Institute for Research on Poverty

Special Report no. 68

Comparison of National and Poor Households:
Results of a Survey of Consumer Knowledge and Risk Perceptions
of Food-Related Biotechnologies
Robin Douthitt
University of WisconsinMadison
Lydia Zepeda
University of WisconsinMadison
Deana Grobe
Oregon State University
January 1996
Acknowledgments
Financial support for this project was provided by the U.S. Department of Agriculture
Cooperative State Research Service, National Research Initiative Competitive Grants Program
#9503405. The authors thank USDA for its support. Any opinions contained herein are solely those of
the authors and do not necessarily represent any views or interpretation by the sponsoring agency.
Table of Contents
Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
SURVEY DESCRIPTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Survey Research Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Survey Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Sample Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
National Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Wisconsin Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Vermont Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Poor Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Response Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
FINDINGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Data Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Calculation of Weights for the DGZ Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Demographic Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
DGZ Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Base-Level Poverty Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Consumer Awareness, Knowledge, and Approval of Biotechnology, rpGH, and rbGH . . . . . 13
Awareness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Approval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Consumers Perceived Concern for Discovery of Human Ill Health Effects . . . . . . . . . . . . . 21
Current Health Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Future Health Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Risk Perception Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Beliefs about Milk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Trustworthiness of Information Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Knowledge of rbGH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Locus of Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Group Affiliation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Consumers Demand Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Consumers Preference for Treated versus Untreated Milk . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Consumer Support for Labeling Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Gender Comparison Regarding Food-Related Biotechnologies . . . . . . . . . . . . . . . . . . . . . . . 46
CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Endnotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
i
Executive Summary
The primary goal of this study was to better understand what factors influence consumers risk
perceptions toward recombinant bovine growth hormone (rbGH) and recombinant porcine growth
hormone (rpGH), and to determine whether these risk perceptions differ between poor and nonpoor
samples. This report summarizes the findings of a national survey of household food shoppers
measuring consumer knowledge and risk perceptions of rbGH and rpGH, one year after the FDA
approved rbGH for commercial use. A total of 1,910 interviews, averaging sixteen minutes in length,
were completed. The findings were reported by comparing poor and nonpoor samples, where the poor
sample was composed of those respondents who met the USDA poverty guideline requirements.
We found that the poor respondents, when compared to the nonpoor respondents, were less
aware of biotechnology and rbGH. A smaller percentage of the poor sample approved of biotechnology
than the nonpoor sample, and a larger percentage of poor respondents disapproved of the use of rbGH.
But awareness of and willingness to purchase rpGH-treated pork if approved by the FDA was similar
for both samples. In both samples, the majority stated there had been no change in the amount of milk
they bought since the FDA approved rbGH. However, 10.6 percent of aggregate fluid milk
consumption was identified by respondents as having come from untreated herd milk. Similarly, a
majority of the poor and nonpoor samples stated they thought milk should be labeled so consumers
could distinguish between milk from treated and untreated herds. More respondents from the nonpoor
sample preferred untreated milk than from the poor sample.
Respondents from the poor sample were more likely to be concerned about the current safety of
rbGH with respect to human ill health effects, and they were also more likely to be concerned about the
future discovery of human ill health effects than the nonpoor respondents. In order to understand the
differences in risk perception between the two samples, we compared the poor and nonpoor
ii
respondents who were concerned to those who were not concerned about the future discovery of health
risks. The findings are as follows:
For both samples, the concerned group disagreed with the milk-belief statements milk is a
natural product and milk is a pure product more so than the not concerned group.
The nonpoor sample ranked physicians as their most trustworthy food-related information
source, followed by the USDA and FDA, while the poor sample ranked the USDA as most
trustworthy, followed by farmers organizations and the FDA. Both samples ranked
advertisements as being the least trustworthy food-related information source.
The concerned group from both samples disagreed that bGH is naturally found in milk;
more respondents from the not-concerned group tended to agree with this statement.
The concerned group had strong agreement with the locus-of-control questions regarding
the future of todays children and feeling helpless in the world, for both samples.
The concerned group members were also more likely to identify with environmentalists and
animal rights groups. A larger percentage of poor respondents strongly identified with these
two groups than nonpoor respondents.
The report also analyzed the differences based on gender. To summarize:
Males were more aware of biotechnology and rbGH.
Males had a greater overall approval of biotechnology and a more favorable opinion of
rbGH.
Females were less likely to want to purchase rpGH-treated pork if it is approved by the
FDA.
Females were more concerned about current and future discovery of rbGHs health effects.
A greater percentage of females preferred milk from untreated herds.
iii
More females felt there should be labels to distinguish between milk from treated and
untreated herds.
Comparison of National and Poor Households:
Results of a Survey of Consumer Knowledge and Risk Perceptions
of Food-Related Biotechnologies
INTRODUCTION
This report summarizes the results of a national survey of food shoppers which measured
consumer knowledge and risk perceptions of two food-related applications of biotechnology. By
developing risk-perception measures, the authors will be able to determine which factors influence
consumers risk perceptions and suggest what role regulators have in influencing those perceptions.
Although biotechnology is being applied to food-production processes, we limit our work specifically to
consumers risk perceptions of recombinant bovine somatotropin (rbGH) and recombinant porcine
somatotropin (rpGH). An evaluation of rbGH and rpGH provides a comparison of whether or not there
are greater risk perceptions for one biotechnology in particular, or if there are similar perceptions of
risk for both biotechnologies. They provide an interesting contrast, particularly because they are similar
substances; rpGH use endows pork with benefits that consumers value, leaner pork products, while
consumers do not perceive a benefit from rbGH (Senauer, Asp, and Kinsey 1991, p. 264).
Previous research has shown inconsistencies regarding incomes effect on consumers risk
perceptions of rbGH-treated herd milk (McGuirk, Preston, and Jones 1992; Grobe and Douthitt 1995).
Thus, a primary goal of the study was to better understand the conceptual relationship between income
and consumers risk perception of two specific food-related biotechnologies. The supporting objectives
were to analyze the correlates to consumer risk perception and the impact of commercial use of rbGH
on consumer demand for fluid milk, one year after rbGHs approval by the Food and Drug
Administration (FDA).
This report will specify the survey design and interview, sample composition, and response
rates. We then present our findings in a descriptive and demographic summary of the poor and nonpoor
samples. Findings are presented for the following issues:
2
awareness, approval, and knowledge of food-related biotechnology applications
concern about current health effects and the future discovery of health effects
perceptions of risk associated with biotechnology food applications, including consumer
beliefs about food product characteristics, confidence in federal food regulators, and
personal experiences and circumstances that may influence risk perceptions
consumer demand for fluid milk in response to the commercial use of rbGH
preference for milk from rbGH-treated or untreated herds,
response to labeling to distinguish between milk from rbGH-treated versus untreated herds,
and
gender comparison regarding food-related biotechnologies.
SURVEY DESCRIPTION
Survey Research Methodology
Focus group sessions were first conducted to explore qualitative insights regarding the effects
of income and gender on the perceived risks of biotechnologies. Three convenience samples of
individuals who lived in or near Benton County, Oregon, were selected, and focus groups were
conducted on November 17, 28, and 29, 1994. Different target groups were represented in each of the
samples, including a sample of low-income women (n=8), middle- to high-income women (n=9), and
middle- to high-income men (n=7). The focus group results were used to help clarify vague empirical
measures of risk-perception antecedents, as well as to assist in formulating more concise questions for
the survey instrument.
1
Survey questions were designed according to theories of risk perception (Eom 1993; Hadden
1989; Grobe and Douthitt 1995) and the results of the focus group sessions. Variables defining risk
perception were incorporated into the question design. Surveys from other studies that have evaluated
3
rbGH and food safety were used as references for particular question wording (Sterngold, Warland, and
Herrmann 1994; McGuirk, Preston, and Jones 1992; Research Alliance 1990; Slusher 1990; Smith and
Warland 1992). The survey instrument was designed to analyze whether there were certain factors or
experiences (personal, health-related, beliefs, attitudes) influencing consumers to feel there was a risk
associated with the rbGH product. The survey instrument design also incorporated the impact of the
commercial use of rbGH one year after its approval by the FDA. Thus, the survey instrument included
items to evaluate: personal circumstances and health risk factors that might influence consumers risk
perceptions, factors of risk perception specific to the biotechnologies (rpGH and rbGH), the effects of
commercial adoption of rbGH on consumer demand for fluid milk, consumer self-protection or risk-
averting response, and support for and availability of product labeling.
Three different survey techniques were used to allow for validity testing of responses for a
particular question order or word usage. For four of the questions, a randomization of responses was
used. For example, when asking about milk beliefs, four belief statements were randomized for each
2
respondent. The second survey technique alternated the use of the terms administering and
injecting rbGH for two questions. The third technique alternated the block of questions for rbGH and
3
the block of questions for rpGH, thus allowing testing for possible influence of the rbGH discussion on
4
the discussion of rpGH and vice versa.
Survey Interview
A nationwide survey of primary household food purchasers attitudes toward the use of rbGH
and rpGH was conducted by the Letters and Science Survey Center (LSSC) from March 1 through June
27, 1995, approximately one year after the FDA approved the commercial use of rbGH. (The LSSC is a
unit of the College of Letters and Science at the University of WisconsinMadison.) A total of 1,910
interviews averaging sixteen minutes in length were completed. The adjusted response rate for the
entire sample frame was 56.1 percent (see Table 1).
