Agriculture and Human Values

, Volume 34, Issue 4, pp 921–932 | Cite as

Get real: an analysis of student preference for real food

  • Jennifer Porter
  • David Conner
  • Jane Kolodinsky
  • Amy Trubek
Article

Abstract

The Real Food Challenge is a national student movement in the United States (U.S.) that aims to shift $1 billion—roughly 20%—of college and university food budgets across the country towards local, ecologically sound, fair, and humane food sources—what they call “real” food—by 2020. The University of Vermont (UVM) was the fifth university in the U.S. to sign the Real Food Campus Commitment, pledging to shift at least 20% of its own food budget towards “real” food by 2020. In order to examine student preference for “real” food on the UVM campus, we analyzed a survey of 904 undergraduate students that used contingent valuation to evaluate students’ willingness-to-pay (WTP) for the “real” attribute. We found that a majority of students are willing to pay a positive premium for “real” food. Furthermore, we found that student characteristics and attitudes significantly influence WTP. Specifically, gender, residency, college, and attitudes about price and origin of food are significant predictors of WTP.

Keywords

Willingness-to-pay Real food challenge Contingent valuation University dining Student values Credence attribute 

Abbreviations

CALS

College of agriculture and life sciences

CAS

College of arts and sciences

CEMS

College of engineering and mathematical sciences

CESS

College of education and social services

CNHS

College of nursing and health sciences

CV

Contingent valuation

IRB

Institutional review board

RFC

Real food challenge

RFWG

Real food working group

RUB

Rubenstein school of environment and natural resources

SBA

School of business administration

USA

United States of America

UVM

University of vermont

WTP

Willingness to pay

Introduction

A growing number of higher education institutions across the United States (U.S.) have adopted sustainable food initiatives on their campuses (Barlett 2011). The various sustainable food initiatives present on higher education campuses and elsewhere, such as local, sustainable, organic and fair trade, are all components of a larger social movement that contests the dominant agro-food system. The Real Food Challenge (RFC)—a national, student movement—was launched in 2008 as way to support and amplify the efforts of all of the seemingly disparate food initiatives on campuses across the U.S.

The RFC aims to address the social, economic, and environmental problems associated with the dominant agro-food system by harnessing the purchasing power of U.S. higher education institutions, which collectively spend about $5 billion annually on food (Steel 2012). The goal of the RFC is to shift 20% (measured by value) of existing higher education food budgets—or $1 billion—towards “real” food by 2020. “Real” food is defined as:

Food that truly nourishes producers, consumers, communities and the earth. It is a food system—from seed to plate—that fundamentally respects human dignity and health, animal welfare, social justice and environmental sustainability. Some people call it “local”, “green”, “slow”, or “fair”. We use “Real Food” as a holistic term to bring together many of these diverse ideas people have about a values-based food economy. (The Real Food Challenge n.d.)

As this definition is rather holistic and non-specific in terms of the balance of credence attributes actually present, the RFC developed criteria, which are third party certifications and producer characteristics, that are divided into four categories: local/community based, ecologically sound, fair, and humane. In order to qualify as “real”, food must meet the criteria of at least one of the four categories:

  • Local and Community-Based: These foods can be traced to nearby farms and businesses that are locally owned and operated. Sourcing these foods supports the local economy by keeping money in the community and builds community relations. The food travels fewer miles to reach consumers. The food is seasonal, and when it is fresh, it often has a higher nutrient content.

  • Fair: Individuals involved in food production, distribution, preparation and other parts of the food system work in safe and fair conditions, receive a living wage, are ensured the right to organize and the right to a grievance process, and have equal opportunity for employment. Fair food builds community capacity and ensures and promotes socially just practices in the food system.

  • Humane: Animals can express natural behavior in a low-stress environment and are raised with no hormones or unnecessary medication.

  • Ecologically Sound: Farms, businesses, and other operations involved with food production practice environmental stewardship that conserves biodiversity and preserves natural resources, including energy, wildlife, water, air, and soil. Production practices should minimize toxic substances as well as direct and indirect petroleum inputs. (The Real Food Challenge n.d.)

The primary way the RFC intends to meet its goal of shifting $1 billion of higher education food budgets towards “real” food is through the Real Food Campus Commitment. This commitment, which was modeled in part on the President’s Climate Commitment, asks presidents of higher education institutions to formally prioritize “real” food. It commits the institution to procure more “real” food (at least 20% by 2020), improve institutional transparency around food purchasing, and increase student and community engagement (The Real Food Challenge n.d.).

To date, 40 institutions (including one high school) have signed the Real Food Campus Commitment, pledging to shift over $60 million towards “real” food (The Real Food Challenge n.d.). Furthermore, students from over 200 institutions have utilized the Real Food Calculator, a tool designed to measure an institution’s “real” food purchasing percentages (The Real Food Challenge n.d.). As the Real Food Movement spreads to campuses across the U.S. and gains popularity amongst students it becomes clear that the term “real” is emerging as a new attribute for food products.

The “real” attribute, like the four attributes encompassed by the term, is a credence attribute, meaning that consumers cannot differentiate it from non-credence alternatives at the point or purchase, or consumption, based on its physical properties alone. Credence attributes may provide private benefits to consumers, but their production often has “affiliated public dimensions” (Lusk et al. 2007). These fairly intangible attributes often have outcomes related to public health, environmental conservation, creation of employment, supporting small-scale agriculture and local rural communities, farm incomes, especially in marginal and/or disadvantaged conditions, and workers’ rights (Moser et al. 2011).

It is important to understand student preference for “real” food if the RFC is to become an established, widespread, national movement. Students are a critical consumer demographic to understand because they are the consumers of tomorrow. The preferences that they develop during their time at university will translate into purchasing behavior in the marketplace once they graduate. Furthermore, it is important to understand students’ preference for “real” food because students are the driving force behind the massive purchasing power of universities. The way in which higher education institutions spend their $5 billion food budget is heavily influenced by student preferences. As such, understanding student preference for “real” food will provide some insight as to how successfully the goals of the RFC can be met.

