of
Samples
Frequency
45
40
35
30,
25
20 "
15
10
5
»
!
i
i
*
i
!
i
i
*
i
i
,
«
i
i
i
i
4
1
1
1
t
*
;
•
•
*
i
i
•
i
- *
!
i
i
i
i
i
i
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
4
•
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
4
*
*
*
*
*
*
*
*
A
*
*
*
*
*
* *
* *
* *
* *
* *
* *
* *
* *
* *
* *
* *
* *
* *
* *
* *
* *
* *
* *
* *
* 4
* *
* *
* *
* *
* *
.
*
*
Percent
Chrysotile
1
5
15
25
35
45
55
65
75
-4
- 14
- 24
- 34
- 44
- 54
- 64
- 74
- 84
Total
* ***** *****
*****
Number of
** Samples
35
45
0
1
1
1
2
2
1
88
*****
Cumulative
%
39.8
51.2
0.0
1.1
1.1
1.1
2.3
2.3
1.1
100.0
*****
*****
%
39
91
91
92
93
94
96
98
100
;
.8
.0
.0
.1
.2
.3
.6
.9
.0
*****
***** *****
1-4
5-14
15-24 25-34 35-44 45-54 55-64 65-74. 75-84 .
Percent Chrysotile
Figure I.3.B. Distribution of the Percent Chrysotile in the 88 Non-Friable Samples Containing Chrysotile
-------
Percent
Chrysotile
1
7
19
31
43
55
67
79
- 6
- 18
- 30
- 42
- 54
- 66
- 78
- 90
Number of
Samples
7
20
18
13
1
24
14
21
%
5.9
17.0
15.2
11.1
0.8
20.3
11.9
17.8
Cumulative
%
5.9
22.9
38.1
49.2
50.0
70.3
82.2
100.0
Frequency
20
Total
118
100.0
Number
of
Samples
15
10
I
+ *
1 •
1 »
1 .
1 .
« »
! >
! »
'
1 .
1 4
J » «. * * • .
1 . * . * « *
1 « A * • * *
(
i
* * *
* * *
• * * » * * *
* * * * * * *
• * « • * * *
* * * * * * *
* * * 4 * * *
*** * A 4 » *****
*** 4*** *****
*** *** *****
*** *•* *****
*** *** *****
• ** *** *****
4*4 * * 4 *****
£ £
,
1-6 7-18 19-30 31-42 43-54 55-66 67-78 79-90
Percent Chrysotile
Figure I.3.C. Distribution of the Percent Chrysotile in the 118 Friable Samples Containing
Chrysotile
-------
APPENDIX I.A
-------
CODES FOR THE ASBESTOS DATA
TYPEMAT - Type of Material
1 Acoustic Plaster
Acoustic Plaster - Hard
Acoustic Plaster - Soft
Cement Plaster
Plaster - Light - Dense
Textured Plaster
2 Beam Fireproofing
Fireproofing
Soft Spray-on Fireproofing
Spray Fireproofing
3 Acoustic Material
Acoustic Material (Soft)
Acoustic Spray (Hard)
Acoustic Spray (Soft)
Insulation on Cone.
Soft Asbestos
Soft Fiber
Soft Material Insulation
Soft Spray-on Material
Spray-on Asbestos
Spray-on Fibers
Spray-on Insulation
Spray-on Material
Thermal Insulation (Soft)
Thermal Material
TYPEMAT - Type of Material
4 Acoustic Board
Acoustic Paint
Acoustic Tile
Asbestos Covering
Asbestos Heating
Insulation (Soft)
Asbestos Pipe Covering
Boiler Breeching
Boiler Insulation
Boiler Jacket #3
Breeching Material
Cabinet Backing
Cabinet Lining
Cem. Asb. Lay-in Tile
Chimney Breeching
Duct Cover (Insulation)
Heating Insulation
Insulation
Pipe Covering
Pipe Insulation
Soft Asbestos Covering
Soft Plaster Back for
Acoustic Tile
Tank Jacket
Thermal Insulation
Transite Fiberboard
Vent Hood Insulation
ASBESTOS - Was any asbestos in the sample?
1
2
No
Yes
PCTASB1 - Percent of first type of asbestos in the sample
TYPEASB1 - First type of asbestos in the sample
1 Amosite
2 Chrysotile
3 Crocidolite
PCTASB2 - Percent of second type of asbestos in the sample
TYPEASB2 - Second type of asbestos in the sample
1 Amosite
2 Chrysotile
3 Crocidolite
TOTPCTAS - Total percent asbestos in the sample
16
-------
DATA LISTING
Sample
Number TYPEMAT ASBESTOS PCTASB1 TYPEASE1 PCTASB2 TYPEASB2 TOTPCTAS
1310 C 0
2 210 0 0
3"32601 0 60
4 110 G 0
5410 00
6 110 0 c
7 2 2 30 2 0 39
8 3 2 30 3 0 30
•9 32302 ? 3 J
10 110 C 0
11 2 1 C 0 0
12 110 3 0
13 110 C 0
14 12320 3
40
16 110 C 0
17 22520 5
IS 4 1 0 0 0
19 110 3 ' 0
3
2i 410 G 0
22
23
24 3 2 70 2 C 7G
25 2 2 35 2 G 35
26
27
28
33
2 2 65 2 •?
