Expressing Individuality via Food Choices Amanda Simpson Weaver and Jayson Lusk* Abstract A new means to the end of expressing one’s identity or individuality has grown in popularity in recent years; food is much more to consumers than the basic physiological needs of food. Consumers have diversified into a wide range of food personality types with different perceptions of the role food should play in their lives. This paper uses factor analysis and compares these food personality factors with food attributes factors consisting of non-price features of food products. Results show that identity is expressed via food at differing levels and income level does have some influence. Keywords: identity, food personality, food attributes, hierarchy of needs *Amanda Simpson Weaver is a doctoral student and Jayson Lusk is a Regents Professor and Willard Sparks Endowed Chair at Oklahoma State University in the Department of Agricultural Economics. Introduction While the general population has become more and more segregated from agriculture, the food market has rapidly evolved. Food consumed away from home has increased in both calories consumed and total food expenditures (Stewart et al. 2011; Stewart et al. 2004; Lin, Frazao, and Guthrie 1999), and food consumed at home has changed its meaning. Home cooked meals are becoming less common, and the microwave is the new cook of the family offering the meal of choice for each member at the time they wish. Needless to say, the primary shopper of the household has a totally different cart of food than before; from choosing foods that everyone will eat to choosing foods individually for each member. Eating family style where the dishes are passed around the table is being replaced with single-serving, convenient foods. This could be seen as agribusinesses bringing some of the appeal of restaurants into the home: variety to please everyone. This allows for individuality in food choice, which uniqueness and nonconformity has become a social and cultural norm in American society. Therefore, food markets must continue to adapt to the changing market as consumers use food choices to express their individuality. How do consumers express individuality through food? Bell et al. (1981) find personal food beliefs dictate the foods that consumers select when analyzing six subjects with very different “food habits” such as semi-vegetarian, European diet, average Australian diet, lacto-ovo vegetarian, health food, and fast food (pg. 22). They preliminarily found that the complex food beliefs between their subjects could be factored into three variables to explain a majority of the variation. Alternatively, another study looked at food choice through food motives – health, weight concern, pleasure and identity – among women using personality variables (Lindeman and Stark 1999). The expression of identity motive was “best predicted by vegetarianism, magical beliefs about food and health, and personal strivings for ecological welfare and for understanding self and the world” (pg. 141). Thus, these studies suggest that beliefs about different types of food and what the food can offer to achieve supplementary goals are supposedly becoming more important to consumers food decisions. Over a decade ago, Antle (1999) described in his Presidential Address to the American Agricultural Economics Association (AAEA) that demand was no longer just a function of own price and 1 prices of other goods, income, and population, but also demographics of that population and other nonprice attributes of the good. Nonprice attributes he listed are nutrition, convenience, safety, where the food is grown, naturalness (produced without modern technologies), and environmental impact. In addition, taste has been cited by many to be the top food quality of interest when choosing food products (Senauer 2001; Glanz et al. 1998; Demeritt 2002). The portion of the population who view food as a means to satiate hunger and thirst alone is declining as society becomes further removed from agriculture; food is now fulfilling much more for consumers through these nonprice attributes. Considering the original Maslow’s hierarchy of needs, food is included in the base of the triangle as one of the main physiological needs for survival (Maslow and Lewis 1987), and moving towards the top of the pyramid are safety, love/belonging, esteem and selfactualization. In 1997, Belonax created a hierarchy of needs in relation to food as can be seen in Figure 1, which illustrates how consumers express themselves through the food they consume. Senauer (2001) describes how in the past, demand was almost a linear function of prices, income and a few demographics. Now, he warns that while low income households will be motivated to meet physical needs of food, it would be incorrect to assume someone moves or does not move up the hierarchy based on income alone. Food is a trend that is attracting attention from all ages, and instead of clothing or traveling, the new way to exhibit trendiness is foodiness or keeping up with the latest food trends. “What you eat for dinner has become the definitive marker of social status” (Miller 2010, pg. 44). The way people consume food is widely varied for a potentially endless number of reasons or combinations of reasons and will continue to change and branch off. Senauer (2001, pg. 5) states that “demand has shifted toward high-quality, processed and prepared foods that provide convenience and other desired attributes,” but the trend in food consumption more than ten years later, is pushing consumers to steer away from the center aisles of the supermarket or the supermarket altogether. Some use social pressure to influence people to steer away from the fast food lifestyle and promote consumers to spend more time on food planning, shopping, preparing, cooking and consuming or enjoying food. One example of the trendy