In this paper we seek to understand students’ preference for the new “real” attribute and ask what factors influence students’ preference for “real” food. To address this question we use students’ willingness-to-pay (WTP) any positive premium for a meal plan consisting of at least 20% “real” food as a proxy for their preference for the “real” attribute. A meal plan at UVM takes multiple forms, but is, generally, a prepaid allotment of meals in campus dining halls, which is required for first and second-year students and is optional for more senior students. We measure students’ WTP for “real” food and explore how personal characteristics and attitudes influence the likelihood of WTP a positive premium for it. By understanding what factors influence students’ preference for “real” food we can identify potential leverage points for increasing support for the RFC.

Background

Demand for “real” food attributes

While the attributes encompassed by “real” food—local, ecologically sound, fair, and humane—are not novel, the usage of a catchall term for them is. Since the Real Food movement is relatively new there has not yet been, to our knowledge, any literature devoted to consumer demand and preference for the new “real” attribute. There has, however, been a wealth of literature devoted to consumers’ preference and demand for various credence attributes of food, including the four attributes encompassed by “real” food. As such, in this section we review the literature on consumer demand for these four credence attributes individually.

Below, Table 1 summarizes the literature examining consumers’ WTP for the four credence attributes incorporated into “real” food. We include studies from the past 10 years that use stated preference methods, as that is the method we use in our survey to examine student preference. As Table 1 demonstrates, consumers are willing to pay a positive premium for these attributes, but premiums vary both by credence value and type of food product. Though the literature has demonstrated a preference for these attributes, it is unclear whether or not the premiums consumers state they are willing to pay are enough to cover the increased costs associated with providing these attributes. Furthermore, it is not known whether the premiums for these attributes are additive, or if the premium for “real” food is less than the sum of its parts.

Table 1

Summary table of studies examining consumer WTP for “real” food attributes

Author(s) and date of publication

Country of study

Sample population

Credence value

Avg. price premium

Type of food

Carpio & Isengildina-Massa (2009)

U.S

General consumers

Local

25%

Produce; meat

Li et al. (2012)

U.S

General consumers

Local; ecologically sound

43%; 48%

Unspecified

Loureiro & Hine (2002)

U.S

General consumers

Local; ecologically sound

$0.09/lb; $0.06/lb

Potatoes

Onozaka & McFadden (2011)

U.S

General consumers

Local

9–15%

Produce

Loureiro & Lotade (2005)

U.S

General consumers

Ecologically sound fair

$0.16/lb −$0.20/lb; $0.22/lb

Coffee

Lusk et al. (2007)

U.S

General consumers

Ecologically sound humane

$0.67/lb; $0.84/lb

Pork

Mahé (2010)

Switzerland

General consumers

Fair

$0.47/lba

Bananas

Trudel & Cotte (2009)

U.S

General consumers

Fair

$1.40/lb

Coffee

Naald & Cameron (2011)

U.S

General consumers

Humane

$0.35/lb

Chicken

Tonsor et al. (2009)

U.S

General consumers

Humane

$2.11/lb

Pork

aConverted from CHF/kg

In a national survey of college students in the U.S., Feenstra et al. (2011) found that slightly more than half of students considered it important that their food was humanely raised (about 62%) and was produced by workers earning a living wage (about 51%). Less than half of students considered it important that their food was grown sustainably (about 41%), locally (about 30%), certified organic (about 25%), or on a small farm (about 18%). These results suggest that a significant proportion of students in the U.S. value the attributes incorporated into “real” food.

Beyond demonstrating that students have positive attitudes toward the attributes of “real” food, there is evidence suggesting that students may prefer—as indicated by WTP—campus food that has one or more of the attributes of “real” food. Of students surveyed at Clark University, 30% were very willing to pay more (an undetermined amount) for a meal plan that had a higher percentage of local, organic, sustainable food, and 51% were somewhat willing to pay more (Clark University 2010). Feenstra et al. (2011) found that over 40% of students would be willing to pay a 14% premium for a salad if it was organic, local, sustainably produced, or produced in accordance with living wage guidelines.

As the literature demonstrates, there exists varied demand for the credence attributes encompassed by “real” food. Furthermore, the literature supports the claim that students, in particular, prefer these attributes. Indeed, the existence and widespread popularity of the RFC supports this, as it is a student-led initiative. Our study contributes to the literature by examining student preference for these four attributes aggregated under the “real” attribute.

Theoretical framework

The theoretical framework we use for this analysis is Lancaster’s (1966) theory of consumer demand, which suggests that utility is derived from bundles of attributes of goods rather than from goods themselves. Thus, the utility derived from a good is a function of the good’s attributes, given the consumer’s preferences. This can be represented as:
$$U{\text{ }} = {\text{ }}f{\text{ }}(X_{1} ,X_{2} , \ldots ,X_{n} ;P)$$
where U is utility, P is the consumer’s preferences, and Xi are attributes of the good in question. We assume that a consumer will choose the goods with attributes that maximize his or her utility, given his or her preferences. Because the utility derived from attributes is a latent construct, it is not directly observable. A proxy measure of utility, however, can be estimated by WTP, because it is assumed that a consumer will be willing to pay a price premium for a given attribute if he or she derives utility from that attribute (or bundle of attributes).

In order to understand what factors influence students’ preference for “real” food, we first need to measure weather or not a student has a preference for the “real” attribute. Theoretically, students may prefer “real” food because they derive utility from the “real” attribute. We use WTP as a proxy to empirically measure the unobservable utility derived from the “real” attribute. The “real” attribute is just one of many attributes of food that may increase utility for students. Choosing to pay a premium for the “real” attribute necessarily requires students to negotiate tradeoffs between other attributes that they may derive utility from. We chose WTP as a proxy for preference precisely because it requires a negotiation of competing attributes. We are interested in whether students value the “real” attribute, given other attributes, enough to pay a price premium for it, because students’ purchasing power is their primary method for indicating preferences and signaling demand to the market. We therefore expect that positive WTP suggests the value of “real”. We seek to measure how many students in our sample value “real” as expressed by positive WTP, as well as which attitudinal and demographic attributes are associated with positive WTP in order to inform marketing and educational efforts.