31
34-
35
37
28
4:
41 i 1 0 n • C
42 4 2 78 2 0 7H
43 326G2 3 6 C
* 4 4 2 13 2 0 10
45 i 1 0 C *
4fc 2271 3 7
47 110 0 'J
48 110
3
2
3
1
X
4
1
2
3
3
i
2
I
1
1
1
1
2
4
1
1
4
4
4
3
n
£.
4
1
4
~
2
1
4
1
4
1
a
-y
4
1
1
2
1
1
1
2
2
2
1
1
1
i
2
2
1
2
1
1
2
1
1
1
2
2
1
1
i
2
2
2
2
1
2
1
1
1
2
C
0
60
0
0
0
30
30
30
0
n
r\
0
3
40
0
K
0
0
3
0
a
0
70
35
c
C
Q
30
65
1
25
G
20
r
0
0
C
1
2
3
2
2
2
2
2
/«»
«
£-
2
2
2
2
2
i
V
0
0
G
0
0
0
0
n
0
r.
«*
J
c
0
0
r
o
0
n
0
G
0
i?
C
£
,*;
0
^/
0
n
0
0
0
Q
J^
?
n
0
4? i i
TO O
± •> .f ./
2 80 2 0 80
10 o r
i i**1
"^ -.I
54 i i n a
17
-------
DATA LISTING
Sample
Number . .' TYPEMAT ASBESTOS PCTASE1 TYPEASE1 PCTASR2 TYPEASB2 TOTPCTAS
55 1 2 3 2 C 3
56 110 0 p
57 2 2 30 2 C 30
58 110 H . C ,
59 4 2 60 2 C 6C
6f, 110 0 0
61 410 C 0
62 2 2 7 2 C 7
63 110 3 C
64 1 2 3 ' 2 C 3
£5 110 G C
66 4 2 ' 78 2 0 78
£7 310 C 0
68 110 0 C
<•
-------
DATA LISTING
Sample
Number TYPEMAT ASBESTOS PCTAS?: TYPEASP! PCTASE?2 TYPEAS32 TOTPCTAS
l "9 110 r> o
nr 11:. : o
111 liG
112 11 G :: 0
113 lie - :•
115 ] 1 0 '• ' ••
lit- 4 2 £ r 2 : £:
117' 2 2 70 1 •? 7 •:
118 2 2 65 2 : 61.
119 • 1 ' 1 G G '•
120 113 C . C-
121 2 2 7 2 C' 7
122 1272T 7
1 23 4 2 PO 1 •: 80
12* * 2 ft. 0 1 f- * 0
125 2 2.3:- 2 : 3"
1 x f. 1 1 "i " '
127 1 1 S : ."•
128 11 o n •:
129 12720 7
13G i 1 0 -• D
131 110
132 i i : r D
133 1 2 3 2 '.:, 3
134 1 1 D * C
135 12520 5
13b 125?": 5
137 IIG ., C
138 2 2 e: i c ft:-
139 1 2 3 2 C 3
1 H C, . 3 2 *" 2 T- 41
: * i i i o
1*2 i ' i o :. a
143 1 2 J 2 :; 7.
1*4 1 2 8 2 - H
145 310 C "
1*6 1 2 f - ' " 2 C1 3
1*7 2 2 7 l : 7
148 : i o •: ' ?
1*9 • i i o : :
153 1 1 G : C
151 1 5 5 2 5
its 2 2 ic i i: . 2 ?r
153 . 12 11 12 2
16* 12120 1
155 210 j :
156 110 : -:
157 11? 0
1 i 8 1 2 S 2 ~! 8
159 2 2 12 1 36-2 46
ibG i i :• : :.
161 3 2 15 1 "• 15
162 i 2 70 2 : 70
19
-------
DATA LISTING
Sample
Number TYPCMAT ASBESTOS PCTASP1 TYPEASP1 PCTASU2 TYPEASB2 TOTPCTAS
H>3 12210 2
16<* 110 C 0
165 1 1 0 : 0
166 12 11 C • •!
167 110 v C
1*8 lie r
169 1 2 1 2 C 1
17C . £ 2 8 1 7, 2 11
171 22 82 82 16
172 22 £5 2 f- 65
173 ft » 2 18 2 - 0 18
174 12620 fc •
175 2 2 3C 1 0 33
176 11C G /
177 110 0 C
178 12 1 2 C . 1
179 II'? C r'
ISO 2252" 5
181 2 2 4" 2 0 4C
182 2;0 C: ,;
183
134 £ 2 4,"! 2 0 40
165
166
187
188
1C9
5
191 2 2 70 2 C 7L
7
193
194
195 2 2 60 2 0' 60
196 1 1 ' C • •: C
197 2 2 £5 2 2?. 1 85
19S 12520 5
199 2 2 70 1 0 70
2Ci
65
5
1 2 ? 2 G
2 .'4
2;.'5
2C6 i 1 C ,' C
207 4 2 75 2 0 75
75
2'~ 4 2 sr. 2 o ar
21 v
211 a. .-J b .-. -J b
7
213
214
235
20
2
r
1
j.