4
Thirty-eight survey interviewers were briefed on the background and goals of the study, the
funding source, and a detailed description of the instrument. A pretest consisting of nineteen completed
interviews was performed in mid-February 1995, and the survey instrument was revised based on
interviewers and pretest respondents input. The interview was conducted using a Computer-Assisted
Telephone Interview (CATI) system. The text of each question appeared on the screen for the
interviewer to read. The routing through the interview was computer-determined, based on
programmed skip patterns. Question wording could be adapted according to answers previously given
in the interview. The computer allowed only valid responses; when an invalid response was entered, the
computer asked the interviewer to reenter the response.
Telephone calls were conducted at all times of day and night, including weekends. However,
most interviews were completed in the evening and on weekends. When each telephone number was
called, the interviewer would determine whether or not a working residential number had been reached.
Each residential number was then screened to verify that it was associated with a household. Working
residential households located in the continental United States were then further screened to determine
whether there was at least one household resident who was 18 years or older. Finally, the person
selected as the interview respondent was the person identified as a household resident who is age 18 or
older and primarily responsible for the households food purchasing decisions. Only that person could
be interviewed; no substitutions were allowed.
5
TABLE 1
Response Rate Calculations
Sample Response Rate
Entire Sample: National, WI, VT, Poor
1910 1910
=
1910+1038+145+156+5+(.602*248) 3403
56.1%
National Sample
969 969
=
969+585+80+77+1+(.576*131) 1787
54.2%
Wisconsin Sample
187 187
=
187+77+6+8+(.584*14) 286
65.3%
Vermont Sample
186 186
=
186+73+5+15+(.513*29) 294
63.3%
Poor Sample: 20th Percentile
394 394
=
394+209+37+22+(.702*47) 695
Poor Sample: 10th Percentile
174 174
=
174+94+17+34+4+(.693*29) 343
56.7%
50.7%
Note: The adjusted response rate was calculated as follows:
completed
= .
completed + refused +away for duration +R not available + other + (adjusted)(no answer)
6
Sample Composition
The 1,910 completed interviews consisted of 969 completions from a National sample frame,
187 completions from a Wisconsin sample frame, 186 completions from a Vermont sample frame, and
568 completions from a Poor sample frame. The states of Wisconsin and Vermont were oversampled
because of their food labeling regulations on rbGH. Wisconsin has established voluntary labeling
regulations of products from untreated herds, and Vermont has established mandatory labeling of
products from both treated and untreated herd milk. Results from Grobe and Douthitt (1995) showed
that low-income respondents were more apprehensive toward rbGH-treated herd milk; therefore, poor
households were also oversampled to ensure sufficient degrees of freedom for multivariate analysis of
their behavior.
National Sample. Nine hundred sixty-nine completed interviews were obtained from a National
sample frame of telephone numbers purchased by LSSC from Nielsen Media Research. The adjusted
response rate for the National sample was 54.2 percent (see Table 1). The sample is representative of
currently working residential telephone numbers in the continental United States, including both listed
and nonlisted numbers. Nielsen updates the sample three times a year. It is estimated that
approximately 5 to 7 percent of U.S. households do not have telephones, and would therefore not be
represented in the sample.
Nielsen Media Research begins with a file of all residential telephone numbers that are listed in
published telephone directories. This file is, in effect, sorted by exchange and number within exchange.
Next, within each exchange, ten thousand potential telephone numbers (XXX-0000 through XXX-
9999) are generated and divided into one hundred blocks of one hundred consecutive numbers. If any of
these blocks do not contain listed residential numbers, the block is eliminated. A sample is then drawn
from the remaining numbers. Thus, the sample includes telephone numbers that are listed in the
published directories, those that are unlisted, and numbers within those blocks that have been assigned
7
since the most recent issue of the telephone directory. Use of this sampling scheme is more efficient
than a simple random digit-dialing procedure. The time and expense of making calls to blocks that do
not have currently assigned numbers, or to blocks with nonexistent or nonresidential exchanges, is
avoided.
Wisconsin Sample. One hundred and eighty-seven completions were acquired from the
Wisconsin sample frame. LSSC purchased this sample frame of telephone numbers from Nielsen
Media Research. The adjusted response rate for the Wisconsin sample was 65.3 percent (see Table 1).
The same sampling strategy described above for the National sample frame was utilized by Nielsen
Media Research to select the state of Wisconsin sample frame. The sample frame is representative of
currently working listed and nonlisted residential telephone numbers in the state of Wisconsin.
Vermont Sample. One hundred and eighty-six completions were acquired from the Vermont
sample frame. This sample frame was also purchased by LSSC from Nielsen Media Research. The
adjusted response rate for the Vermont sample was 63.3 percent (see Table 1). Again, the same
sampling strategy described in the section above for the National sample frame was utilized to select
the state of Vermont sample frame. The sample frame is representative of currently working listed and
nonlisted residential telephone numbers in the state of Vermont.
Poor Sample. Five hundred sixty-eight completions were acquired from the Poor sample frame.
Two sample frames of telephone numbers were purchased by LSSC from Survey Sampling, Inc. (SSI).
The samples targeted geographic areas in the continental United States where average household
incomes fall below a specific level. One sample frame was drawn from exchanges within areas where
the average household income was within the lowest 20 percent of U.S. household income, and the
other was from exchanges where the average household income was within the lowest 10 percent of
household income. The response rate for the 20th percentile was 56.7 percent and 50.7 percent for the
5
10th percentile (see Table 1). In the low-income area sample frames, interviews were attempted with
8
all sample cases. Cases were not screened for meeting poverty guidelines before the interviews. The
USDA poverty guidelines were used post-interview to determine the respondents level of poverty.
SSI used the following selection process for random digit telephone samples: (1) identified all
working telephone exchanges and working blocks (the first two digits after the exchange); (2) assigned
each exchange to a single county; (3) stratified the sampling frame by exchange, and within exchange
by working block; and (4) systematically selected the sample for the geographic area specified.
Response Rates
A total sample frame of 5,815 telephone numbers were used in this study. This resulted in:
1,910 Completed interviews
2,313 Non-sample (including not eligible, not working numbers, business numbers, etc.)
1,592 Non-response
1,038 Refusals
145 Away for duration of study
156 Contact respondent not available
5 Other non-response
248 No answer
The response rate was adjusted to compensate for the never answered numbers. The never-answered
6
numbers consist of residential numbers that were never answered in any of twenty calls that were
made. We assumed that the ratio of working residential numbers to other numbers in this subset is the
same as for numbers that were answered (see Table 1 for specific calculations).
9
FINDINGS
Data Sets
The findings are reported by comparing combined survey samples and subsamples of the
previously discussed survey samples. When analyzing national trends, all survey samples were
combined; that data set will hereafter be referred to as the DGZ (Douthitt, Grobe, Zepeda) sample. It
consists of the 1,910 completed surveys, including the National, Wisconsin, Vermont, and Poor (10th
and 20th percentile) samples. The DGZ sample results are weighted to ensure findings are
representative of the U.S. population. To analyze poor consumers behavior, another data set was
constructed, the Base-Level Poverty (BLP) sample. The BLP is a subsample of the 1,910 completed
surveys. It includes all the respondents who qualified as poor under the February 9, 1995, USDA
poverty guidelines. The 217 respondents who met the USDA poverty guideline requirements make up
7
the BLP sample.
Calculation of Weights for the DGZ Sample. The DGZ sample contained five separate samples
that were combined by appropriate weighting: (1) a National random sample, (2) a Wisconsin random
sample, (3) a Vermont random sample, (4) a sample drawn from low-income areas (20th percentile),
and (5) a second sample drawn from low-income areas where average income is lower than in the first
(10th percentile). Because the National and Poor samples were independent (that is, they were not
mutually exclusive and exhaustive subsets of the total population), we assume that the low-income
households in each of the three samples (National, 20th percentile, 10th percentile) represent low-
income persons in general.
To combine the Poor samples with the random National sample, a weight was computed such
that the income distribution of the random National sample was maintained (that is, depending on the
reported income, Poor area cases were added to deflate or inflate the weight of all cases). Household
income was divided into 13 categories and weights were computed.
10
To add the two state samples (Wisconsin and Vermont) to the National and Poor samples,
weights were computed that make the Vermont oversample represent .0023 of the total sample and the
Wisconsin cases (from both the state and the National samples) represent .0197 of the total sample.
8
According to recent Census Bureau state population estimates, Wisconsin includes 1.97 percent of U.S.
households; Vermont includes 0.23 percent. The sum of the weights is then adjusted to 1,910, the total
number of sample cases.
Demographic Characteristics
DGZ Sample. We compared the DGZ sample to U.S. Census household demographic
characteristics to assess whether the DGZ sample was representative of U.S. households. Household
data were taken from the 1990 Census of Population, General Population Characteristics of the United
States. Table 2 summarizes these comparisons. Almost 72 percent of the respondents were women,
reflecting the screening question for primary household food purchaser. The DGZ sample was more
educated than U.S. households, with mean age of householder a few years younger than U.S.
households. Close to 60 percent of the DGZ respondents were married, with an average household size
of 2.87; for U.S. households, the respective figures are 54.9 percent and 2.63. Median income was
higher than the U.S. households figure. Eighty percent of the DGZ respondents were Caucasian,
consistent with U.S. households, with approximate representation of other ethnic groups.
Base-Level Poverty Sample. Table 3 summarizes the BLP sample demographic characteristics
compared to the U.S. poverty households. U.S. poverty household data were taken from U.S.