Methods

We define the “real” attribute as meeting at least one of the following criteria (as previously defined): local, ecologically sound, fair, or humane. Since there is currently no market for the specific bundle of attributes that define “real” food, we use the contingent valuation (CV) method to estimate students’ WTP for it. The CV method circumvents the absence of a market for “real” food by presenting students with a hypothetical market in which they have the opportunity to buy a meal plan that offers at least a minimum amount (20%) of “real” food. CV has been widely used to value consumer demand for the individual characteristics of “real” food, as demonstrated in Table 1. In this study we use CV to measure students’ preference for the “real” attribute as a bundle of the four characteristics it encompasses.

As summarized by Breidert et al. (2006) there are some criticisms of measuring WTP with direct consumer surveys, like the CV method: consumers can over or underestimate their WTP; stated WTP does not necessarily translate into purchasing behavior; and focusing on price can displace some of the product’s other attributes. These criticisms must be weighed against the convenience and cost-effectiveness of the CV method. While there may be some concerns about the hypothetical nature of the CV method, Loureiro et al. (2003) found that consumers who reported a WTP a premium for a product was actually more likely to purchase that product.

Although models based on stated preference are not as reliable as models based on revealed preferences, we had no way of obtaining students’ revealed preferences for “real” food on campus. Though UVM purchased 13% “real” food, annually, at the time of our survey, students were not always informed about which foods did and did not qualify as “real”. Furthermore, the “real” food was mixed into unlimited meal plan options as well as retail dining locations. These two factors made it difficult for students to demonstrate their preference for the “real” attribute, because either the attribute was concealed from them or they were not given a choice whether or not to purchase an item with the attribute. Because of these limitations, we utilize CV to obtain students stated preference for a meal plan consisting of at least 20% “real” food.

The type of CV method we utilize in this study is the payment card method, which involves a survey question that has an ordered set of threshold values and the respondent is asked to choose the highest value they would be willing to pay for the good in question (Mitchell and Carson 1989). The payment card method is very convenient (no need for interviewer prompts) and it largely avoids the problem of non-response with open-ended questions (Cameron and Huppert 1989). This method is, however, vulnerable to biases associated with the range of values and the intervals between the values (Mitchell and Carson 1989). To mitigate these biases, we minimized the intervals between values and allowed respondents to choose not to pay any premium, or pay less than or greater than our range of premiums.

Survey instrument

Our research was conducted at The University of Vermont (UVM), a land grant university located in Burlington, Vermont. In the fall of 2013, when this research was conducted, UVM had 11,781 students enrolled in degree programs, 9970 of whom were undergraduate students. UVM was one of the first institutions to pilot the Real Food Calculator in 2009 and was the fifth institution in the U.S. to sign the Real Food Campus Commitment in 2012. At the time this research was conducted, UVM had been implementing the Campus Commitment for about a year and a half and was already spending 13% of its annual food budget on “real” food (The University of Vermont n.d.).

In the fall of 2013 the Real Food Working Group (RFWG)—a multi-stakeholder group on campus tasked with implementing the Real Food Campus Commitment—partnered with a UVM undergraduate social research methods class in a service-learning context to develop and conduct a survey of undergraduate students at UVM. Two of the authors of this paper were members of the RFWG and one served as a representative from the RFWG to advise the students on what type of information to elicit from the survey. UVM’s Institutional Review Board (IRB) approved protocols and all students completed an IRB on-line training program before data collection began.

Though members of the RFWG had input into the creation of the survey, the students ultimately designed the survey as a class. The survey collected the following information: demographic characteristics, awareness of the RFC, most frequent dining location on campus, the importance of several attributes when deciding where to eat on campus, and willingness to pay more for a meal plan that had at least 20% “real” food. The survey was coded prior to distribution to ensure that response data would be valid and useful.

The students in the class were responsible for distributing the survey using convenience-sampling methods in November 2013. Each of the 48 students in the class was given 20 surveys to distribute to UVM undergraduate students on campus and at other student gathering spots. A total of 904 surveys were completed. Table 2 shows demographic characteristics of our sample and compares them to the demographics of the entire UVM undergraduate population.

Table 2

Demographic characteristics of survey respondents

Characteristic

Sample % (n = 904)

UVM % (n = 9970)

p value

Gender (n = 890)

 Male

43.6

44.0

0.81

 Female

56.4

56.0

 

Classification (n = 892)

 First year

26.5

27.6

0.31

 Second year

35.2

24.1

0.00*

 Third year

23.0

22.8

0.98

 Fourth year

15.4

25.5

0.00*

Residency (n = 882)

 In-state

34.5

31.9

0.12

 Out-of-state

65.5

68.1

 

College (n = 892)

 CALS

23.0

13.1

0.00*

 CAS

33.0

45.9

0.00*

 RUB

5.8

6.2

0.83

 SBA

13.1

8.0

0.00*

 CEMS

8.9

10.1

0.07

 CESS

8.3

7.7

0.41

 CNHS

8.0

9.0

0.23

Percentages may not add up to 100 due to rounding

*Significant at α= 0.05

Though the sample was a convenience sample, Table 2 demonstrates that the demographic information of the sample is similar to the UVM undergraduate student population, with two exceptions. The only significant differences were among class years and colleges. Second year students were slightly overrepresented and fourth year students were slightly underrepresented. We hypothesize that fourth year students may have been underrepresented because the majority of them do not live on campus or have a meal plan, and therefore less of them may have been on campus when the surveys were distributed. The College of Agriculture and Life Sciences (CALS) and the School of Business Administration (SBA) were slightly overrepresented and the College of Arts and Sciences (CAS) was slightly underrepresented. CALS may have been overrepresented because most of the students in the class that administered the survey were CALS students.