1
1
j.
1
1
2
1
2
ti
i
o
i
2
2
1
1
i
^
i
i
4
4
4
1
a.
i
2
/t
]
2
2
1
1
2
1
1
2
2
2
i •
2
1
c
2
f
2
2
2
1
/-
c^
1
2
2
2
1
tL
2
2
X
25
4,"!
u
t*j
5
3
C
5
70
7
0
60
C
£5
5
70
s^
65
c
I
c
U
C
75
75.
8 r.
c
7
15
3C
7
2
2
2
2
2
5
ti
2
2
2
1
2
2
2
„
i.
2
2
2
2
^,
2
^
C
r,
*•
0
n
r.
u
C
0
G
^
J
•J'
r
2?.
0
V
•'•>,
r
G
u
^
'-
i
,1
3
'j
^-
;
r
|^
-------
DATA LISTING
Sample
Number
217
2 18
2 IS
22 C
221
222
225
?24
225
22 =
727
228
D29
230
231
232
233
234
?35
236
2?7
'"* * 8
239
24C
241
24L
243
244
245
240
247
248
r..' H ^
^ " f-
251
252
253
2 = 4
255
256
257
258
2 K<=
2SO
2-il
2fa2
2 fo.1
2b"
2c5
2 ?:fc
2 '.- 7
2*S
- - TYPE?-1 AT
T
1
2
1
1
X
•7
1
1
1 •
'1
2.
1
i
i
1
L
X
1
/'
1
4
i
a
^
i
C.
2
?
2
3
c.
i.
i
i
•v
i
4 •
4
4
1
4
it
4
1
T
^
2
1
7
4
i
ASFESTCr:
^1
1
1
1
X
4-
1
,j
1
1
c_
2
1
1
i
i
i
2
1
2
2
2
1
2
1
i
1
1
2
2
1
o
1
2
J.
il
1
d
O
£.
cl
2
f.
j.
2
2
1
£
L-
1
2
^
1
PCTASC1
1C
C
o
r
*
1
c
T.
C
0
3D
r
r.
c
0
0
r-
r.
1
o
65
1
35
r.
HC
u
C
2
Q
15
85
r
L
8T-
C
1
n
c 3
c.
c fs
60
70
5
as
C1
70
3
0
8f:
45
0
15
8C
u
TYFEAS"!
2
2
2
2
1
2
ii
2
2
2
1
2
2
2
1
2
^\
c.
<-.
2
2
2
2
2
2
1
2
PCTASP.2 TYFEASB2 T
n
c
r
."-
•^
0
o
."
.-•
f>
'.
3
0
0
r
^
f:
D
r
c-
c
•J
,**
1C 1
c
p
u
r.
\>
C
;T
C
t.'
n
•
T
"
c
c
12 1
0
'
,",
3
C
"•
L
r
c*
15 2
0
C
OTPC1
1C
C
*\
J
'.,
1
,('
•7
r
ij
3 J
t
n
0
V
••t
^.
^
i.
•™
6b
1
35
^
5C
-;
f;
O
*
•".
15
85
;
8'.
:_
1
;'•
2-:-
r "
35
6f;
82
5
85
~.
7:
8
u
85
1 5
f;
"* ".
so
27: t 2 6C
21
-------
DATA LISTING
Sample
Number TYP'EMAT ASBESTOS PCTASB1 TYPEAS31 PCTASB2 TYPEASB2 TOTPCTAS
271
272
273
274
275
276
277
?78
279
2 8 •: •
2fil
282
263
284
285
236
287
268
' 289
2 9 ::
291
292
293
296
297
298
299
30G
3C1
3J2
3113
3 "4
31''
311
312
313
314
516
317
3 IS
319
32f:
321
32 2
7 •'/ ~
>J i. w
3T;4
4
2
a
4
1
1
1
i
4
2
1
5
1
1
1
1
1
1.
1
I
1
7
1
1
j^
p
X
1
3
3
4
1
1
i •
?.
1
i
4
4
3
1
I
4
i
i
J;
1
1
J
4
4
T,
1
-
T
2
2
2
2
1
2
2
1
2
2
1
2
1
1
2
1
i
j.
a
1
i
2
*>
1
1
1
2
2
2
1
2
2
1.
1
2
£-
2
1
2
c-
2
1
2
2
c.
1
1
2
c.'
2
1
2
£
•t
*
2
15
2C
40
r. Ck
0
. 2
p.
fi
15
30
0
35
C
3
2
n
C
T
0
3
45
70
C
r
n
20
10
3
0
20
65
f^
3
c;
25
5
G
75
18
• 83
5
3
60
60
«
u
8
1C
15
r>
90
1C
f,
4
1
c
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
^
sL
r
"
C
o
r
r.