food movement is a TEDx event in Manhattan on February 16th, 2013 (Changing 2 the Way We Eat 2013). It is titled “Changing the Way We Eat,” and they advocate ten steps to change food habits: getting educated about the food industry where they cite books like The Omnivore’s Dilemma by Michael Pollan, shopping sustainable, asking questions, eating less meat, eating seasonal foods, growing your own food, cooking, drinking local water, getting involved in campaigns, and enjoying meals together. Here the antagonist is the retailer and big agribusiness; they claim to be a proponent of the farmer. Then, there is the Slow Food Movement, which is a direct response to the convenient fast food lifestyle. They promote that slow food is not just about food; it is a lifestyle that links food consumption to social, ethical, political, environmental and spiritual elements (Andrews 2008). The current back to the roots movement may or may not decrease the overall level of time spent on food decisions. Nevertheless, some consumers are not only changing the types of food they eat, but they are also devoting much more time and energy towards what they eat. However, it is worth noting that while some consumers continue to follow the new trends, others may pick and choose which trends adhere to their beliefs. Since food is commonly given the role of expressing identity, none have looked at how particular food attributes increase the probability that an individual has a particular set of what we call food personality. This paper aims to define and cluster food personalities and find the impact of food qualities on each personality type. It is important for food manufacturers and agribusinesses to know what attributes are desired by their target market and for research to continue to study the evolving food consumer. Theory Antle (1999) describes that estimating consumer demand is no longer a simple function of income and price of the good, where price is a function of the quality. Consumer food demand now incorporates other nonprice attributes as a result of the opportunity cost of time involved with food consumption. Thus, our consumer demand model resembles that of Antle’s new economic model for the demand-side: ππ· = π·(π·, πΌ, π, πΆ, π) 3 where ππ· represents the demand for good π₯ and is a function of π or a vector of prices of all goods, πΌ income, π the population of interest, πΆ or characteristics of the consumer population, and π or the nonprice attributes of the good. In our study, the population of interest is the United States and characteristics of this population, which could incorporate a number of issues that influence preferences, but we included food related habits such as the number of times eating fast food per week, hours exercised per week, etc. Nonprice attributes corresponds several nonprice attributes demanded in the food market today, and in nutritional content, convenience, safety, where and how good is produced, environmental attributes in production process, and the process or inputs of production. This corresponds with seven out of nine price and nonprice attributes chosen to include in our study: price, nutrition, convenience, safety, where the food is grown, naturalness (produced without modern technologies), and environmental impact. Taste and appearance were included as well, which could also be considered as hedonic attributes. Considering the Maslow’s hierarchy of needs in relation to food, we aim to sort these nonprice attributes into a similar pyramid of food needs. Senauer (2001) states that income levels motivates the level that individuals sit on this hierarchy, and that as income increases, society tends to use food to fill needs higher up in the pyramid. We hypothesize that income will not be the only variable that effects the desirability of upper hierarchy levels. Valuation of food attributes can also reveal how important it may be to individuals to fulfill needs other than physiological needs. Also, if an individual highly ranks upper levels in the hierarchy, we still expect they value the bottom tier or physiological needs. However, when choosing a high level over the bottom level, they may assume that meeting physiological needs is given and not an issue. Data Given our interest in food perceptions and food attribute rankings, an internet survey was chosen to collect primary data. A random sample of the United States population was sought through Qualtrics, an online research survey manager. The site solicits self-registered participants who meet the qualifications of the targeted population via email, usually with a monetary incentive. Here, they were given a $1.50 4 payment into their account that can be used for gift certificates or a prepaid debit card, and the individuals can only fill out a survey every 10 days, which reduces problems with professional respondents. This study acquired 774 observations in March 2011. While the exact response rate could be higher due to a number of reasons including spam filters keeping the survey from getting to potential panelists or minimal usage of the email account, the rate available to report is 13.5% completing the survey. While this may be a lower response rate and a high response rate would be preferred to reduce the chance of sample selection bias, the means of demographics are compared to the US population via the Current Population Survey in Table 1. As can be observed, the sample is similar to the US population, which provides assurance of a random sample. Even so, a weight was employed using iterative proportional fitting techniques (Izrael, Hoaglin, and Battaglia 2004, 2000), which made the sample means equal to the US Census data in terms of age, education, gender, ethnicity, US Census region, and income level. The questionnaire first lists a series of statements on “Food Personality,” which is in Figure 2, and they are asked to state their