Department of Housing and Urban Development and the Census Bureau (Grall 1994). Almost 80
percent of the primary household food purchasers in the BLP sample were female. The BLP sample
11
TABLE 2
Comparison of U.S. Household Demographics and DGZ Sample Demographics
DGZ Sample
U.S. Households (N=1910)
Education (1993)
<= 8 years 9.3% 4.0%
13 years of high school 10.5% 6.2%
High school graduate 32.0% 21.2%
13 years of college 17.7% 23.5%
Associate 6.5% 15.2%
Bachelors 15.0% 18.9%
Masters 5.6% 7.9%
Ph.D. 1.3% 1.2%
Professional 1.7% 1.9%
Marital Status (1991)
Married 54.9% 62.9%
Widowed 12.4% 8.5%
Divorced 13.2% 9.6%
Living with a partner 4.0%
Separated 4.0% 1.3%
Never married 15.3% 13.2%
Household Size (1990)
Average per household 2.63 2.87
Income (1993)
Median Income $31,241 $40,000
Ethnic Background (1990)
Caucasian 80.0% 81.2%
African American 10.8% 10.4%
Native American 0.6% 1.0%
Asian 2.1% 1.5%
Hispanic 6.5% 3.6%
a
other 2.5%
Age (1993)
Mean age of adult householder 48.2 45.6
Note: Parenthetical dates in the first column refer to the year of the most recent census data available
for that category.
For U.S. households, all races (Caucasian, African American, Native American, Asian) were asked
a
whether they were of Hispanic origin, while for the DGZ sample only those who indicated they were
Caucasian were also asked if they were of Hispanic origin (Mexican American, Latin American, Puerto
Rican, or Cuban).
12
TABLE 3
Comparison of U.S. Poverty Household Demographics
and Base-Level Poverty Sample Demographics
U.S. Poverty Households Base-Level Poverty Sample
(1991) (N=217)
Education
<= 8 years 22.6% 20.4%
13 years of high school 21.4% 22.7%
High school graduate 34.9% 22.2%
13 years of college 12.9% 17.1%
College degree or more 7.9% 17.6%
Age
Median age of adult householder 45.3 46
Household Size
Median household size 2.1 2
Income
Median income $5,581 $9,800
Ethnic Background
Caucasian 59.7% 65.1%
African American 25.2% 23.3%
Other Ethnic 4.8% 5.6%
Hispanic 10.1% 6.0%
a
For U.S. households, all races (Caucasian, African American, Native American, Asian) were asked
a
whether they were of Hispanic origin, while for the DGZ sample only those who indicated they were
Caucasian were also asked if they were of Hispanic origin (Mexican American, Latin American, Puerto
Rican, or Cuban).
13
was more educated than U.S. poverty households, with median age of householder one year older than
the U.S. poverty households. The median household size was 2.0, consistent with the median U.S.
poverty household size of 2.1. Median income was higher than that of U.S. poverty households. Sixty-
five percent of the BLP sample were Caucasian, approximating with the U.S. poverty households, with
consistent representation of other ethnic groups.
Consumer Awareness, Knowledge, and Approval of Biotechnology, rpGH, and rbGH
Awareness. Respondents were asked about their awareness of biotechnology, recombinant
porcine growth hormone (rpGH), and recombinant bovine growth hormone (rbGH) (see Table 4). They
were asked whether they had heard of or read anything about each particular technology (see Figure 1
for question wording). Awareness of biotechnology and rbGH was high: approximately 60 percent of
the DGZ sample were aware of biotechnology, and 66 percent were aware of rbGH. DGZ respondents
were less familiar with rpGH: approximately 24 percent had either heard of or read anything about the
use of rpGH.
The BLP sample respondents had significantly lower awareness levels regarding biotechnology
(approximately 35 percent were aware) ( = 50.26, p value .001) and rbGH (54 percent were
2 9
1 df
aware) ( = 17.96, p value = .001), but had comparable awareness of rpGH (24 percent). The above
2
1 df
results were similar for the total respondents of both samples, as well as among those expressing an
opinion. Respondents who were not aware of rbGH or rpGH were read a brief information statement
before proceeding to the next question (see Figure 1).
Knowledge. In addition to awareness of rbGH, respondents were read six different statements
to acquire data on their knowledge of rbGH. They were asked to what extent they agreed or disagreed
with each statement. Table 5 indicates that, among the entire DGZ sample, 60 percent disagreed with
the statement that rbGH-treated herd milk is the same as untreated herd milk; 73 percent among
14
TABLE 4
Awareness of Biotechnology, rbGH, and rpGH
DGZ Sample BLP Sample
(N=1910) (N=217)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
Awareness of biotechnology**
Yes 56.7% 56.9% 35.0% 35.5%
No 43.0% 43.1% 63.6% 64.5%
Dont know 0.4% 1.4%
Awareness of rbGH**
Yes 63.7% 64.1% 53.9% 54.7%
No 35.7% 35.9% 44.7% 45.3%
Dont know 0.7% 1.4%
Awareness of rpGH
Yes 23.9% 24.0% 24.0% 24.3%
No 75.6% 76.0% 74.7% 75.7%
Dont know 0.5% 1.4%
Note: ** p value < .01; chi-square analysis comparing DGZ sample and BLP sample respondents who
expressed an opinion.
15
Figure 1
Wording of Question Items
Awareness
Biotechnology
As you may know, biotechnology refers to the use of technology to create new plant or animal
species, or to create chemicals. In agriculture, biotechnology has been used to create new
disease-resistant plants and to economically produce chemicals to increase farm production.
Have you heard or read anything about agricultural use of biotechnology?
rpGH
Have you heard or read anything about the use of rpGH, a synthetic pork hormone used to
stimulate the growth of hogs to produce leaner pork?
rbGH
Do you recall having heard or read anything about the use of a synthetic bovine growth
hormone, commonly called bGH or bST, that is used by farmers to increase cows milk
production?
Information
rpGH
Porcine somatotropin, or rpGH, is a growth hormone, which when injected stimulates the
growth rate of hogs. Its use causes reduced fat deposit and hence, leaner pork.
rbGH
Bovine somatotropin (rbGH) is a growth hormone, which when administered/injected in fully
grown, lactating cows, increases their milk production, thereby improving dairy farm profits.
The milk from cows given rbGH has the same product characteristics as the milk from
untreated cows.
(figure continues)
16
Figure 1 (continued)
Health Risk
Current rpGH
Although scientists have not discovered any ill health effects for humans from eating pork
treated with rpGH, some consumers believe that the meat from treated pigs is not safe. How
about you, do you share such concerns, or not?
Would you say you are concerned only a little, moderately concerned, or very concerned?
Current rbGH
Although the U.S. Food and Drug Administration has stated that there are no ill health effects
for humans associated with consuming dairy products made of milk from rbGH- treated cows,
some consumers believe that rbGH-treated cows milk is not safe. How about you, do you share
such concerns, or not?
Would you say you are concerned only a little, moderately concerned, or very concerned?
Future rbGH
Some consumers are further concerned that although there are currently no known human ill
health effects associated with consuming milk from rbGH-treated cows, that ill health effects
may later be discovered. How about you, do you share such concerns, or not?
Would you say you are concerned only a little, moderately concerned, or very concerned?
Demand Response
Since February of 1994, it has been legal for farmers to increase their cows milk production by
administering/injecting them with rbGH. How, if at all, has the approval of rbGHs use
influenced the amount of milk you buy?
17
TABLE 5
Knowledge of rbGH
DGZ Sample BLP Sample
(N=1910) (N=217)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
rbGH-treated herd milk is
the same as untreated herd
milk*
Disagree strongly 38.2% 46.8% 39.6% 48.6%
Disagree somewhat 21.6% 26.5% 19.8% 24.3%
Neither agree nor disagree 9.9% 12.2% 8.8% 10.7%
Agree somewhat 8.5% 10.4% 6.9% 8.5%
Agree strongly 3.4% 4.1% 6.5% 7.9%
Dont know 18.4% 18.4%
The long-run health
implications of rbGH are
not known**
Disagree strongly 3.5% 4.2% 6.0% 7.7%
Disagree somewhat 7.0% 8.3% 13.0% 16.7%
Neither agree nor disagree 8.4% 9.9% 7.4% 9.5%
Agree somewhat 19.8% 23.5% 17.1% 22.0%
Agree strongly 45.5% 54.1% 34.3% 44.0%
Dont know 15.8% 22.2%
Increasing milk production
by farmers using rbGH has
benefitted consumers*
Disagree strongly 27.3% 32.9% 27.3% 33.9%
Disagree somewhat 20.5% 24.8% 20.8% 25.9%
Neither agree nor disagree 12.6% 15.2% 8.3% 10.3%
Agree somewhat 17.4% 21.0% 14.8% 18.4%
Agree strongly 5.1% 6.2% 9.3% 11.5%
Dont know 17.2% 19.4%
(table continues)
18
TABLE 5, continued
DGZ Sample BLP Sample
(N=1910) (N=217)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
bGH is naturally found
in milk*
Disagree strongly 25.9% 38.7% 22.1% 32.7%
Disagree somewhat 16.1% 24.0% 18.9% 27.9%
Neither agree nor disagree 13.6% 20.3% 10.1% 15.0%
Agree somewhat 8.4% 12.5% 12.9% 19.0%
Agree strongly 3.0% 4.5% 3.7% 5.4%
Dont know 33.0% 32.3%
rbGH use is not harmful
to cows**
Disagree strongly 23.8% 33.0% 32.3% 46.7%
Disagree somewhat 20.7% 28.8% 14.7% 21.3%
Neither agree nor disagree 14.4% 20.0% 9.7% 14.0%
Agree somewhat 10.7% 14.8% 10.6% 15.3%
Agree strongly 2.5% 3.4% 1.8% 2.7%
Dont know 28.0% 30.9%
rbGH use has had negative
effects on small dairy farms
Disagree strongly 7.0% 10.2% 6.9% 10.2%
Disagree somewhat 11.4% 16.7% 14.3% 21.1%
Neither agree nor disagree 14.8% 21.7% 11.1% 16.3%
Agree somewhat 20.9% 30.6% 20.7% 30.6%
Agree strongly 14.2% 20.8% 14.7% 21.8%
Dont know 31.7% 32.3%
* p value < .05; ** p value < .01; chi-square analysis comparing DGZ sample and BLP sample
respondents who expressed an opinion.