0
o
7
r
V
o
c
•*
r,
G
•w
2
C
r.
^
25 1
0
G
-*
Ci
n
1
0
10 1
7 1
G
••.
L
0
c
*;
•;
u
^
r.
0
0
f
\J
•^
v:
0
c
V
C
G
("»,
.-.
.;;
15
20
4C
55
r
A.
e
c
15
3C
G
35
r\
J
2
s.
G
3
0
ij
45
95
0
j
V
20
i;
*_/
r
30
72
Tj
•.•>
c
25
5
•;:
75
18
80
C
7
6?
60
1
0
H
10
15
r-
80
1 Z
j
4
22
-------
DATA LISTING
Sample
Number TYPEMAT ASBESTOS PCTAS81 TYPEASB1 PCTASB2 TYPEA.SB2 TOTPCTAS
C
326 12520 5
327
328
29 4 2 75 2 G 75
330 .110 CO
331 4 2 79 2 C 79
332
334
335
336
337 2 2 10 2 0 10
338 110 0 0
339 210 0 0
34 "i
342
343 1 1 D ' 0 0
34* 3 2 15 2 0 15
345 110 0 C
346 i 1 ' 0 0 D
347
348
3 4 3
0
351
°l R O 1 1 1 A *5 <•* 1
> D £. ^ C. 1 1: 2 U 1
T 53
354 1 2 55 2 C 55
2T
356
357
3 58
210 D
85
361 1 2 80 20 80
362 2 ' 2 . 65 2 D 65
363
364 1 2 5 2 C 5
365 3 2 '60 1 D 6u
3=6 3 2 75 2 0 75
367 4 2 8C 2 D 80
368
369
37C
371
372
373
374
375
376
377
378 425
23
1
1
1
1
4
1
4
1
2
1
1
1
2
1
2
1
t
I
1
i
1
i
1
1
1
1
1
J.
1
1
^
3
1
1
1
2
2
1
2
1
1
t.
2
4
A
1
1
1
1
1
1
1
1
-
1
2
1
1
2
1
2
1
2
1
1
1
2
1
1
1
2
1
1
2
1
1
1
2
1
1
4
X
2
1
2
2
1
1
1
1
r-.
c.
2
2
1
C.
2
2
2
1
1
1
1
1
1
1
1
1
1
n
5
J
r
75
0
79
0
3
C
0
,n
10
0
0
0
6C
0
3
15
C
0
0
5
0
0
•^
'j
1C
0
55
20
,";
o
0
0
85
80
. 65
0
5
'60
75
ac
0
0
c
0
C
r.
0
c
C
^
2
2
2
2
2
n,
2
2
. 2
2
2
2
2
2
2
1
2
«.'
0
0
0
n
G
C
c
D
U
0
0
c
0
'J
r-,
t-
£
V
0
o
0
r;
o
G
fl
*•>
G
..
;..
n
0
r,
C
i\
;•>
,-;
0
o
"
0
0
u
n
i.-
u
0
D
,"
r\
r
o
r
o
0
p.
c
n
-------
DATA LISTING
Sample
Number TYPEHAT ASBESTOS PCTASB1 TYPEASP1 PCTASH2 TYPEASB2 TOTPCTAS
379 4 2 40 2 0 40
383
381 12720 7
3S2 1 2 5 2-0 -5
383 * 2 70 2 G 70
71
w «J
38b 2 2 20 1 C 23.
t
367 1 ? 7C 2 0 70
3i
7,g9 4 2 8f: 1 C 8 D
39C 11 0 0
7 c i L i r: r
w «• i-
392 1 2 8 2 C H
393 4 2 10 1 G 13
394 2 2 15 2 G 15
3 9C
396 '4 2 &C- 2 0 fe:
397 1' 2 1 2 G 1
39B 1 2 8 2 C 8
399 4 2 12 1 0 12
65
421 12 32 0
0
413 1 1 G C
-;4 12 12 C
4 05
4G£-
4,7
C9 * 2 40 2 ' 4C>
41r i' 2 15 2 0 15
411 « 2 fa 5 ? C 65
412 " 2 65 2 0 65
413 1 1 G 0
C
415
4!6
417
4 18
419
42:;
421
422
423
424 2 2 25 2 22 1 37
42:
427
428
429
43' 3 2 9Q
431 1 1 C
43" 1 1 0
24
4
1
•1
X
1
fc
4
2
1
1
^
4
1
4
1
/4
2
•4
4
1
1
4
<
1
1
1
1
X
1
-------
Sample
Number
453
474
4T.-5
4 3k
437
4 38
't 59
44..
441
442
443
444
445
446
447
44 «
44°
45 •-'
4 3 i
452
453
454
455
456
457
450
459
46 J
461
462
4b3
4fc4
4i5
466
4b7
4-58
4 (•• 9
47v
471
472
473
474
T Y P E H A T
3
1
A.
