Ch. 3— Patterns of Animal i > . 
e • 59 
Table 3-3.— Reliability of Various Data Sources (Continued) 
Years 
Confidence 
Source 
covered 
rating 
Strength(s) 
Limitation(s) 
Alex Brown & Sons 
1981 
Poor 
At the time, it was thought to 
represent best estimate for lab 
animals in U.S. market 
Data based on a few inter- 
views, and mostly broad 
estimates 
Andrew N. Rowan 
1985 
Poor 
Data distinguishes between 
production, acquisition, and 
actual use 
Broad analysis with many 
assumptions. Based mainly on 
one breeding facility 
Amphibians: 
Emmons 
1969 
Indeterminate 
Global estimates 
Culley 
1981 
Indeterminate 
Many assumptions 
Nace 
1974-81 
Fair 
Fair detail for basis of 
Difficult to know actual 
estimates 
numbers due to large number 
of users 
Various, on fish 
1983 
Fair 
Data consistent 
Global estimates only 
usage 
Various, on bird 
usage 
1983 
Poor 
Good detail by institutions 
Uncertainty about nonreporting 
institutions, and fraction of 
fowl used by lab experimentation 
Data on animal trends: 
Wadsworth Center, NY 
1980-83 
Poor 
Good detail of different 
species used 
Difficult to predict any trends 
Johns Hopkins, MD 
1975-85 
Poor 
Limited data that are 
impossible to analyze 
SOURCE: Office of Technology Assessment. 
million in fiscal year 1967. (The results for all re- 
spondents , while not mentioned in the report, were 
compiled and reported for comparison purposes 
in the ILAR 1980 survey.) The confidence rating 
was "fair.” 
The ILAR National Survey of Laboratory Ani- 
mal Facilities and Resources (22) also appears to 
be a thorough and statistically solid report, al- 
though the data (for fiscal year 1978) are now 8 
years old. Since it also was funded by NIH, pri- 
mary attention was given to nonprofit biomedical 
research institutions eligible for Federal grants. 
In addition, data were received from Federal orga- 
nizations, commercial research labs, and the phar- 
maceutical industry. Seventy -two percent of the 
2,637 questionnaires were returned; 47 percent 
of those were acceptable, thus providing 1,252 re- 
spondents (including 992 nonprofit Federal-grant- 
eligibles, 137commercial laboratories, 25 compo- 
nents of the DOD, 21 units of NIH, and 77 compo- 
nents of other Federal agencies). Although the 
individual identities of the respondents are un- 
known, the biomedical research expenditures of 
the nonprofit organizations are known. Since their 
data are reported separately from all respondents, 
an extrapolation to the unknown cases can be at- 
tempted based on the known national (meaning 
“all use in the United States”) biomedical research 
expenditures. This source was assigned a confi- 
dence rating of “fair.” 
W.B. Saunders & Company 
W.B. Saunders & Company (41) surveyed the lab- 
oratory animal market in 1965 and projected fig- 
ures for 1970. The survey and its estimates are 
widely quoted as one of the first estimates of ani- 
mal use. The survey methodology is unclear and 
the company no longer exists, so these data fall 
under the “indeterminate” category. 
Foster D. Snell, Inc., for 
Manufacturing Chemists Association 
A study performed by Foster D. Snell, Inc., for 
the Manufacturing Chemists Association (25) esti- 
mated that 35 million mice and 40 million rats were 