19
those expressing an opinion disagreed. Two-thirds of the DGZ respondents agreed with the statement
that the long run health implications of rbGH are not known, and almost half disagreed that
increasing milk production by farmers using rbGH has benefitted consumers. Among those
expressing an opinion, 77 percent agreed that we do not know rbGHs long-run health implications, and
27 percent agreed that rbGHs use has benefitted consumers. These differences between all
respondents and those expressing an opinion are reflected in the dont know responses, which ranged
from 15 percent to 18 percent for the first three statements. The respondents were even more unsure of
the final three statements, as dont know responses escalated to around 30 percent. The majority of
the DGZ sample who did express an opinion disagreed with the statements bGH is naturally found in
milk, and treating cows with rbGH is not harmful to them, while tending to agree more than
disagree that rbGH use has had negative economic effects on small dairy farms. BLP sample
respondents who expressed an opinion had statistically different results for knowledge of rbGH than the
DGZ sample respondents (see Table 5 for results). However, among those who expressed an
2
opinion, BLP respondents were more likely than DGZ respondents to express disagreement with the
statement the long-run health implications of rbGH are not known.
Approval. To get an idea of consumers overall approval of biotechnology and rbGH, we asked,
Overall, do you or do you not approve of agricultural uses of biotechnology? and Overall, what is
your opinion about treating cows with rbGH to increase their milk production? The results in Table 6
show varying approval opinions between biotechnology and rbGH. Over half of the total DGZ
respondents approved of biotechnology uses, whereas over half of the total DGZ respondents felt the
use of rbGH was a poor idea, with 30 percent feeling it was a fair idea. Of those DGZ respondents who
expressed an opinion, 60 percent approved of biotechnology uses, and the same percentage felt rbGH
was a poor idea. These results are somewhat different, but not significantly different for the BLP
respondents, of whom 45 percent approved of biotechnology uses and 60 percent felt rbGH was a poor
20
TABLE 6
Overall Approval of Biotechnology and rbGH
DGZ Sample BLP Sample
(N=1910) (N=217)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
Overall approval of biotechnology
Approve 53.8% 59.9% 44.7% 50.7%
Disapprove 36.0% 40.1% 43.4% 49.3%
Dont know 10.1% 11.8%
Overall opinion of rbGH
Poor idea 53.8% 58.3% 59.4% 66.5%
Fair idea 27.1% 29.4% 19.4% 21.6%
Good idea 8.9% 9.7% 7.8% 8.8%
Excellent idea 2.5% 2.7% 2.8% 3.1%
Dont know 7.7% 10.6%
21
idea, with an additional 20 percent rating rbGH as a fair idea. Fifty percent of the BLP respondents
expressing an opinion approved of biotechnology uses, and 66 percent felt rbGH was a poor idea.
Overall opinions toward rpGH were not asked because the product is not currently approved for
commercial use. However, the respondents were asked if they would consider purchasing rpGH-treated
pork if it were approved by the FDA. Table 7 shows that 50 percent of the total DGZ and BLP samples
would purchase pork produced with rpGH, and approximately 60 percent of those expressing an
opinion in the two samples would purchase pork produced with rpGH if approved by the FDA.
Consumers Perceived Concern for Discovery of Human Ill Health Effects
Previous research has explored the long-run risk perceptions of rbGH; that is, whether
consumers were concerned that human health risks would be discovered in the future (Grobe and
Douthitt 1995). However, additional work was needed to evaluate current consumer risk perception.
Consumers may or may not have as great a concern for risks in the short run as they appear to have in
the long run. This survey evaluated both current concern for health risks and future discovery of health
risks associated with rbGH, and only the current concern of health risks associated with rpGH, given
rpGHs stage of product development (see Figure 1 for question wording).
Current Health Risks. Table 8 presents the results where over 75 percent of the respondents
from both samples acknowledged some level of concern for current human ill health effects associated
with rpGH and rbGH. Two-thirds of those respondents were moderately to very concerned, while
approximately 20 percent had no concern about current health risks from either of these
biotechnologies. These results were similar regardless of whether all respondents or only respondents
expressing an opinion (excluding dont knows) were considered. The BLP sample respondents were
equally concerned with rpGH, and those expressing an opinion were significantly more likely ( =
2
3 df
22
TABLE 7
Respondents Willingness to Purchase rpGH-Treated Pork If Approved by the FDA
DGZ Sample BLP Sample
(N=1910) (N=217)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
Purchase rpGH-treated
pork if approved by FDA
Yes 50.1% 56.4% 49.7% 59.3%
No 38.8% 43.6% 34.1% 40.7%
Dont know 11.1% 16.2%
23
TABLE 8
Concerns about Current Human Ill Health Effects Associated with rpGH and rbGH
DGZ Sample BLP Sample
(N=1910) (N=217)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
Current health concern
rpGH
No concern 23.8% 23.9% 22.9% 22.9%
Concerned a little 9.8% 9.8% 10.9% 10.9%
Moderately concerned 29.7% 29.8% 24.5% 24.5%
Very concerned 36.5% 36.6% 41.7% 41.7%
Dont know 0.2%
rbGH**
No concern 21.4% 22.5% 16.6% 18.0%
Concerned a little 8.4% 8.9% 7.8% 8.5%
Moderately concerned 30.1% 31.7% 19.8% 21.5%
Very concerned 35.0% 36.9% 47.9% 52.0%
Dont know 5.0% 7.8%
* p value < .05; ** p value < .01; chi-square analysis comparing DGZ sample and BLP sample
respondents who expressed an opinion.
24
16.78, p value = .001) to be concerned about the current safety of rbGH with respect to human ill health
effects (82 percent), compared to the DGZ respondents.
Future Health Risks. Respondents had greater concern for the discovery of human ill health
effects associated with rbGH than with current human health effects (see Table 9). Eighty-five percent
of all the DGZ respondents and of those expressing an opinion acknowledged some level of concern,
with over 70 percent indicating their level of concern as moderate or very concerned. The BLP
respondents expressing an opinion were significantly more likely ( = 12.6, p value = .006) to be
2
3 df
concerned about the future discovery of human ill health effects than the DGZ respondents.
Risk Perception Factors
In order to understand the differences in risk perception, we compared the respondents who
were concerned to those who were not concerned about the future discovery of health risks for both
samples (DGZ and BLP) (Tables 1014). Those respondents who acknowledged some level of concern
about the future discovery of current health risks were compared with the respondents with no concern.
Some of the variables believed to influence risk perceptions are beliefs about milk, trustworthiness of
information sources, knowledge of rbGH and non-rbGH treated herd milk, locus of control, and group
affiliation. We discuss our findings for each variable for the DGZ and the BLP respondents.
Beliefs about Milk. Researchers have suggested that one must consider the central role of
beliefs in shaping how consumers perceive risk (Groth 1990). Thus, this survey explored respondents
beliefs about milk. They were asked to state the extent to which they agreed or disagreed with four
statements. The majority of both groups (no concern and concerned) for both samples agreed with the
first two statements, children must have milk for proper growth and development and milk is
nutritious (see Table 10). When asked whether they agreed or disagreed with the statement that milk
is a natural product, significantly more of the not concerned group for both samples (around 80
25
TABLE 9
Concern with the Future Discovery of Human Ill Health Effects Associated with rbGH
DGZ Sample BLP Sample
(N=1910) (N=217)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
Future health concern
rbGH**
No concern 12.2% 12.5% 14.7% 15.7%
Concerned a little 11.0% 11.3% 12.9% 13.7%
Moderately concerned 31.9% 32.7% 19.4% 20.6%
Very concerned 42.3% 43.5% 47.0% 50.0%
Dont know 2.6% 6.0%
* p value < .05; ** p value < .01; chi-square analysis comparing DGZ sample and BLP sample
respondents who expressed an opinion.
26
TABLE 10
Respondent Milk Beliefs by Those Who Are Not Concerned and Those Who Are
Concerned with Future Discovery of Health Risks Associated with rbGH
DGZ Sample BLP Sample
No Concern Concerned No Concern Concerned
(N=233) (N=1628) (N=32) (N=172)
Children must have milk
Disagree strongly 1.2% 4.2% 3.1% 3.5%
Disagree somewhat 6.0% 9.1% 3.1% 4.1%
Neither agree nor disagree 2.8% 2.6% 2.9%
Agree somewhat 18.6% 22.9% 28.1% 17.4%
Agree strongly 71.4% 61.2% 65.6% 72.1%
Milk is nutritious
Disagree strongly 1.9% 1.5% 3.1% 1.2%
Disagree somewhat 2.3% 5.6% 6.0%
Neither agree nor disagree 2.4% 2.8% 1.8%
Agree somewhat 30.0% 34.5% 43.8% 29.2%
Agree strongly 63.3% 55.5% 53.1% 61.9%
Milk is a natural product**
Disagree strongly 2.9% 8.5% 3.1% 16.6%
Disagree somewhat 11.5% 18.9% 15.6% 17.2%
Neither agree nor disagree 3.0% 3.6% 3.1% 2.4%
Agree somewhat 38.4% 37.7% 37.5% 33.1%
Agree strongly 44.2% 31.3% 40.6% 30.8%
Milk is a pure product**
Disagree strongly 5.8% 13.7% 9.7% 16.8%
Disagree somewhat 17.7% 29.0% 9.7% 19.8%
Neither agree nor disagree 11.2% 5.5% 6.5% 6.6%
Agree somewhat 40.5% 34.1% 38.7% 30.5%
Agree strongly 24.8% 17.7% 35.5% 26.3%
* p value < .05; ** p value < .01; chi-square analysis comparing respondents who are concerned and
those who are not concerned with future discovery of health effects. Analysis comparing the DGZ
sample and the BLP sample was not performed.