level of agreement on a Likert scale coded from 1 through 5, or strongly agree to strongly disagree. The second question of interest asks the respondent to place the nine food attributes into three categories: most important, somewhat important, and least important. The survey forced the respondent to place just three attributes per category. This eliminates respondents from being able to deem all categories as important, though future research could allow for more flexibility or only require at least one per category. Finally, a number of questions were asked that relate to food consumption like if the respondent is the primary shopper of the household, how often they eat fast food, if they have ever received food stamps, number of hours spent exercising per week, etc. The question on food stamps is comparable to a question in the National Health and Nutrition Examination Survey (NHANES), which asks if the participant has ever received food stamp benefits. The average in 20072008 was approximately 25%, which is close to the 29% average found in the survey. The means of food personality values, average food attribute rank, factors of both, food stamp participation, and percent who have a college degree are found in Table 2. 5 Procedures Before setting up a regression model, combining some of the variables into common factors was utilized by factor analysis for both food personality variables and food attribute variables. Tables 3 and 4, respectively, show the results of the factor analysis in SAS. For the food personality variables, the factors are meant to signify the same sort of increasing order as Maslow’s Hierarchy of Needs with food personality types. So the “Bottom Tier” is represented by factor 2 which contains necessity and fuel. Then factor 3 contains comforting and enemy, which we will consider to be the “Middle Tier” needs. They are beyond the physical needs, but still physically related. Factor 1 contains the “Top Tier” needs or symbol, passion, social and sacred – none of which meet physiological or physical needs, but needs of selfactualization or expression of identity. The food attribute variables are factored into three categories as well, where Factor 3 contains price and safety. We will call this factor “Necessity” attributes because those two factors meet two of the most basic or necessary food needs. Then, factor 2 contains nutrition, convenience, and appearance which we will call “Value-Added” because these all contribute to the individuals benefit in health and time. Finally, Factor 1 contains naturalness, grown, environment, and taste which we give the name “Luxury” in that the first three do not contribute to the food consumption experience directly, which explains the inverse relationship of the three with taste. So within each factor, the respondent likely ranked the positive (negative) attributes in a higher/lower importance level. Thus, if taste was highly rated, the other three were not as highly rated and vice versa. Now that the factors are defined, they can be created by taking the average of the variables involved in each factor of food personality variables and food attribute variables. For example, the food personality factor “Bottom Tier” was created by taking the average of both “food is fuel” and “food is necessity” ratings. The same process of finding the average of the variables involved in each factor was executed for the food attribute variables, though the inverse relationship was taken into account as well. For example, to find the Necessity factor, safety was subtracted from price to find the average. Without subtracting, the inverse relationship between each factor is not taken into account. 6 Discussing the difference between the food personality and food attribute factors, one is a characteristic of the person’s perception of the importance of an attribute and the other is the person’s perception of themselves as it relates to food. Thus, we hypothesize that in a grocery aisle, consumers in the necessity category are likely to base their choice on the lower price tag or the brand they trust to be safer; they are inversely correlated so giving one high importance means the other is not as important. This consumer could be someone with low income or any income level if they do not wish to fulfill other needs beyond food consumption. They potentially spend more time on food by not spending more money on convenient products and fast food; they prepare more meals at home, spend more time searching for cheaper item like calculating cost per unit, etc. The value-added attributes are likely to be chosen by consumers who base their choice on their needs being met on the nutrition panel or the way the food or packaging of the food looks, which could enhance the perceived taste. However, there is a negative connection between nutrition and appearance and the convenience of the food or the time required to prepare/level of difficulty to prepare. Thus, if an individual highly rates convenience and appearance, they are likely to rate nutrition lower. With a high ranking of convenience, they could be indirectly demanding more time available or less time spent preparing food or exercising to burn excess calories could be some deeper reasoning. They likely spend the least amount of time on food, i.e. more meals away from home or brought home and more packaged food from the store. Considering the positive correlation with appearance, food packaging could be an interpretation of appearance, and an appealing package or trendy logo could be associated with a newer product that is easy to prepare. With a negative relationship between nutrition and convenience, this could be from a mostly true observation that nutritious foods are not convenient foods, though there are always exceptions. Finally, consumers who highly rate the luxury attributes are likely to base their choice on the taste