Before presenting students with the CV question during the survey we included a brief description of “real” food, as it was assumed that not all students were aware of the definition of it. The CV question we used to elicit students’ WTP for real food was as follows: Consider the resources you and your parents/guardians have to pay for your meals at college. How much more would you be willing to pay for your meal plan if over 20% of the food was defined as “real” using the qualifications above (circle one)? The options they were able to choose from were presented in dollars/semester and represented a less than 1, 1, 3, 5 and 10%, or greater than 10% premium. There was also an option to not pay any more per semester. The premiums were calculated based on the average meal plan price (for financial aid) of $1883.00. We chose to set the minimum amount of “real” food in the hypothetical scenario at 20%, because that is the minimum amount UVM had committed to reaching by 2020.

Statistical analysis

The data were analyzed with Statistical Package for Social Sciences (SPSS) Version 21. We used a binary logistic model to identify how personal characteristics and attitudes influence students’ preference for “real” food. Competing models, such as tobit and ordered probit, revealed few, if any, significant predictors explaining the variability in WTP, given there was a positive WTP. As such, we chose to use a binary logistic regression to determine which predictors influenced whether a respondent was willing to pay any premium or not.

Logistic regression enables us to predict whether or not an individual is a member of a group (yes or no) based on a set of explanatory variables (X). The dependent variable (y) is dichotomous and can take the value of 1 (member of the group) or 0 (not a member). In logistic regression the relationship between independent and dependent variables is not linear. Rather, the dependent variable is transformed by the logit function as such:
$${\text{Logit}}\left( {y\left( x \right)} \right)=\alpha +{\beta _1}{X_1}+{\beta _2}{{\text{X}}_2} \ldots +{\beta _i}{{\text{X}}_i}$$
where \({\alpha }\) is the constant term and β is the coefficient of independent variables.

Logistic regression predicts the odds ratio for each independent variable, which is a measure of association between the presence of an independent variable and membership in the group (y = 1). An odds ratio of one indicates that the given independent variable has no effect on an individual’s membership in the group. An odds ratio above one indicates that increasing the independent variable by one unit increases the odds that the individual will be in the group (y = 1) by a magnitude of the odds ratio, holding all other independent variables constant. Conversely, an odds ratio below one indicates that increasing the independent variable by one unit will decrease the odds that the individual will be in the group (y = 1) by a magnitude of the odds ratio, holding all other independent variables constant.

The dependent variable in the model is a binary indicator of a student’s WTP a positive premium of any magnitude for a meal plan that consists of at least 20% “real” food (i.e. zero or non-zero WTP). We use students’ positive WTP as a proxy for preference for “real” food. If a student is willing to pay any positive premium we assume that he or she derives utility from the “real” attribute and, thus, has a preference for “real” food. We believe students that are unwilling to pay any premium are fundamentally different than those that are willing to pay even a very small (less than 1%) premium. We built the model by including demographic characteristics that were hypothesized to influence preferences as well as variables that measure students’ attitudes. A description of the explanatory variables included in the model can be found in Table 3. The model was specified as:

Table 3

Explanatory variables used in binary logistic regression model

Variable code

Description of variable code

FEMALE

1 = student is a female; 0 = male

SOPHOMORE

1 = student is a sophomore; 0 otherwise

JUNIOR

1 = student is a junior; 0 otherwise

SENIOR

1 = student is a senior; 0 otherwise

INSTATE

1 = student is a Vermont resident; 0 otherwise

CALS

1 = student is in College of Agriculture and Life Sciences; 0 otherwise

RUB

1 = student is in Rubenstein School of Environment and Natural Resources; 0 otherwise

ORIGINMOST

1 = student considers the origin of food to be very important to them when deciding where to eat; 0 otherwise

PRICEMOST

1 = student considers price to be very important to them when deciding where to eat; 0 otherwise

$$y={\beta _0}+{\beta _1}{\text{FEMALE}}+{\beta _2}{\text{INSTATE}}+{\beta _3}{\text{CALS~}}+{\beta _4}{\text{RUB}}+{\beta _5}{\text{ORIGMOST}}+{\beta _6}{\text{PRICEMOST}}+{\beta _7}{\text{SOPHOMORE}}+{\beta _8}{\text{JUNIOR}}+{\beta _9}{\text{SENIOR}}+ \in$$
We include an indicator variable for whether or not a student is a Vermont resident (INSTATE), because we hypothesize that Vermont residents may be more likely to prefer “real” food, given the strong local and sustainable food movements in Vermont (U.S. Department of Agriculture National Agricultural Statistics Service 2007; U.S. Department of Agriculture National Agricultural Statistics Service 2008; Vermont Sustainable Jobs Fund 2013). One of the demographic characteristics represented in our model is the college that students are enrolled in. UVM has seven colleges, but we only include two of them in the model. We chose to include the College of Agriculture and Life Sciences (CALS) and the Rubenstein School of Environment and Natural Resources (RUB) because those two colleges offer the majority of classes at UVM that pertain to food systems.

Information on students’ class year and meal plan were collected in the survey, but bivariate analyses reveal that these two variables are significantly associated (p = 0.000). Furthermore, bivariate analyses reveal that neither variable is significantly related to students WTP a positive premium (p = 0.605 for class year and p = 0.766 for meal plan). Therefore, we decided to only include one of them in the model. We chose to use class year instead of meal plan, because in addition to implying some meal plan information (e.g. all freshmen must be on an unlimited plan), class year also implies the level of education a student has received. It is hypothesized that a students’ level of education may influence their WTP, as other studies have found that education influences preference for organic, fair trade, local, or humane food products (Loureiro and Hine 2002; Loureiro and Lotade 2005; Onianwa et al. 2005; Naald and Cameron 2011).

Some of the students in our sample had no meal plan because they lived off-campus. We were initially concerned that these students would not be willing to pay any premium and might bias the results. However, bivariate analyses revealed that the frequency of students willing to pay a premium did not vary by the presence of a meal plan (p = 0.966). Therefore, students who did not have a meal plan were left in the sample.