27
percent) agreed with the statement, compared with the concerned groups (approximately 66 percent)
( = 24.8, p value = .0001). Approximately half of the concerned respondents from both samples
2
4 df
agreed with the final statement, milk is a pure product, and less than half (40 percent) disagreed with
the statement, significantly different from the no concern respondents ( = 34.17, p value = .0001).
2
4 df
More than 65 percent of the no concern respondents from both samples agreed with the statement.
Trustworthiness of Information Sources. Respondents were asked about the trustworthiness of
various food-related information sources (see Table 11). In both samples, the group with no health risk
concerns tended to believe in the trustworthiness of the FDA, significantly more so than the concerned
groups, who rated the FDA as only somewhat trustworthy ( = 32.66, p value= .0001). Similar
2
3 df
results were also found on the USDAs trustworthiness as a food-related information source. Each
group in both samples rated farmers organizations, nutritional information labels, and journalists as
somewhat trustworthy. Both groups, especially in the DGZ sample, rated physicians as being very
trustworthy. Advertisements received the least trustworthy rating: approximately 60 percent of both
groups in the DGZ sample stated they were not very to not at all trustworthy. Of the BLP sample, 40
percent of the no concern and 50 percent of the concerned rated advertisements as not very to not at all
trustworthy. Furthermore, food and drug businesses were rated by approximately 60 percent of each
group in both samples as somewhat trustworthy, with almost a quarter regarding food and drug business
as not very to not at all trustworthy.
A comparison of DGZ and BLP respondents rankings of food-related information sources is
found in Table 12. Physicians were ranked the highest by the DGZ respondents, as 89.8 percent of the
respondents felt physicians were a somewhat to very trustworthy source of information. The DGZ
respondents ranked the USDA as second, followed by the FDA. Conversely, the BLP sample ranked
28
TABLE 11
Respondent Trustworthiness in Information Sources by Those Who Are Not Concerned and
Those Who Are Concerned with the Future Discovery of Health Risks Associated with rbGH
DGZ Sample BLP Sample
No Concern Concerned No Concern Concerned
(N=233) (N=1628) (N=32) (N=172)
FDA**
Very trustworthy 37.9% 21.7% 36.0% 20.7%
Somewhat trustworthy 54.5% 66.0% 52.0% 62.0%
Not very trustworthy 5.5% 9.6% 12.0% 11.3%
Not at all trustworthy 2.1% 2.7% 6.0%
Farmers Organizations
Very trustworthy 16.9% 20.1% 19.2% 25.8%
Somewhat trustworthy 71.4% 63.9% 65.4% 58.9%
Not very trustworthy 10.3% 13.8% 7.7% 10.5%
Not at all trustworthy 1.4% 2.3% 7.7% 4.8%
Physicians*
Very trustworthy 45.7% 41.2% 32.1% 39.8%
Somewhat trustworthy 48.4% 47.8% 60.7% 39.8%
Not very trustworthy 5.4% 8.4% 3.6% 13.7%
Not at all trustworthy 0.6% 2.6% 3.6% 6.8%
Advertisements
Very trustworthy 5.5% 3.2% 7.4% 9.8%
Somewhat trustworthy 36.0% 36.1% 51.9% 36.6%
Not very trustworthy 41.5% 39.2% 22.2% 26.8%
Not at all trustworthy 17.0% 21.5% 18.5% 26.8%
Nutrition Information Labels
Very trustworthy 35.3% 31.7% 37.9% 33.1%
Somewhat trustworthy 52.4% 55.1% 51.7% 46.5%
Not very trustworthy 8.3% 11.1% 3.4% 15.9%
Not at all trustworthy 4.0% 2.1% 6.9% 4.5%
(table continues)
29
TABLE 11, continued
DGZ Sample BLP Sample
No Concern Concerned No Concern Concerned
(N=233) (N=1628) (N=32) (N=172)
Journalists*
Very trustworthy 10.2% 10.1% 3.6% 8.1%
Somewhat trustworthy 50.3% 61.1% 57.1% 56.1%
Not very trustworthy 29.3% 20.6% 14.3% 26.4%
Not at all trustworthy 10.2% 8.2% 25.0% 9.5%
Food and Drug Business*
Very trustworthy 11.0% 8.1% 10.3% 16.8%
Somewhat trustworthy 63.4% 61.1% 72.4% 57.7%
Not very trustworthy 21.9% 21.4% 10.3% 16.1%
Not at all trustworthy 3.7% 9.4% 6.9% 9.4%
USDA
Very trustworthy 32.4% 25.2% 34.5% 34.4%
Somewhat trustworthy 57.7% 63.3% 51.7% 54.5%
Not very trustworthy 7.7% 8.4% 6.9% 6.5%
Not at all trustworthy 2.2% 3.1% 6.9% 4.5%
* p value < .05; ** p value < .01; chi-square analysis comparing respondents who are concerned and
those who are not concerned with future discovery of health effects. Analysis comparing the DGZ
sample and the BLP sample was not performed.
30
TABLE 12
Respondents Ranking of the Trustworthiness of Food-Related Information Sources
DGZ Sample BLP Sample
(N=1910) (N=217)
Physicians 89.8% USDA 88.0%
USDA 88.7% Farmers Organizations 84.3%
FDA 88.4% FDA 84.0%
Nutrition Information Labels 87.0% Physicians 81.9%
Farmers Organizations 84.4% Nutrition Information Labels 81.0%
Journalists 69.8% Food & Drug Business 75.6%
Food & Drug Business 69.7% Journalists 64.3%
Advertisements 39.8% Advertisements 49.0%
Note: Percentages represent the respondents who indicated that a particular food-related information
source was somewhat to very trustworthy.
31
the USDA as first, followed by farmers organizations and the FDA. Both samples ranked
advertisements as the least trustworthy food-related information source.
Knowledge of rbGH. To measure consumer knowledge about rbGH, this survey asked
respondents to what extent they agreed or disagreed with six statements. The overall results in Table 13
show that, for both samples, the concerned group has significantly stronger opinions (that is, agree
strongly or disagree strongly) toward the statements than those who are not concerned. Additionally,
dont know responses were high, especially for the last three statements.
Almost 65 percent of the concerned group in both samples disagreed with the statement that
milk from cows treated with rbGH is just like milk from untreated cows. Thirty-two percent of the
not concerned DGZ sample and 43 percent of the BLP sample disagreed, and 38 percent and 22 percent
agreed with the statement, respectively. A greater percentage of the concerned group in both samples
disagreed that rbGH benefits consumers. Furthermore, 70 percent of the DGZ concerned group and 57
percent of the BLP concerned group agreed with the statement, the long-run health implications of
rbGH are not known, compared to 44 percent of the DGZ and 32 percent of the BLP no concern
group.
Nearly half of the concerned respondents in both samples disagreed that bGH is naturally
found in milk, and approximately 30 percent did not know. Similar results were found for the
statement, treating cows with rbGH is not harmful to them. Fifty percent of the concerned groups
disagreed, while about 25 percent of the no concern groups disagreed. Finally, results were
approximately equally divided between the agree and disagree responses for both groups and samples
concerning the statement, rbGH use has had negative economic effects on small dairy farms.
Locus of Control. Locus of control refers to a persons sense of control over life events.
Questions to ascertain respondents locus of control were (a) I worry about the future that todays
children are facing, and (b) More and more, I feel helpless in the face of whats happening in the
32
TABLE 13
Respondent Knowledge of rbGH by Those Who Are Not Concerned and Those Who Are
Concerned with the Future Discovery of Health Risks Associated with rbGH
DGZ Sample BLP Sample
No Concern Concerned No Concern Concerned
(N=233) (N=1628) (N=32) (N=172)
Milk from rbGH-treated
herds is the same as from
untreated herds**
Disagree strongly 12.4% 42.8% 12.5% 45.9%
Disagree somewhat 19.9% 22.2% 31.3% 18.6%
Neither agree nor disagree 11.7% 9.9% 6.3% 9.9%
Agree somewhat 22.1% 6.5% 9.4% 5.8%
Agree strongly 15.8% 1.7% 12.5% 5.8%
Dont know 18.0% 16.8% 28.1% 14.0%
Dont know rbGHs long-run
health implications**
Disagree strongly 5.3% 3.3% 6.5% 5.8%
Disagree somewhat 17.8% 5.6% 25.8% 11.0%
Neither agree nor disagree 14.9% 7.5% 9.7% 7.6%
Agree somewhat 25.0% 19.2% 25.8% 15.1%
Agree strongly 19.3% 50.5% 6.5% 41.9%
Dont know 17.7% 13.9% 25.8% 18.6%
rbGHs use has benefitted
consumers**
Disagree strongly 11.4% 30.1% 6.5% 32.0%
Disagree somewhat 14.7% 21.9% 32.3% 19.8%
Neither agree nor disagree 17.0% 12.2% 12.9% 8.1%
Agree somewhat 34.4% 15.3% 16.1% 14.5%
Agree strongly 11.0% 4.4% 6.5% 10.5%
Dont know 11.4% 16.2% 25.8% 15.1%
(table continues)
33
TABLE 13, continued
DGZ Sample BLP Sample
No Concern Concerned No Concern Concerned
(N=233) (N=1628) (N=32) (N=172)
bGH is naturally found in milk**
Disagree strongly 15.3% 28.0% 9.4% 25.6%
Disagree somewhat 15.4% 16.4% 34.4% 16.3%
Neither agree nor disagree 15.0% 13.8% 6.3% 11.6%
Agree somewhat 13.2% 7.9% 9.4% 13.4%
Agree strongly 6.6% 2.4% 3.1% 4.1%
Dont know 34.5% 31.5% 37.5% 29.1%
rbGH use is not harmful to cows**
Disagree strongly 9.4% 26.3% 15.6% 36.6%
Disagree somewhat 14.7% 22.2% 15.6% 15.7%
Neither agree nor disagree 16.0% 14.4% 9.4% 10.5%
Agree somewhat 24.1% 8.9% 18.8% 9.3%
Agree strongly 7.8% 1.8% 3.1% 1.7%
Dont know 28.1% 26.3% 37.5% 26.2%
rbGH use has had negative effects
on small dairy farms**
Disagree strongly 9.3% 6.8% 6.3% 7.0%
Disagree somewhat 16.2% 11.0% 15.6% 15.1%
Neither agree nor disagree 20.6% 14.4% 12.5% 11.6%
Agree somewhat 23.1% 21.1% 25.0% 20.9%
Agree strongly 6.4% 15.5% 6.3% 16.9%
Dont know 24.4% 31.1% 34.4% 28.5%
* p value < .05; ** p value < .01; chi-square analysis comparing respondents who are concerned and
those who are not concerned with future discovery of health effects. Analysis comparing the DGZ
sample and the BLP sample was not performed.