1
1
1
4
1
1
4
4
i
4
4
4
4
4
£
j_
/.
1
«
X
4
1
1
4
~z
1
*
1
4
1
7-
1
^
i
A.
J.
n
I
t>
i
3
1
ASBESTOS
2
1
1
1
i
1
i
2.
1
\
2
2
o
2
2
2
2
2
•
o
c.
1
2
2
2
1
2
1
I
i
2
?
1
C
fc
1
1
2
1
2
2
i
1
PCTAS01
4 3
".
,"
c
r-
o
7 n
1
V
A
85
c;
'a 5
&;;
8-
75
tO
2 u
r
5
r»
33
a
e.
D
80
0
r.
0
2
60
o
SO
•»
o
•J
G ^
r
50
3
r.
:,,>
-.
TYF-E
2
1
2
2
2
2
2
2
1
2
?
1
2
tl
2
2
2
2-
^
2
1
c
2
DATA LISTING
iAScl POTASS? TYPEASB2 TCTPC'
r 4'
35
1
n
85
c
H5
an
8C
75
f-8
u
3:
5
r>
8C
P
6C
J
s:
7,
25 .
-------
II. MARYLAND
Data Supplied By:
Dr. Max Eisenberg
Division of Health and Mental Hygiene
Environmental Health Administration
State of Maryland
-------
II.1 Introduction
The State of Maryland has developed an asbestos survey
program for its public schools in response to the recently
identified hazards of some building materials. The actual
identification and quantitation of processed asbestos is not
a routine or simple matter. Discrepancies are not uncommon
between analytical laboratories analyzing the same sample.
This situation exists in part because the determination of
asbestos is a relatively new technique for some facilities
and there is as yet no standard analytical protocol or
universal reference material available.
Nevertheless, the need to ascertain the mineral content
of sprayed insulation materials, pipe insulation, and the
like is of central importance to a preventative measures
program. If the material is falsely identified as positive
for asbestos, there will be an unnecessary expenditure of
funds for remedial action and a loss of confidence in the
program when this is discovered. If the material is falsely
identified as negative for asbestos, exposure to the poten-
tial health hazard is left unresolved.
Considering the costs involved and the current level of
public concern, asbestos must be identified with a high
degree of certainty. To this end the State of Maryland
health officials have employed three competent laboratories
and sent split-samples to two or more of the facilities.
Only those results which show agreement between laboratories
are considered valid in their initial survey and worthy of
recommendations for action. The following is a listing and
analysis of data from the first phase of this comparative
testing program.
II.2 Data
The data set consists of the laboratory determinations
of two types of asbestos (amphibole and chrysotile) by three
laboratories (Maryland Geological Survey (MGS), Occupational
Health Laboratory (OCCH), and Maryland Mineral Associates
(MDMA)) on 37 samples. A listing of the data is given in
Table II.1 (eighty-six lab reports in all).
The table shows that
(a) all three laboratories reported data on 10 samples
(samples 1, 11, 12, 14, 15, 19, 21, 24, 25 and
30),
(b) lab determinations by at least two of the three
laboratories were done on all but three samples
(55, 80 and 85) ,
26
-------
(c) the Maryland Geological Survey lab reported data
on all except two samples (.80 and 85) , and also
reported two determinations each on five samples
(12, 14, 19, 24 and 25),
Cd) the Occupational Health Laboratories reported data
on thirty samples, and
(e) the Maryland Mineral Associates (MDMA) lab reported
data only on fourteen samples.
The laboratory reports include an overall assessment on
presence or absence of asbestos for each of the samples
analyzed by that laboratory.
The following conventions and modifications were used
in converting the data into computer readable form:
1. " " was treated as zero
2. When a range is provided by a lab, the mean value
of this range was used.
For those readers not interested in the statistical details,
Section II.5 (page ) presents the summary and conclusions
of the analysis.
II.3 Statistical Methodology
II.3.1 Introduction
The data were essentially analyzed in two ways:
1. To determine if there were differences between
labs with regard to presence or absence of as-
bestos, and
2. To determine if there were differences between
labs with regard to the actual amount (level) of
asbestos found.
Within each of these two basic analyses tests were performed
on two groups of samples; namely those samples with data on
all three labs and samples with data on only two labs ajt one
time.
II.3.2 Model
The data were analyzed with a view to determining
whether the differences among the laboratories with respect
to presence/absence or amount of asbestos were statistically
significant. The data were analyzed using both an analysis
of variance approach and nonparametic approaches based on
ranks. The following basic model was employed for the
analysis:
27
-------
Yij = » + Li + sj + eij
where
Y. . is the laboratory determination (amphibole or
chrysotile) of the presence/absence or amount of
asbestos in the j sample at the i laboratory;
y is a constant term;
L.(IL.=0) is the effect associated with the i
laboratory;
S.(IS.=0) is the effect associated with the j
sample; and
e. . is an error term.