We include two variables that serve as a proxy for consumers’ attitudes. The PRICEMOST variable represents how important price is to a student when he or she decides where to dine on campus. The ORIGMOST variable represents how important the origin of food (i.e. local, organic, fair trade, humane) is to a student when he or she decides where to dine on campus. Both of these attitudes were measured on a Likert scale from one to five (“not at all important” to “very important”). The PRCIEMOST and ORIGMOST variables are indicator variables for whether or not a student responded that the characteristic of the food was very important to them (a 5 on the Likert scale). It is hypothesized that these attitude variables may influence students’ preference for “real” food, as previous studies have demonstrated that attitudes can influence WTP for “real” food characteristics, such as locally produced (Zepeda and Li 2006; Campbell et al. 2014). The survey measured a few additional attitudinal variables, including the importance of the following factors in determining where a student dines on campus: location; atmosphere; availability of to-go options; dietary restrictions; and selection/variety of food. These variables were found to be poor predictors of WTP and were thus omitted from the model.

Results

Table 4 displays the distribution of premia students were willing to pay for a meal plan consisting of at least 20% “real” food. A majority (70.8%) of students were willing to pay a positive premium. The median premium of all students surveyed was $18.00/semester (a 1% increase) and the average premium was $45.00/semester (a 2.4% increase). Of students that were willing to pay any premium over $0.00, the median premium was $56.50/semester (a 3% increase) and the average was $63.55/semester (a 3.4% increase).

Table 4

Distribution of WTP for “real” food

Additional $/semester

(%) Price increase

Students WTP (%)

0.00

0

29.2

<18.00

<1

10.1

18.00

1

21.8

56.50

3

18.4

94.17

5

11.8

188.35

10

3.7

>188.35

>10

5.1

Percentages may not add up to 100 due to rounding

The results for the logistic regression predicting students’ WTP a premium for a meal plan consisting of at least 20% “real” food are displayed in Table 5. Overall, the model was significant (p = 0.000) and correctly assigned 71.1% of students to their correct group (willing to pay or not willing to pay). However, while it correctly identified most (99.5%) of the students who are willing to pay a positive premium, the classification of students that were not willing to pay was quite poor (only 1.2% correctly predicted).

Table 5

Logistic regression predicting students’ WTP

Predictor

β

SE of β

d.f.

p value

eβ (odds ratio)

CONSTANT

0.724

0.180

1

0.000*

2.062

FEMALE

0.311

0.159

1

0.050*

1.364

INSTATE

0.289

0.168

1

0.085*

1.335

RUB

1.065

0.426

1

0.012*

2.900

CALS

0.326

0.196

1

0.097*

1.385

CLASSYEAR

  

3

0.369

 

SOPHOMORE

−0.334

0.202

1

0.098*

0.716

JUNIOR

−0.095

0.229

1

0.679

0.910

SENIOR

−0.083

0.260

1

0.749

0.920

PRICEMOST

−0.572

0.187

1

0.002*

0.564

ORIGMOST

0.712

0.270

1

0.008*

2.039

*Significant at α = 0.10

The purpose of our model is to identify characteristics of students that increase their probability of being willing to pay a premium for “real” food, and our model is very good at predicting these types of students. There is considerable variability among students who are not willing to pay a premium, and our model does not do well at classifying this group. According to the model, females, Vermont residents, RUB students, and CALS students are more likely to be willing to pay a positive premium. Furthermore, students that consider the origin of food to be very important to them when deciding where to eat are more likely to be willing to pay a premium. Conversely, students who consider price to be very important are less likely to be willing to pay a premium. The greatest odds ratio was for RUB students, who the model predicts are nearly three times as likely to be willing to pay a premium than students in other colleges. Though just barely significant, sophomores are slightly less likely to be willing to pay a premium than freshmen. Juniors and seniors, however, do not have significantly different odds than freshmen.

Discussion

Feenstra et al. (2011) found that 40% of students were willing to pay a 14% premium for a salad that was produced sustainably or locally. We only found about 9% of students willing to pay a comparable premium (10% or more) for “real” food. It may not be fair, however, to directly compare our results with other studies, such as Feenstra et al.’s, because we asked students to consider a premium on their meal plan for the entire semester rather than for just one meal. Students may be willing to pay a higher premium for a single meal because it is a smaller incremental cost to consider at the given time and does not lock them into paying that premium each time they want to eat. Our finding that 70.8% of students are willing to pay a positive premium for a meal plan consisting of at least 20% “real” food does align with Clark University’s finding that about 81% of students are somewhat or very willing to pay an undetermined premium for a meal plan consisting of more local, organic, or sustainable food.

We were not surprised to find that in-state students (i.e. Vermont residents) are more likely to be willing to pay a premium, given the strong presence in the market place of food marketed on the basis local and ecological credence values (U.S. Department of Agriculture, National Agricultural Statistics Service 2007; U.S. Department of Agriculture, National Agricultural Statistics Service 2008). In-state students may be more familiar with these types of values and products and may already be in the habit of paying a premium for them. Furthermore, these students may be willing to pay a premium for “real” food because they associate it with local food, and they have a desire to support their local food economy. Previous studies have found that consumers who are motivated to purchase local food by perceptions of support for the local food economy are willing to pay higher premiums (Thilmany et al. 2008; Carpio and Isengildina-Massa 2009).

Our finding that female students are more likely to prefer “real” food echoes Loureiro and Lotade’s (2005) finding that females are more likely to pay a premium for both Fair Trade and organic coffee. Other studies, however, have found that gender does not significantly influence WTP for local or organic food (Loureiro and Hine 2002) or females have a lower WTP for local food or humanely raised animal products (Onianwa et al. 2005; Naald and Cameron 2011).