34
world today. Interestingly, although the majority of both groups for the two samples felt less in
control, the respondents with concerns about the safety of rbGH had significantly stronger agreement
with both statements ( = 104.2, p value = .0001; = 52.541, p value = .0001, respectively),
2 2
3 df 3df
especially when asked about feeling helpless in the world today (see Table 14). Nearly all (97.3
percent) concerned respondents from both samples agreed with the first statement, and 80 percent
agreed with the second. For the not concerned group in each sample, about 86 percent and 65 percent
agreed with the first and second locus-of-control questions, respectively.
Group Affiliation. To gain insight into the relationship between technology acceptance and the
respondents group affiliation, questions were asked to determine whether or not the respondent
identified with the following groups: (1) environmentalists, those who wish to protect our natural
resources, (2) religious groups, who believe in a strict interpretation of the Bible, and (3) animal
rights groups, those who oppose using animals in experimental studies. A greater percentage of the
concerned groups for each sample were significantly more likely to identify with environmentalists and
animal rights groups ( = 56.0, p value = .0001; = 51.2, p value = .0001, respectively) (see
2 2
2 df 2df
Table 15). The concerned BLP respondents had a larger percentage of strong identification with all
three groups compared to the concerned DGZ respondents. Religious affiliation was unrelated to safety
concerns.
Consumers Demand Response
An essential question of interest to the study was How, if at all, has the approval of rbGHs
use influenced the amount of milk you buy? The respondents demand response was either (a)
increased the amount of milk they purchase, (b) no change in their milk purchase, (c) decreased the
amount of milk they purchase, or (d) stopped buying milk altogether. The majority (approximately 90
percent) of the total DGZ and BLP respondents stated they made no change in the amount of milk they
35
TABLE 14
Respondent Locus of Control by Those Who Are Not Concerned and Those Who Are Concerned
with Future Discovery of Health Risks Associated with rbGH
DGZ Sample BLP Sample
No Concern Concerned No Concern Concerned
(N=233) (N=1628) (N=32) (N=172)
Worry about future for kids**
Strongly agree 51.7% 75.4% 48.4% 75.6%
Agree 33.3% 21.5% 38.7% 23.8%
Disagree 10.2% 2.2% 9.7% 0.6%
Strongly disagree 4.8% 0.8% 3.2%
Feel helpless in the world**
Strongly agree 32.6% 48.2% 38.7% 57.4%
Agree 28.2% 31.6% 29.0% 27.2%
Disagree 29.0% 16.2% 22.6% 10.1%
Strongly disagree 10.2% 4.0% 9.7% 5.3%
* p value < .05; ** p value < .01; chi-square analysis comparing respondents who are concerned and
those who are not concerned with future discovery of health effects. Analysis comparing the DGZ
sample and the BLP sample was not performed.
36
TABLE 15
Respondent Group Affiliation by Those Who Are Not Concerned and Those Who Are Concerned
with Future Discovery of Health Risks Associated with rbGH
DGZ Sample BLP Sample
No Concern Concerned No Concern Concerned
(N=233) (N=1628) (N=32) (N=172)
Identify with environmentalists**
Not at all 19.0% 6.3% 22.6% 7.1%
Somewhat 58.4% 55.6% 38.7% 39.6%
Strongly 22.6% 38.1% 38.7% 53.3%
Identify with religious groups
Not at all 36.4% 36.2% 25.8% 26.2%
Somewhat 39.2% 41.4% 35.5% 37.8%
Strongly 24.4% 22.4% 38.7% 36.0%
Identify with animal rights groups**
Not at all 49.3% 28.0% 44.8% 28.4%
Somewhat 40.3% 46.3% 41.4% 35.2%
Strongly 10.4% 25.6% 13.8% 36.4%
* p value < .05; ** p value < .01; chi-square analysis comparing respondents who are concerned and
those who are not concerned with future discovery of health effects. Analysis comparing the DGZ
sample and the BLP sample was not performed.
37
bought since the FDA approved rbGH (see Table 16). However, of the respondents who have not
changed the amount of milk they buy, 8 percent of the DGZ respondents and 13.7 percent of the BLP
respondents usually purchased milk identified as coming from nontreated cows.
For the DGZ respondents who usually purchase milk from nontreated cows, 52 percent know it
was milk from untreated herds because it was a store policy, 60.7 percent because it was labeled
untreated, 65.9 percent because it was a brand policy, and 26 percent because it was labeled organic
(see Table 17). For the BLP respondents, 65.5 percent knew because it was a store policy, 58.6 percent
because it was labeled untreated, 65.5 percent because it was a brand policy, and 34.5 percent because
it was labeled organic. For this study, respondents who stated they were buying milk identified as
coming from nontreated herds made up 10.6 percent total fluid milk consumption.
Five percent of the total DGZ respondents decreased the amount of milk they bought, and 0.3
percent stopped buying milk altogether. Of the DGZ respondents who have decreased or stopped
buying milk, 50 percent have substituted other products for milk. Similar results were found for the
BLP respondents. Five percent decreased their consumption, and 0.5 percent stopped buying milk. Of
these respondents, 39 percent have substituted other products for cows milk.
Another question of interest was whether or not concern for future discovery of health effects
influenced demand response. Table 18 shows that the concerned group for both samples were more
likely to decrease or stop buying milk, compared to the no concern group. The results also show that a
smaller percentage of the concerned group for both samples indicated no change in their demand
response.
Consumers Preference for Treated versus Untreated Milk
A series of questions were asked to see if consumers have a preference for milk from an
untreated or treated herd. First, respondents were asked, Do you have a preference for whether the
38
TABLE 16
Influence of rbGHs Approval on Consumers Demand Response
DGZ Sample BLP Sample
(N=1910) (N=217)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
Demand Response
Increase 0.6% 0.6% 0.9% 1.0%
No change 91.4% 93.7% 87.1% 92.6%
Decrease 5.4% 5.5% 5.5% 5.9%
Stop buying milk 0.3% 0.3% 0.5% 0.5%
Dont know 2.4% 6.0%
39
TABLE 17
Consumers Who Usually Purchase Milk Identified as Coming from Untreated Herds
DGZ Sample BLP Sample
(N=165) (N=29)
Respondent knew milk was
from untreated herd because of
Store policy 52.0% 65.5%
Milk was labeled untreated 60.7% 58.6%
Policy of the brand 65.9% 65.5%
Milk was labeled organic 26.0% 34.5%
40
TABLE 18
Respondent Demand Response by Those Who Are Not Concerned and Those Who Are
Concerned
with Future Discovery of Health Risks Associated with rbGH
DGZ Sample BLP Sample
No Concern Concerned No Concern Concerned
(N=233) (N=1628) (N=32) (N=172)
Demand response**
Increase 0.5% 0.6% 1.2%
No change 97.1% 90.9% 90.6% 86.6%
Decrease 0.6% 6.2% 6.3% 5.8%
Stop buying milk 0.3% 0.6%
Dont know 1.8% 2.0% 3.1% 5.8%
* p value < .05; ** p value < .01; chi-square analysis comparing respondents who are concerned and
those who are not concerned with future discovery of health effects. Analysis comparing the DGZ
sample and the BLP sample was not performed.
41
milk you buy comes from rbGH-treated cows, or not? If so, they were asked, Which do you prefer,
milk from the treated or untreated cows? Many (68.7 percent and 60.8 percent) of the total DGZ and
BLP samples, respectively, preferred milk from untreated herds (see Table 19). The DGZ sample was
significantly different from the BLP sample ( = 22.82, p value = .0001). Of the total DGZ sample,
2
2 df
only 1 percent preferred milk from treated cows and 29 percent had no preference while almost 4
percent of the total BLP sample preferred milk from treated cows and 34 percent had no preference. Of
those who preferred untreated herd milk, 60 percent from the total DGZ sample and 70 percent from
the total BLP sample indicated they usually purchase milk identified as coming from untreated herds.
