Food Protection Trends - January/February 2018 - 28

to a portion of the unpasteurized milk products in return for
their financial investment (8).
Regardless of the way in which products are obtained,
unpasteurized dairy products are a low proportion of overall
consumption in the USA. Only 1.6% of English-speaking
pregnant women reported consuming unpasteurized milk,
whereas 10.4% of Spanish-speaking women reported unpasteurized milk consumption (10). The Centers for Disease
Control and Prevention (CDC) report that dairy products
are the second most frequent food associated with foodborne
illnesses, many of which result from consumption of unpasteurized products (17). From 1998 to 2011, CDC reported
2384 illnesses resulting from consuming unpasteurized milk
products; 284 required hospitalization and two resulted in
death (3). Data from Minnesota indicated that 3.7% of all
foodborne illnesses were attributed to previous consumption
of unpasteurized milk products (19). These estimates are
complicated by the underreporting of unpasteurized milk
consumption among consumers and their families (15),
and in states where legal sale is prohibited but consumption
continues (11).
An approach to better understand why Americans consume unpasteurized milk products is to develop a mental
model of the behavior. Mental models are most common
in social sciences, where characterizations of behavior are
derived from qualitative study (16). The models describe
underlying assumptions and learned factors that influence
behavioral action, often relying on subconscious information more often than deliberative thought (12). Educating
through public health messaging that essentially communicates, "Don't do that," has little effect on the targeted practice
and the resulting economic impacts of the behavior. A mental
models approach has the potential to explain the hidden,
less public roots of behavior that could prove to be motivational to the individual who is resistant to behavior change.
Combining the qualitative methods of social science with
quantitative data analysis could measure the scope and depth
of behavioral motivators (4). Potentially, this is an approach
that may reveal innovative strategies for educating people as
to why consumption of unpasteurized milk products is not in
the interest of human health (5, 6).
With this approach in mind, the goal of this study was
to research and create mental, behavioral models that give
insight into why Americans consume milk products that are
a known public health risk. Two models were chosen as the
base theories for the construction of the Mental Models of
Milk Product Consumption, The Risk Information Seeking
and Processing (RISP) model (5) and the Theory of Planned
Behavior (TPB) (1) (Fig. 1). The objective was to create
mental models of Americans who consume unpasteurized
milk products, and, for comparison, models of those Americans who consume pasteurized milk products. The beneficiaries of this study are instructors of food safety information
through government regulatory agencies and educational

Food Protection Trends January/February

institutions, those who provide instruction to the public, and
ultimately the consumers of dairy products.
MATERIALS AND METHODS
Categorical variables
The categorical variables in the experimental design
were used to identify the respondent's milk preference and
residency. The Milk Preference categorical variable was
constructed to organize respondents into two groups based
on milk product consumption: Pasteurized (the exclusive
consumers of pasteurized milk) and Unpasteurized (the
combination of exclusive and occasional consumers of
unpasteurized milk). The Residence categorical variable was
the location in which respondents resided (urban or rural),
according to standard definitions from the U.S. Bureau of
the Census (21). Data for Milk Preference and Residence
were then combined to form a single categorical variable
named Milk/Residence, with four sub-groups: Pasteurized/
Urban, Pasteurized/Rural, Unpasteurized/Urban, and
Unpasteurized/Rural. Additionally, a behavioral outcome
variable was constructed from two survey items probing the
type of milk currently consumed. Data were coded to a single
variable with three response choices (Exclusive pasteurized
milk consumer, code = 1; Occasional unpasteurized milk
consumer, code = 2; or Exclusive unpasteurized milk
consumer, code = 3). The variable was used as the behavioral
outcome variable for the TPB portion of the Mental Models
of Milk Product Consumption.
Recruitment and sampling
A modification of snowball and chain-referral sampling
of hidden populations was used to reach respondents (20).
These techniques were selected because the distribution of
unpasteurized milk product consumers in the USA differs
according to state laws (11), and because of our knowledge
that unpasteurized milk consumers are reluctant to openly
participate in research studies (15). The first phase of
recruitment was to advertise the survey in markets known
to appeal to or to have direct contact with consumers of
unpasteurized milk/dairy products. E-mails were sent to
leaders in dairy science and milk marketing, food safety
professionals at educational and governmental institutions,
websites for health professionals and trade associations,
and email listservs. The email explained the study goals
and scope, who we wanted to participate, the method by
which interested persons could participate, and a live link
to the online survey. Participants were compensated and
told of the method to receive compensation through the
online survey. In the next phase of recruitment, first and
subsequent round responders were encouraged to forward
the recruitment email on to individuals who might be
interested or to post it in locations on social media where
potential participants might see the information and the link
to the study. The snowballing that occurred from multiple

Table of Contents for the Digital Edition of Food Protection Trends - January/February 2018