The e..'s are assumed to be independent indentically
distributed random variables. The analysis of variance
approach assumes that the e^'s have a normal distribution
2
with mean zero and standard deviation = a* The nonparam-
etric method assumes that the EJ^'S come from a continuous
population. Thus the assumptions on the e..'s are less
stringent for the nonparametric approach. In both types of
analysis, the null hypothesis is the equality of laboratory
effects,
L, = L2 = L_; i.e., is percent present or the level
of asbestos found the same for
each of the labs.
II.3.3 Analysis of Variance (ANOVA)
The data on all the 37 samples (excluding the data on
the second determination on five samples at MGS) were used
for this analysis. The ANOVA was only used to test that the
amount of asbestos was the same for the various labs and to
estimate measurement errors.
A separate analysis of variance was carried out with
the data on the five MGS samples with duplicate determina-
tions for the purpose of estimating measurement errors. .
28
-------
II.3.4 Nonparametric Methods
The significance of the differences among laboratory
levels of asbestos was examined with the data from the ten
samples analyzed by all the three laboratories. The sample
was treated as a "block" and the data were analyzed like a
randomized block design with three treatments and ten blocks
using the rank order statistic proposed by Friedman (see
pages 262-265 of the reference). For each sample, the
amounts of asbestos for laboratories were ranked (separately
for chrysotile and amphibole) from one to three and these
ranks were analyzed using the procedure.
The differences between pairs of laboratories for
amounts of asbestos were analyzed using all the samples with
determinations from both the laboratories of the pair. The
number of samples available for these pairwise comparisons
are 30 for MGS vs. OCCH, 14 for MGS vs. MDM and 10 for OCCH
vs. MDM. The data were analyzed using the Wilcoxon signed-
rank Test (see pages 124 - 131 of the text referenced) using
a normal approximation.
The data on presence or absence of asbestos on the ten
samples analyzed at all the laboratories were analyzed using
the Cochran statistic (see page 267 of the text referenced).
The McNemear Test (see page 268 of the text referenced) was
used to compare pairs of laboratories on the presence or
absence of asbestos.
II.4 Results
II.4.1 Comparisons of the Data on Presence/Absence of
Asbestos Among Labs
There were no disagreements in presence/absence of
asbestos between the duplicate determinations made by MGS on
the five samples (12, 14, 19, 24, and 25).
The distribution of the data on the ten samples analyzed
by all the three labs is as follows:
No. of Samples
All reported asbestos present 6
All reported asbestos absent 1
Two labs reported asbestos present 2*
Two labs reported asbestos absent 1**
Total 10
* For samples 12 & 14 MGS and MDMA reported asbestos
present while OCCH reported asbestos absent
** For sample 1 OCCH and MDMA reported asbestos absent
while MGS reported asbestos present
29
-------
This data was analyzed to test the difference among the
three labs with respect to their assessment of presence or
absence of asbestos using the Cochran procedure (see page
267 of the reference). The differences were not significant
at the 0.05 level (test statistic of 4.667, P = 0.10).
The distribution of the data on the 24 samples analyzed
by only two laboratories is as follows:
Both reported asbestos 11
Both reported no asbestos 13
Total 24
Thus, there were no disagreements between the pairs of labs
involved in the analysis of these twenty-four samples.
The cross tabulations of the data, on all the 34 samples
analyzed by at least two labs, for pairs of labs are given
in Tables II.2a, II.2b, and II.2c. None of the differences
are statistically significant (by the McNemear Test).
To sum up, there were disagreements between the labora-
tories only in three of the 34 (8.8%) samples. The diffe-
rences among the labs was not significant at the 0.05 level.
II.4.2 Summary Statistics
Table II.3 presents the summary statistics for the
determinations of levels of asbestos made by each of the
laboratories. Note that since all laboratories did. not
analyze all the samples, the comparison of laboratories
based on these summary statistics may be misleading.
II.4.3 Comparisons of Laboratory Determinations Among
Laboratories
The test, statistics related to the comparison of labo-
ratories are given in Tables II.4 and II.5.
The Friedman statistics in Table II.4 are based on the
data on the ten samples analyzed by all the three labora-
tories. We note that the differences among the laboratories
are significant at the 0.05 level for chrysotile determina-
tions.
The test statistics based on the Wilcoxon signed-rank
sum procedure for pairwise comparisons of the labs are given
in Table II.5. We note that the differences between MGS and
MDMA are significant for both chrysotile and amphibole.
Compared to the determinations at MGS, the % asbestos level
determinations at MDMA are significantly lower for chryso-
tile and higher for amphibole.
30
-------
II.4.4 Analysis of Variance
The ANOVA on the differences between labs on levels of
asbestos for all 37 samples was not statistically signifi-
cant. However, the assumptions underlying this analysis are
suspect (e.g., the data is not normally distributed); and
therefore, the non-parametric analyses given above are
probably more appropriate.
The estimates of the measurement errors using ANOVA and
other summary statistics are given in Table II.6. The
square-root of the residual mean square error under the
model is shown as the measurement error in the table. The
measurement error is an estimate of the variability of
repeated measurements on the same sample.