The results in Table 20 show the concerned groups have a significantly different milk
preference than the no concern groups ( = 263.4, p value = .0001). The concerned group in both
2
samples prefer untreated herd milk, and more of the no concern group has no preference between
treated or untreated herd milk. Similar results were found when exploring whether or not preferences
influence demand response (see Table 21). Those respondents who stated no preference or who
preferred treated herd milk indicated a demand response of no change or an increase in the amount of
milk they bought. Respondents with a preference for untreated herd milk were more likely to state they
had decreased or stopped buying milk than those who had no preference or preferred treated herd milk.
Consumer Support for Labeling Policy
Another interesting result was the percentage of respondents answering yes to the question,
Do you think milk should be labeled in such a way that you could distinguish between milk from
treated and untreated cows? Ninety-four percent of the total DGZ sample and 93.5 percent of the total
BLP sample responded yes to this question on labeling (see Table 22). Only 4.4 percent of the total
42
TABLE 19
Consumers Preference for Treated versus Untreated Milk
DGZ Sample BLP Sample
(N=1910) (N=217)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
Milk preference**
No preference 29.6% 29.8% 34.4% 34.8%
Prefer treated 1.1% 1.1% 3.8% 3.8%
Prefer untreated 68.7% 69.1% 60.8% 61.4%
Dont know 0.6% 1.1%
* p value < .05; ** p value < .01; chi-square analysis comparing DGZ sample and the BLP sample
respondents who expressed an opinion.
43
TABLE 20
Respondent Preference for Treated versus Untreated Milk by Those Who Are Not Concerned
and Those Who Are Concerned with Future Discovery of Health Risks Associated with rbGH
DGZ Sample BLP Sample
No Concern Concerned No Concern Concerned
(N=233) (N=1628) (N=32) (N=172)
Milk preference**
No preference 76.0% 22.8% 63.0% 28.3%
Prefer treated herd milk 1.7% 0.9% 7.4% 2.0%
Prefer untreated herd milk 21.7% 75.7% 25.9% 69.1%
Dont know 0.6% 0.6% 3.7% 0.7%
* p value < .05; ** p value < .01; chi-square analysis comparing respondents who are concerned and
those who are not concerned with future discovery of health effects. Analysis comparing the DGZ
sample and the BLP sample was not performed.
44
TABLE 21
Respondents Demand Response by Preference for Treated versus Untreated Herd Milk
DGZ Sample BLP Sample
(N=1765) (N=184)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
No preference demand response
Increase 1.0% 1.0% 1.6% 1.7%
No change 96.7% 98.3% 89.1% 96.6%
Decrease 0.7% 0.7% 1.6% 1.7%
Stop buying milk
Dont know 1.6% 7.8%
Prefer treated herd milk
demand response
Increase 6.0% 6.6%
No change 84.6% 93.4% 100% 100%
Decrease
Stop buying milk
Dont know 9.4%
Prefer untreated herd milk
demand response
Increase 0.3% 0.3% 0.9% 0.9%
No change 90.0% 91.3% 87.6% 90.0%
Decrease 7.8% 7.9% 8.0% 8.2%
Stop buying milk 0.4% 0.4% 0.9% 0.9%
Dont know 1.4% 2.7%
45
TABLE 22
Consumer Support for Labeling that Distinguishes between Milk
from rbGH-Treated versus Untreated Herds
DGZ Sample BLP Sample
(N=1910) (N=217)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
Should there be labels to
distinguish between milk from
treated and untreated cows?
Yes 94.2% 95.5% 93.5% 94.4%
No 4.4% 4.5% 5.5% 5.6%
Dont know 1.3% 0.9%
46
DGZ sample and 5.5 percent of the total BLP sample said there should not be labels to distinguish milk
from treated and untreated cows. Similar results were found among those expressing an opinion.
Gender Comparison Regarding Food-Related Biotechnologies
Several patterns emerged when the data were examined according to gender (see Table 23). A
significantly larger percentage of males were aware of biotechnology and rbGH than females ( =
2
1 df
31.33, p value = .0001; = 10.64, p value = .001, respectively). Two-thirds of the men were aware
2
1 df
of biotechnology, whereas just over half of the females were aware. Females were more aware of rbGH
(61 percent) than they were of biotechnology (52.6 percent), but less aware than males (69.3 percent)
regarding rbGH. Awareness of rbGH was similar for both males and females.
Among the males expressing an opinion, over 70 percent approved of biotechnology; 50 percent
of females approved. Of those who expressed their opinions, a significantly larger percentage of
females than males felt rbGH was a poor idea (61.3 percent and 50.6 percent, respectively) ( =
2
3 df
49.16, p value = .0001), while more than half of both males and females said they would purchase pork
produced with rpGH if it were approved by the FDA.
More females than males said they were very concerned about current health effects and the
future discovery of human ill health effects associated with rbGH. Eighty percent of females and
around 60 percent of males expressed some level of concern about current human ill health effects
associated with rpGH and rbGH ( = 52.7, p value = .0001; = 77.76, p value = .001,
2 2
3 df 3 df
respectively). Furthermore, a significantly larger percentage of females (87.8 percent) expressed a
concern for the future discovery of human ill health effects ( = 44.26, p value = .0001).
2
3 df
Males and females had similar demand responses, with women more likely to believe there
should be labels to distinguish between milk from untreated and treated herds ( = 21.23, p value =
2
1 df
.0001). Further, a significantly greater percentage of females preferred milk from untreated herds
47
TABLE 23
Gender Comparison Regarding Food-Related Biotechnologies
Male Female
(N=539) (N=1371)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
Awareness
Awareness of Biotechnology**
Yes 67.0% 67.0% 52.6% 52.9%
No 33.0% 33.0% 46.9% 47.1%
Dont know 0.5%
Awareness of rbGH**
Yes 69.3% 69.8% 61.5% 61.9%
No 29.9% 30.2% 37.9% 38.1%
Dont know 0.8% 0.6%
Awareness of rpGH
Yes 25.6% 25.8% 23.2% 23.3%
No 73.8% 74.2% 76.3% 76.7%
Dont know 0.5% 0.5%
Approval
Overall approval of biotechnology**
Approve 66.6% 72.1% 47.4% 53.5%
Disapprove 25.7% 27.9% 41.2% 46.5%
Dont know 7.7% 11.4%
Overall opinion of rbGH**
Poor idea 47.1% 50.6% 56.4% 61.3%
Fair idea 27.1% 29.1% 27.1% 29.5%
Good idea 13.6% 14.6% 7.1% 7.7%
Excellent idea 5.4% 5.7% 1.4% 1.5%
Dont know 6.9% 8.0%
Purchase rpGH treated-pork
if approved by FDA*
Yes 56.5% 60.7% 47.5% 54.5%
No 36.5% 39.3% 39.7% 45.5%
Dont know 6.9% 12.8%
(table continues)
48
TABLE 23, continued
Male Female
(N=539) (N=1371)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
Current Health Concern
rpGH**
No concern 35.8% 35.8% 19.2% 19.2%
Concerned a little 8.9% 8.9% 10.1% 10.2%
Moderately concerned 25.0% 25.0% 31.5% 31.6%
Very concerned 30.3% 30.3% 38.9% 39.0%
Dont know 0.2%
rbGH**
No concern 34.6% 36.1% 16.1% 17.1%
Concerned a little 6.8% 7.1% 9.0% 9.6%
Moderately concerned 26.4% 27.6% 31.6% 33.4%
Very concerned 28.0% 29.2% 37.8% 40.0%
Dont know 4.1% 5.4%
Future Health Concern
rbGH**
No concern 18.5% 19.0% 9.7% 10.0%
Concerned a little 13.8% 14.2% 9.9% 10.2%
Moderately concerned 31.5% 32.4% 32.0% 32.8%
Very concerned 33.4% 34.3% 45.9% 47.0%
Dont know 2.8% 2.5%
Demand response
Increase 0.5% 0.5% 0.6% 0.6%
No change 92.3% 94.7% 91.1% 93.2%
Decrease 4.2% 4.3% 5.8% 6.0%
Stop buying milk 0.4% 0.4% 0.2% 0.2%
Dont know 2.6% 2.3%
Milk preference**
No preference 40.3% 40.3% 25.4% 25.6%
Prefer treated 0.6% 0.6% 1.3% 1.3%
Prefer untreated 59.1% 59.1% 72.5% 73.0%
Dont know 0.8%
(table continues)
49
TABLE 23, continued
Male Female
(N=539) (N=1371)
Among Those Among Those
Total Expressing Opinion Total Expressing Opinion
Should there be labels to
distinguish between milk from
treated and untreated cows?**
Yes 90.5% 92.0% 95.7% 96.9%
No 7.9% 8.0% 3.1% 3.1%
Dont know 1.6% 1.2%
* p value < .05; ** p value < .01; chi-square analysis comparing males and females for those
respondents who expressed an opinion.
50
than males (around 73 percent and 60 percent, respectively) ( = 37.83, p value = .0001). Forty
2
2 df
percent of the males had no preference, compared to a quarter of the females.
CONCLUSIONS
The primary goal of this study was to understand what factors influence risk perception
associated with rbGH and rpGH, and whether these perceptions differ by income. Previous research has
shown inconsistencies on the effect of income in discussing consumers risk perceptions toward rbGH-
treated herd milk (McGuirk, Preston, and Jones 1992; Grobe and Douthitt 1995). Results from
McGuirk, Preston, and Jones (1992) found that individuals with incomes in the $20,000$50,000 range
were the most worried about the long-term health effects of milk from cows treated with rbGH, and
those individuals with incomes over $30,000 were most skeptical of the governments ability to ensure
that milk supplies are safe. In a previous study (1990), they also found that individuals with annual
incomes of $10,000$20,000 wanted approval of rbGH. Conversely, Grobe and Douthitt (1995) found
low-income respondents to be more apprehensive toward rbGH-treated herd milk. This report
compared the differences between a national sample (DGZ sample) and a poverty sample (BLP
sample). The results indicate a significant difference between the national and poverty samples
concerning awareness of biotechnology and rbGH, knowledge of rbGH, current concern for rbGHs
health effects, concern for the future discovery of rbGH health effects, and preference for milk from
treated or untreated herds.