Our finding that class year, overall, is not a significant predictor of students’ WTP implies that 1–3 additional years of education do not change students’ preferences for “real” food. Although some studies have found education to be positively related to WTP for attributes of “real” food (Loureiro and Hine 2002; Loureiro and Lotade 2005; Onianwa et al. 2005; Naald and Cameron 2011), others have found a negative relationship (Giraud et al. 2005; Jekanowski et al. 2000). Zepeda and Li (2006) found demographics, and education in particular, to be poor proxies for preferences.

Although class year, overall, is not a significant predictor of WTP, second-year students are marginally less likely to be willing to pay than first-year students. This may be because second-year students are tired of eating food included in the meal plan, as all students living on campus (mandatory for first and second-year students) are required to have a meal plan. They may not be willing to pay any premium for a meal plan with “real” food, because they do not want to have a meal plan at all. Third and fourth-year students, however, may not have a significantly different WTP than freshmen because they are not required to have a meal plan and thus feel freer to pay or not pay as they choose.

We found that RUB students and CALS students are more likely to be willing to pay a premium for “real” food than students in other colleges. It is in these two colleges that courses with most connection to environmental and food systems are offered. Therefore, it appears that education may increase preference for “real” food, if that education is ‘relevant to the issue’ (i.e. provides an understanding of the functioning of the food chain). The odds ratio for CALS students may be less than the odds ratio for RUB students because CALS includes a much broader variety of majors and disciplines than RUB. Although CALS houses the Food Systems program, it also houses majors such as public communication and chemistry. Conversely, almost all students in RUB are required to take environmentally-oriented courses that likely expose them to concepts such as ecosystem services and raise awareness of the negative externalities of conventional food production. Furthermore, these students may not only be more likely to be educated about environmental issues as they relate to food production, but they may also be more likely to place a higher than average value on the environment, as they selected a college that has a focus on environmental sciences.

Both of the attitude variables we included in the model were significant. We found that students who consider price to be a ‘very important’ are less likely to be willing to pay a premium for “real” food. This is not surprising, as previous studies have found price consciousness to negatively influence preference for food with credence attributes, such as locally produced (Zepeda and Li 2006; Campbell et al. 2014). Unfortunately, our study was not able to definitively determine whether students who are not willing to pay a premium are not willing to do so because they cannot afford to or because they do not perceive the value of “real” food to warrant a price premium. We assume, however, that almost every student would be able to afford an additional $0.01–$18.00 (<1%) per semester, which was one of the choices in the CV question. Therefore, we assume that since students had the option to pay such a small premium and chose not to, they must not derive any utility from the “real” attribute.

It was also not surprising that we found that students who consider origin of food to be ‘very important’ are more likely to be willing to pay a premium. Essentially, this indicates that students with strong attitudes about the origin of food are willing to act on those attitudes. The high odds ratio of this variable also indicates that a strong attitude towards the origin of food is one of the best predictors of preference for “real” food.

Of our sample, 12.5% reported that the origin of food is ‘very important’ to them, which is considerably less than the 44% of students who reported issues of local, organic, and sustainable food to be very important to them at Clark University (Clark University 2010). However, when we include students that reported the origin of food to be ‘important’ to them (as opposed to ‘very important’) the proportion increases to 35.6%. This number is somewhat in line with Feenstra et al.’s (2011) finding that between 25 and 61% of students consider it important that their food be organic, local, sustainable, fair, or humane. It is difficult to compare findings, however, because we did not ask students about the importance of each of the individual characteristics of “real” food, as Feenstra et al. did.

Implications for practice

The purpose of this paper is to better understand what factors influence students’ preference for “real” food. Our findings indicate that values regarding price and origin of food and awareness of environmental and/or food systems issues are the strongest predictors of students’ preference for “real” food. Fortunately, these are also the only mutable variables in our model. As such, these are the leverage points that offer the most promise for increasing student demand for “real” food and driving momentum for the RFC.

Values are often considered to be enduring, but college can be a “coming-of-age” time in students’ lives when they begin to question their values and beliefs. As such, universities may be particularly effective places to influence students’ values surrounding food. Our results indicate that students who highly value the price of food are less likely to prefer “real” food. It may be quite difficult to change the importance of price in students’ decision making, given constrained budgets. Therefore, it may be more realistic to influence the importance of the origin of food in students’ decision-making processes.

There are two ways universities may influence students’ values with regards to the origin of food. First, universities can lead by example. That is, by modeling behaviors based on specific values, a university may influence students’ values. For example, if a university demonstrates that the origin of food is important by consistently providing students with information about the origin of the food served then students may be more likely to value that information in their food decision-making processes in the future.

The second way universities may influence students’ food-related values is through education, which brings us to the other leverage point—awareness of environmental and/or food systems issues. By informing students of the positive benefits of “real” food and the negative consequences of the “conventional” food it seeks to replace, universities may influence students to weigh the origin of food more heavily in their food decision-making processes. A Chi square test reveals that there is a significant association between the ORIGIN variable and college (p = 0.048), with CALS and RUB having the highest percentage of students who consider the origin of food to be important or very important in their food decision-making process. This suggests that awareness of environmental and/or food systems issues may influence how students value the origin of food. Therefore, education may be the more useful tool for increasing student preference for “real” food.

Our finding that education is one of the best leverage points for influencing students’ preference for “real” food has significant implications for the future of the Real Food movement. If the aim of the RFC is to garner widespread preference for “real” food amongst students across the country, then focusing the campaign on education about food systems issues may be an effective tool. Moreover, individual institutions that wish to increase support for “real” food on campus can do so by focusing on education. Future research could explore the most effective educational methods for influencing student preferences.

Given that values and awareness of environmental and/or food systems issues are the best predictors or preference for “real” food, they may also be the best indicators of a campus’s readiness for the RFC. In order to predict if the RFC will be successful on a campus, it may be worthwhile to gauge how the student population values price and the origin of food and to what extent they are aware of environmental and/or food systems issues. These indicators may predict how willing the student population is to accept and support the RFC. This type of information could be useful for selecting campuses at which to promote the RFC and could help prevent the expenditure of time and resources on campuses that are not ready.