II.5 Summary and Conclusions
The data on presence/absence of asbestos were analyzed
to determine whether there were differences among the labo-
ratories. There were disagreements between the labs only in
three of the 34 (8.8%) samples analyzed by two or more labs.
Tests of significance between labs for presence or absence
of asbestos using the Cochran procedure or the McNemear Test
were not significant.
The data on laboratory % asbestos levels of chrysotile
and amphibole carried out by three laboratories on 37 samples
were also analyzed with a view to determine whether there
were any differences among the laboratories. The analysis
indicated that there were differences between MGS and MDMA.
Compared to the laboratory levels at MGS, the % asbestos
levels by MDMA were lower for chrysotile and higher for
amphibole. Thus, % asbestos levels difference between labs
were noted but lab differences on the presence/absence of
asbestos were not significant.
31
-------
REFERENCE
Lehman, E.L., "Nonparametrics: Statistical Methods Based on
Ranks", Holden-Day, Inc., San Francisco, 1975.
32
-------
Table II.1 DATA LISTING
LAB
CHRY
PRESENT
1
1
t
*
2
2
3
3
a
a
5
5
b
b
7
7
e
3
10
10
11
11
11
12
12
12
12
13
13
1*«
1«*
l<»
I '4
1 5
l5
15
10
lb
17
17
13
13
19
I?
19
19
2«
2U
el
21
21
22
22
23
23
2 a
2«
M DMA
MHS1
UCCH
MCS1
CCCH
f*GSl
IjCCH
MGSi
uCCH
MGSl
QCCH
MGSi
QCCH
MGSl
OCCH
MGSl
UCCH
MGSl
UCCH
MO HA
MGSl
GCCH
MDf-jA
MGSl
MGSl
QCCH
ilGSl
OCCH
M DMA
MGSl
MGSl
UCCH
MD^A
MGol
OCCH
MGSl
UCC*
MGSl
OCCH
MGSl
UCCH
MDMA
MGSl
MGSl
UCCH
MGSl
UCCH
MO MA
MGSl
. OCCH
MGSl
OCCH
MGSl
OCCH
MOM A
•iGSl
0.0
0.5
0.0
«5.0
30.0
0.0
0.0
o.o
o.o
30.0
25.0
35.0
15.0
0.0
0.0
0.0
0.0
0.0
0.0
2.5
7.5
o.o
0.0
15.0
10.0
0.0
o.o
o.o
0.0
25.0
30.0
o.o
0.5
1.5
5.0
0.0
0.0
30.0
30.0
o.o
0.0
90.0
100.0
100.0
50.0
o.o
0.0
0.5
7.5
5.0
0.0
o.o
25.0
37.5
0.0
0.0
c.o
0.0
o.o
12.5
30.0
0.0
0.0
o.o
0.0
U5.0
25.0
20.0
15.0
o.o
0.0
o.o
0.0
' 0.0
o.o
1.5
o.o
10.0
10.0
0.0
o.o
0.0
0.0
o.o
10. 0 .
o.o
o.o
0,0
0.0
0.0
10,0
o.o
0.0
o.o
0,0
0,0
0.0
o.o
o.o
o.o
0.0
0.0
o.o
o.o
0.0
o.o
o.o
o.o
0.0
o.o
0.0
0,0
Nn
YES
MO
YES
YES
NO
NO
MO
NO
YES
YES
YES
YES
. "40
NO
NO
NO
;MO
NO
YES
YES
YES
YES
YES
YES
NO
HO
NO
YES
YES
YES
NO
YES
YES
YES
NH
NO
YES
YES
so
NO
'YES
YES
YES
YES
MO
MO
YES
YE?
YES
•MO
NO
YES
YES
NO
N(J
33
-------
Table 11,1(Continued)
,
30
30
30
35
33
uO
40
a5
<45 -
*o
-jo
">5
60
oU
03
VJT
70
70
75
75
40
43
LAB
MGol
OCCH
MDMA
MGSl
MGSl
OCCH
MGSl
OCCH
M 0 M A
MGSl
OCCH
MDMA
M331
MOMA
MG61
MOMA
MGSl •
MOMA
MGSl
flSSl
MGSl
OCCH
hGSl
OCCH
Ml. 31
rjcuH
MGSl
QCCM
OCCH
UCC'1
CHRy
o.o
0,0
0,0
1.0
1.5
0 , 0
17.5
42,5
35.0
40.0
35. c
0,0
0.0
4-).0
25,0
0.0
0,0
o.o
",0
o.o
o.o
0.0
0,0
0,0
25.0
45.0
15.0
80.0
0 ,0
20.0
AMPH
0,0
:) . o
1.5
0 . .)
O.T
10.0
0,0
0,0
0,0
0,0
0,0
0,0
0.0
0,:)
0. j
0,0
0.0
0,.)