Compared to the DGZ respondents, the BLP respondents had lower awareness levels of
biotechnology and rbGH. Fewer BLP respondents approved of biotechnology than those from the DGZ
sample (45 percent compared to 50 percent, respectively), and more felt rbGH was a poor idea (60
percent) than the DGZ sample (50 percent). Thus, these results do not confirm either of the McGuirk,
Preston, and Jones studies (1990, 1992).
51
More of the BLP respondents were very concerned with the current health effects and future
discovery of health effects associated with rbGH than the DGZ sample. The BLP respondents were less
approving of biotechnology in general than the DGZ respondents (45 percent versus 50 percent,
respectively) and were more likely than the DGZ respondents (60 percent versus 50 percent) to believe
that rbGH was a poor idea.
The supporting objectives were to analyze the correlates to consumer risk perception and the
impact of commercial use of rbGH on consumer demand for fluid milk, one year after rbGHs approval
by the FDA. The findings presented in this report reflect that most consumers are aware of agricultural
uses of biotechnology and rbGH, but less aware of rpGH. These awareness levels are much higher than
those found by Hoban (1994), who also conducted a national survey of households after the FDA
approved rbGH for commercial use. Hoban found only 4 percent of the respondents had read or heard a
lot about bST, 15 percent had heard something, and 62 percent had heard nothing at all. But in its 1990
nationwide survey, the National Dairy Board found that 62 percent of their respondents were aware of
rbGH, a finding similar to that of this study (Smith and Warland 1992).
More of the consumers surveyed approved of the agricultural uses of biotechnology than
approved treating cows with rbGH. Additionally, even though rpGH is not currently approved for
commercial use, consumers were more positive about purchasing pork products treated with it, if it is
approved, than they were about rbGH. These results suggest that consumers are not adverse to
biotechnology applications in general, but exhibit responses specific to each application.
A large percentage of consumers surveyed were also concerned about the discovery of human
ill health effects. To further explore these concerns, several variables thought to influence risk
perceptions were evaluated for those who were concerned and those who were not concerned.
Concerned consumers seemed to believe that milk is nutritious and important for the growth and
development of children. Yet they did not have the same strong opinions about milk being a natural and
52
pure product. It has been suggested that the trustworthiness of information sources might be a factor in
consumers risk perceptions (van Ravenswaay 1995). However, the results of our survey showed that
the two groups had similar responses about the trustworthiness of information sources. Both the
concerned and no concern groups found the FDA, nutrition information labels, and the USDA as
somewhat to very trustworthy. The DGZ sample ranked physicians as their most trustworthy food-
related information source, followed by the USDA and FDA, while the BLP sample ranked the USDA
as being the most trustworthy, followed by farmers organizations and the FDA. Both samples ranked
advertisements as being the least trustworthy food-related information source. Other research findings
on the trustworthiness of information sources has been mixed (van Ravenswaay 1995).
The two groups expresses divergent knowledge of rbGH and non-rbGH milk. The concerned
group felt there is a difference between rbGH and non-rbGH milk, that rbGHs health implications are
not known, and that rbGH does not benefit consumers. All respondents seemed to be less confident
answering the statements on whether bGH occurs naturally in milk, the cows health effects, and the
economic effects on farmers. This was evident by the high number of dont know responses.
The greatest diversity between the concerned and not concerned group was their locus of
control and group affiliation. The concerned group more strongly agreed that they were worried about
the future of children and that they felt helpless in the world. They also more strongly identified with
environmentalists and animal rights groups.
But consumers demand for milk since rbGHs commercial adoption has remained largely the
same. Only a small proportion of consumers have decided to reduce or stop buying milk and to replace
cows milk with other substitutes. However, 8 percent of the DGZ respondents said that they now
usually only buy milk identified as coming from nontreated cows. This represents 10.6 percent of the
total fluid milk consumption by DGZ respondent households. Further, a large percentage (69 percent)
of the DGZ respondents who expressed an opinion, would prefer milk from non-rbGH treated herds. Of
53
the BLP respondents expressing an opinion, 61 percent would prefer milk from untreated herds. This is
also indicated by the result that almost all the consumers surveyed would like to see milk labeled in a
way that would distinguish between milk from treated versus untreated herds. Consumers have
expressed that they would like a choice in their milk purchase decision.
54
55
See Grobe, Douthitt, and Zepeda. Forthcoming 1996a. Exploring Consumers Risk
1
Perceptions Toward Recombinant Bovine Growth Hormone (rbGH) and Recombinant Porcine Growth
Hormone (rpGH) by Income and Gender: A Focus Group Study. Mimeo. Available from Institute for
Research on Poverty, University of WisconsinMadison.
See the Survey Instrument for further information on the randomization sequence, Grobe,
2
Douthitt, and Zepeda. Forthcoming 1996b. Survey Instrument: Measuring Consumer Knowledge and
Risk Perception of Food-Related Biotechnologies. Mimeo. Available from Institute for Research on
Poverty, University of WisconsinMadison.
See the Survey Instrument for variables IBGH and DRES.
3
See the Survey Instrument for the rpGH block of questions APGH through PPRK and the
4
rbGH block of questions ABGH through SCON.
To select a random digit Targeted Income Sample, SSI computes an average of the income
5
predictor score at the household level for each telephone exchange. Survey sampling uses a
sophisticated income predictor to select samples that target households within a specified income range.
The income predictor is derived from a multiple regression analysis of both individual household data
and Census data at the block group level. The individual household data included information such as
automobile ownership, length of residency, and type of dwelling unit; the Census data are based on over
two hundred variables related to income from the U.S. Census. Then the exchanges are ranked by
predicted income. Once a geographic definition has been determined (for this project it was defined as
the continental United States), a particular income level was specified. The sample was selected only
from those exchanges where the average of the income predictor scores was calculated to be at that
level or lower.
Adjusted response rate = (completed + total non-response) / (completed + total non-response +
6
Endnotes
56
total non-sample).
Annual update of the U.S. Department of Health and Human Services Poverty Guidelines.
7
Federal Register 60, no. 27 (Thursday, February 9, 1995): 77727774.
There were, by chance, no Vermont cases in the National sample, although there were
8
Wisconsin cases in the National sample.
In complying with the independence assumption, the DGZ sample needed to be separated so
9
that the BLP sample would no longer be a subsample. Thus, the two groups of comparison were the
BLP Sample (N=217) and the DGZ sample (subtracting out the BLP respondents) (N=1693), for those
respondents who expressed an opinion.
57
References
Bureau of the Census. 1992. 1990 Census of Population, General Population Characteristics of the
United States. Washington, D.C.: U.S. Department of Commerce (CP-1-1).
Eom, Y. S. 1993. Self-Protection, Risk Information, and Ex Ante Values of Food Safety and
Nutrition. Paper presented at the conference Valuing Food Safety and Nutrition, Alexandria,
VA, June 24, 1995.
Grall, T. A. 1994. Households at Risk: Their Housing Situation. (Current Housing Reports, H121/94-2).
Washington, D.C.: U.S. Department of Housing and Urban Development.
Grobe, D., and R. Douthitt.1995. Consumer Acceptance of Recombinant Bovine Growth Hormone:
Interplay between Beliefs and Perceived Risks. Journal of Consumer Affairs 29 (1): 128143.
Groth, E. 1991. Communicating with Consumers on Food Safety and Risk Issues. Food Technology
45 (May): 3234.
Hadden, S. G. 1989. A Citizens Right to Know: Risk Communication and Public Policy. Boulder, CO:
Westview Press.
Hoban, T. J. 1994. Consumer Awareness and Acceptance of Bovine Somatotropin. Washington, D.C.:
Grocery Manufacturers of America.
McGuirk, A. M., W. P. Preston, and G. M. Jones. 1990. Biotechnology and the Consumer: The Case
of Bovine Somatotropin. Department of Agricultural Economics, Virginia Polytechnic and
State University. Blacksburg, VA. Staff paper 90-6.
McGuirk, A. M., W. P. Preston, and G. M. Jones. 1992. Biotechnology and the Consumer: The Case
of Bovine Somatotropin. American Journal of Agricultural Economics 72(5): 13441354.
Research Alliance. 1990. Consumer Reactions to the Use of bST in Dairy Cows. Report prepared for
the National Dairy Promotion and Counseling Service.
58
Senauer, B., E. Asp, and J. Kinsey. 1991. Food Trends and the Changing Consumer. St. Paul, MN:
Eagan Press.
Slusher, B. J. 1990. Consumer Acceptance of Food Production Innovations: An Empirical Focus on
Biotechnology and BST. In Enhancing Consumer Choice: Proceedings of the Second
International Conference on Research in the Consumer Interest, pp. 105106.
Smith, B. J., and R. H. Warland. 1992. Consumer Responses to Milk from bST-Supplemented Cows.
In Bovine Somatotropin and Emerging Issues, ed. M. C. Hallberg, pp. 243264. Boulder, CO:
Westview Press.
Sterngold, A., R. Warland, and R. Hermann. 1994. Do Surveys Overstate Public Concerns? Public
Opinion Quarterly 58: 255263.
Van Ravenswaay, E. 1995. Public Perceptions of Agrichemicals (Task Force Report no. 123). Ames,
IA: Council for Agricultural Science and Technology.