Limitations & future research

There are factors that potentially complicate the measurement of students’ WTP for the “real” attribute. Students may not be responsible for paying for their meal plan, thus making it difficult for them conceive of what type of premium whoever pays for their meal plan would be willing to pay. It also may be difficult for students to conceive how much they (or whoever is paying for their meal plan) would be willing to pay for “real” food over the course of a semester, rather than at just one eating occasion, which is how most studies measure WTP.

Another factor that could have complicated students’ WTP is the fact that the “real” attribute is a catchall term for four distinct credence attributes—local, fair, humane, and ecologically sound. The fact that food could have any one of these credence attributes and be considered “real,” while not revealing which of the four attributes it contains, may lead some students to value the “real” attribute less. These students may value one or more of the four attributes more highly than others and may only be willing to pay a premium (or pay a greater premium) for those specific attributes. Future research could explore how students’ preference for the “real” attribute compares to their preference for each of the four individual attributes encompassed by it. Choice sets would be a more appropriate tool to capture this information than WTP. This type of comparison could inform the future of RFC movement or how individual campuses choose to promote the RFC.

As with any study relying on stated preference measures, such as contingent valuation, there is the risk of over or underestimating WTP. The small incremental increases in payment options presented to students in the CV question may cause “yea-saying” bias and thus increase the risk for value-action gap, or an inconsistency between the actual value students place on the “real” attribute and what they state they are willing to pay for it (Barr 2004). Our study did not identify protest bidders, or students who actually value the “real” attribute but state a zero WTP. It is possible that some students who value the “real” attribute believe it is the responsibility of the university or food service management company to internalize the additional cost of the real food, and, therefore, state their WTP as zero. Since we did not identify protest bidders, any potential protest bids were included in our analysis as zero bids and, therefore, may have caused a downward bias on our WTP estimations (Halstead et al. 1992). It may be useful for future research to explore protests bidders in this context, as it could inform how universities implement the RFC.

We were unable to capture students’ true, observed WTP for “real” food, because the RFC was relatively new to UVM at the time of this study and there was not yet any revealed preference data. Future research could be conducted to corroborate our stated preference data with revealed preference data. This could potentially be challenging, depending on how an institution structures its meal plans and “real” food offerings. For example, at UVM, the only way to measure “real” food preference would be at retail locations, which are frequented less by certain segments of the student population.

Conclusion

The purpose of this study was to characterize student preference for “real” food, as measured by students’ WTP for a meal plan consisting of at least 20% “real” food. We found that the majority of students are willing to pay a positive premium, though only about 20% are WTP a premium of 5% or more. Attitudes towards the price of food and the origin of food significantly influence preference for “real” food, as does enrollment in RUB and CALS, which may be proxy measures for attitude and/or education about the environment and food systems issues. Demographic characteristics, such as gender and residency, are also significant predictors, but class year is not. This study identified values and education to be the best leverage points for influencing student preference for “real” food.

This research is the first of its kind to explore student preference for the new “real” attribute promoted by the RFC. Evidence from the literature has demonstrated demand for the four attributes incorporated into the “real” attribute by general consumers and students alike. This study expands the literature by demonstrating that students also value the newly created “real” attribute as a catchall for the four separate attributes. Our finding that a majority of students have a preference for the “real” attribute has significant implications for the markets for more just and sustainable food; since students are the consumers of tomorrow, their preferences are critical to understand.

By leveraging the purchasing power of higher education institutions, the RFC has the potential to create significant market demand for “real” food and transform the food system. As a student-led movement, success of the RFC relies on student support, thereby making it critical to understand what factors influence students’ preference for the “real” attribute. Beyond demonstrating student preference for the new “real” attribute, this study identifies education and values as key leverage points for influencing student preference, thus contributing to the body of literature around developing markets for more just and sustainable food sources.

This study only examined preferences of undergraduate students at one mid-sized university in the northeastern United States. Our results are a useful first step in understanding the preferences of the general student population, but further research is needed to see if our results hold more widely. There are a number of challenges associated with evaluating student preference for “real” food using WTP, as detailed above in the “Discussion” section. Future research could improve upon the knowledge generated in this study by utilizing different methods to characterize student preference for the “real” attribute. For example, choice sets could be used to explore how students value the different aspects of the “real” attribute along with other attributes, such as price.