0.0
0.0
70.0
90,0
7 0 . j
90.0
0.0
0,0
0,0
0 ,0
0.0
40.0
PRESEN
MO
MO
YES
YES
YES
YES
YES
YES
YES
YES
YES
NO
NO
YES
YES
NO
NO
MO
MO
NO
YES
YES
YES
YES
YES
YES
YES
YES
MQ
YES
34
-------
Table II.2. PRESENCE OF ASBESTOS BY LABORATORIES
(a) MGS vs OCCH
MGS
OCCH
No
Yes
Total
No
11
3
14
Yes
0
16
16
Total
11
19
30
(b) MGS VS MDMA
MGS
MDMA
NO
Yes
Total
No
4
1
5
Yes
0
9
9
Total
4
10
14
(c) OCCH VS MDMA
OCCH
MDMA
No
Yes
Total
No
2
0
2
Yes
2
6
8
Total
4
6
10
35
-------
Table II.3. SUMMARY STATISTICS ON LEVELS OF ASBESTOS BY
LABORATORY
Labo-
ratory
MDMA
MGS*
OCCH
Asbestos
Type
Amphibole
Chrysotile
Amphibole
Chrysotile
Amphibole
Chrysotile
No. of
Determi-
nations
14
14
35
35
32
32
Mean
(%)
1.64
12.04
6.21
12.73
10.0
13.13
Standard
Deviation
(%)
3.58
26.17
18.01
20.55
23.17
20.44
Coefficient**
of Variation
85
125
62
93
92
63
* MGS data includes duplicate determinations on 5 samples.
Only one of these determinations are included here.
** Based on determinations which are non-zero.
36
-------
Table II.4. TEST STATISTIC* BASED ON FRIEDMAN PROCEDURE
FOR COMPARING LEVELS OF ASBESTOS FOR THE THREE
LABORATORIES
Determination
Chrysotile
Amphibole
Test
Statistic
12.45
4.35
P-Value
< 0.025
< 0.12
* The formula for use in presence of ties is used here
(see pages 265 of the reference cited).
Table II.5. TEST STATISTIC BASED ON WILCOXON SIGNED RANK
SUM PROCEDURE FOR COMPARING LEVELS OF ASBESTOS
FOR PAIRS OF LABORATIRIES
Comparison
MGS vs OCCH
MGS vs MDMA
OCCH VS MDMA
Chrysotile
1.219
2.0989*
0
Amphibole
- 1.436
- 1.9936*
0.2210
No. of
Samples
30
14
10
* Significant at 0.05 level (2-tail test).
The formula for use in presence of ties is used
here (see pages 130-131 of the reference cited).
37
-------
Table II.6. RESULTS BASED ON ANOVA
Asbestos
Type
Amphibole
Chrysotile
Chrysotile"1"
Mean (%)
6.5174
13.6686
35.31
ncao uj. ciuc.ii i_
Error-/ (%)
4.686oi/
11. 112 3l/
2.5061"1"1"
Coefficient
of Variation
71.9
81.3
7.10
I/ Square root of the residual mean square error under
the model.
2/ Based on 37 samples.
+ Data on only five duplicate analyses in Maryland
Geological Survey Lab.
++ This is an estimate of the within laboratory vari-
ability for Chrysotile. Amphibole determinations
were all zero.
-------
TECHNICAL HEF'Olir DATA
(I'tcasc read Instructions mi the reverse before
1. REPORT NO.
EPA
2.
4. TITLE AND SUBTITLE
Asbestos Analytical Programs Bulk Sample Analysis:
New York City and Maryland
6. PERFORMING ORGANIZATION CODE
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
7. AUTHORis) A> v> Ra0j corette Parker, Deborah Whitehurst,
Joseph J. Breen, Don Lentzen, Ty Hartwell
8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Research Triangle Institute
Post Office Box 12194
Research Triangle Park, North Carolina 27709
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-01-5848
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Protection Agency
Office of Pesticides and Toxic Substances
Washington, D.C. 20460
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This document describes the statistical analyses of bulk sample data taken from
school buildings in two locations: New York City and Maryland. The bulk samples were
analyzed for asbestos by various laboratories and the results given to the Research
Triangle Institute for analysis. In particular, Section I of this report describes the
analysis of the New York City data which reported the presence or absence of asbestos
(Amosite or Chrysotile) in 474 bulk samples taken from various types of building
material in New York City schools. Section II describes the analysis of the Maryland
data which contains analysis of 37 split-asbestos bulk samples taken from Maryland
public schools. The split-samples in Maryland were sent to two or more laboratories
which determined presence or absence of asbestos (Amphibole and Chrysotile) as well as
actual levels of asbestos found.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
l).IDENTIFIERS/OPtN ENDED TERMS
c. COSATI Held/Group
Asbestos identification in bulk samples
Analysis of split-samples
Asbestos in friable and non-friable bulk
samples
Statistical analysis
8. DISTRIBUTION STATEMENT
19, SECURITY CLASS (This Report)
21. NO. OF PAGES
41
20. SECURITY CLASS (Thispage)
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDI TION is OBSOLETE
------- |