References

  1. Barlett, P.F. 2011. Campus sustainable food projects: critique and engagement. American Anthropologist 113 (1): 101–115.CrossRefGoogle Scholar
  2. Barr, S. 2004. Are we all environmentalists now? Rhetoric and reality in environmental action. Geoforum 35 (2): 231–249.CrossRefGoogle Scholar
  3. Breidert, C., Hahsler, M., and Reutterer, T. 2006. A review of methods for measuring willingness-to-pay. Innovative Marketing 2 (4): 8–32.Google Scholar
  4. Cameron, T. A., and Huppert, D. D. 1989. OLS versus ML estimation of non-market resource values with payment card interval data. Journal of Environmental Economics and Management 17 (3): 230–246.CrossRefGoogle Scholar
  5. Campbell, J., Dipietro, R. B., and Remar, D. 2014. Local foods in a university setting: price consciousness, product involvement, price/quality inference and consumer’s willingness-to-pay. International Journal of Hospitality Management 42: 39–49.CrossRefGoogle Scholar
  6. Carpio, C. E., and Isengildina-Massa, O. 2009. Consumer willingness to pay for locally grown products: the case of South Carolina. Agribusiness 25 (3): 412–426.CrossRefGoogle Scholar
  7. Clark University. 2010. The Sustainable University [class report]. http://www.clarku.edu/offices/campussustainability/pdfs/SustainableUFinal_Report_Dec2010.pdf. Accessed 25 Apr 2014.
  8. Feenstra, G., Allen, P., Hardesty, S., Ohmart, J., and Perez, J. 2011. Using a supply chain analysis to assess the sustainability of farm-to-Institution programs. Journal of Agriculture, Food Systems, and Community Development 1 (4): 1–16.CrossRefGoogle Scholar
  9. Giraud, K. L., Bond, C. A., and Bond, J. J. 2005. Consumer preferences for locally made specialty food products across northern New England. Agricultural and Resource Economics Review 34 (2): 204.CrossRefGoogle Scholar
  10. Halstead, J. M., Luloff, A. E., and Stevens, T. H. 1992. Protest bidders in contingent valuation. Northeastern Journal of Agricultural and Resource Economics 21 (2): 160–169.Google Scholar
  11. Jekanowski, M. D., Williams, D. R., and Schiek, W. A. 2000. Consumers’ willingness to purchase locally produced agricultural products: an analysis of an Indiana survey. Agricultural and Resource Economics Review 29 (1): 43–53.CrossRefGoogle Scholar
  12. Lancaster, K. J. 1966. A new approach to consumer theory. Journal of Political Economy 74 (2): 132–157.CrossRefGoogle Scholar
  13. Li, M., Wang, Q., and Kolodinsky, J. M. 2012. Estimating the optimal premium rates for credential food attributes: a case study in the Northeast United States. Journal of Food Distribution Research 43 (2): 51–63.Google Scholar
  14. Loureiro, M. L., and Hine, S. E. 2002. Discovering niche markets: a comparison of consumer willingness to pay for local (Colorado grown), organic, and GMO-free products. Journal of Agricultural and Applied Economics, 34(03).Google Scholar
  15. Loureiro, M. L., and Lotade, J. 2005. Do fair trade and eco-labels in coffee wake up the consumer conscience? Ecological Economics 53 (1): 129–138.CrossRefGoogle Scholar
  16. Loureiro, M. L., McCluskey, J. J., and Mittelhammer, R. C. 2003. Are stated preferences good predictors of market behavior? Land Economics 79 (1): 44–45.CrossRefGoogle Scholar
  17. Lusk, J. L., Nilsson, T., and Foster, K. 2007. Public preferences and private choices: effect of altruism and free riding on demand for environmentally certified pork. Environmental and Resource Economics, 36(4).Google Scholar
  18. Mahé, T. 2010. Are stated preferences confirmed by purchasing behaviors? The case of fair trade-certified bananas in Switzerland. Journal of Business Ethics 92 (2): 301–315.CrossRefGoogle Scholar
  19. Mitchell, R. C., and Carson, R. T. 1989. Using surveys to value public goods: the contingent valuation method. New York, NY: Resources for the Future.Google Scholar
  20. Moser, R., Raffaelli, R., and Thilmany-McFadden, D. 2011. Consumer preference for fruit and vegetables with credence-based attributes: a review. International Food and Agribusiness Management Review 14 (2): 121–141.Google Scholar
  21. Naald, B. V., and Cameron, T. A. 2011. Willingness to pay for other species’ well being. Ecological Economics 70 (7): 1325–1335.CrossRefGoogle Scholar
  22. Onianwa, O., Wheelock, G., and Mojica, M. 2005. An analysis of the determinants of farmer-to-consumer direct-market shoppers. Journal of Food Distribution Research 36 (1): 130–134.Google Scholar
  23. Onozaka, Y., and McFadden, D. T. 2011. Does local labeling complement or compete with other sustainable labels? A conjoint analysis of direct and joint values for fresh produce claim. American Journal of Agricultural Economics 93 (3): 689–702.CrossRefGoogle Scholar
  24. Steel, A. 2012. A critical mass for real food. Yes! Magazine. http://www.yesmagazine.org/people-power/a-critical-mass-for-real-food. Accessed 16 Jan 2015.
  25. The Real Food Challenge. n.d. http://www.realfoodchallenge.org. Accessed 16 Jan 2015.
  26. The University of Vermont. n.d. Real food calculator. http://www.uvm.edu/foodsystems/?Page=realfoodcalculator.html&SM=realfoodmenu.html. Accessed 27 Jan 2015.
  27. Thilmany, D., Bond, C. A., and Bond, J. K. 2008. Going local: exploring consumer behavior and motivations for direct food purchases. American Journal of Agricultural Economics 90 (5): 1303–1309.CrossRefGoogle Scholar
  28. Tonsor, G. T., Olynk, N., and Wolf, C. 2009. Consumer preferences for animal welfare attributes: the case of gestation crates. Journal of Agricultural and Applied Economics 41 (3): 713–730.CrossRefGoogle Scholar
  29. Trudel, R., and Cotte, J. 2009. Does it pay to be good? MIT Sloan Management Review 50 (2): 61–68.Google Scholar
  30. U.S. Department of Agriculture, National Agricultural Statistics Service. 2007. Market Value of Agricultural Products Sold Including Direct Sales: 2007 and 2002. http://www.agcensus.usda.gov/Publications/2007/Full_Report/Volume_1,_Chapter_2_US_State_Level/. Accessed 16 Jan 2015.
  31. U.S. Department of Agriculture, National Agricultural Statistics Service. 2008. Farms, Land Use, and Sales of Organically Produced Commodities on Certified and Exempt Organic Farms: 2008. http://www.agcensus.usda.gov/Publications/2007/Online_Highlights/Organics/. Accessed 16 Jan 2015.
  32. Vermont Sustainable Jobs Fund. 2013. Farm to Plate Strategic Plan. http://www.vtfoodatlas.com/plan/. Accessed 16 Jan 2015.
  33. Zepeda, L., and Li, J. 2006. Who buys local food? Journal of Food Distribution Research 37 (3): 1–11.Google Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2017

Authors and Affiliations

  • Jennifer Porter
    • 1
  • David Conner
    • 1
  • Jane Kolodinsky
    • 1
  • Amy Trubek
    • 2
  1. 1.Department of Community Development and Applied EconomicsThe University of VermontBurlingtonUSA
  2. 2.Department of Nutrition & Food SciencesBurlingtonUSA

Personalised recommendations