of the food or labels claiming natural, where the food is grown and its environmental impact. Thus, if taste is preferred, naturalness, where the food is grown, and environmental impact are not as highly ranked and vice versa. They are likely to spend more time on food preparation and/or consumption via 7 purchasing or eating foods they know appease their taste buds or seeking foods with more labels. We hypothesize that the top tier food personality variables – passion, social, symbol, sacred – would increase with this factor. Using the common factors, we will run a regression to find the effect of the food attribute levels and other demographics on the food personality factors: (1) ππππ = πΌ0 + πΌ1π π¨ππ + πΌ2 π·π + πΌ3 ππ + πΌ4 π¬πΏπ + πΌ5 ππ + πΌ6 πΊπ + πΌ7 π°ππ + πΌ8 πΈπ + ππππ where π is the food personality factor i where π = 1, 2, 3 tiers or bottom, middle, and top tiers of the hierarchy, π΄ is a vector of food attribute factors where π = necessity, value-added, and luxury food attributes, π·π is a vector for primary shopper in the household where ππ = 0%, 50%, or 100%, ππ is a vector for how many times the individual consumed fast food in a week where ππ = 0, 1.5, 3.5 or 5, π¬πΏπ is a vector for how many hours the individual exercised per week where π¬πΏπ = 0, 1.5, 3.5, 5.5, or 7, ππ is an indicator variable for if ever being a participant of the food stamp program, πΊπ is an indicator variable for if female, πΌππ is a vector of annual household incomes where π = less than $20,000, $20,000-$39,999, $40,000-$59,999, $60,000-$79,999, $80,000-$99,999, and $100,000 or more, πΈπ is an indicator variable for an individual having a Bachelor’s degree or not, and ππππ is an assumed independent error term which is ππππ ~π(π, ππ2 ). Results Some interesting results, first of all, were found just in comparing means overall and between groups. For the most part, the survey respondents follow previous research by placing taste as the number one overall attribute as shown in Table 5, although those who said they have an immediate family member who farms or ranches placed nutrition and price over taste. Then in Table 6, the averages of the traits are listed in ranking order of the highest rating of importance for each personality statement. These averages are interesting to note when considering that the traits are factored into tiers that resemble Maslow’s Hierarchy of Needs. “Social” was the top rated food personality trait which is also a part of the top tier of 8 the hierarchy. The inclination that food is a means towards other ends rather than just physiological needs is affirmed in these results. There is a significant difference in the means for two of the attributes and two personality statements between two groups concerning food stamps as can be observed in Table 7: convenience, taste, necessity and sacred. If an individual was ever on food stamps, they value convenience and taste less, and they tend to agree more with statements that food is a necessity and food is sacred. Why an individual would agree more with the statement that “food is sacred: I eat the way nature intended” is unknown other than interpretation of the statement. Nevertheless, ever having been enrolled in the food stamp program did leave an impression on certain food perceptions and the desirability of a couple food attributes. The regression results in Table 8 provided an interesting picture of what variables will increase or decrease the level of agreement with our food personality groups. As preference for Value-Added attributes increases, the rating of the top level food personality traits decreases, which tells us that those who believe food is their passion, social, sacred or a symbol do not care as much about nutrition, convenience, or appearance of the food. There are a few variables that do increase the agreement level with this top level including if the individual is the primary shopper in the household and how often they exercise, both of which as positively associated. Variables that have a positive influence on the top tier could mean that these individuals spend more time on food choices or compensating for food choices. But then, as they eat more fast food per week, this also increases the agreement level for the top tier, which contradicts their lower ranking of Value-Added attributes. This may be an indicator that these individuals prefer to avoid convenient, fast food, but in reality, they still consume it due to other factors like income. At the other extreme, the bottom tier of the hierarchy is positively influenced by necessity food attributes and low income levels as hypothesized. This suggests that a preference for price and safety increases the agreement level of the bottom tier personality which is the combination of food is fuel and food is a necessity statements. Females and a preference for luxury food items negatively effects the agreement level of the bottom tier as well, meaning females look for more out of food than the basic physiological needs more than males and luxury foods are not desired by those who utilize food for the 9 physiological needs alone. Finally, the middle level is most effected by low income levels as well. Interestingly, necessity food attributes increases the rating of the middle tier, indicating that price or safety are important to this tier who believe food is comforting and my enemy. Could this be an effect of cheap (comforting) junk food? Fast food consumers are also more likely to agree with the statements that food is comforting and food is my enemy, which is somewhat of an expected response. Conclusion The results in this study are quite revealing for the food industry and should provide valuable marketing insight for agribusinesses. For food retailers their primary shoppers are individuals who value foods role in their social lives, believe food is a symbol of who they are, and food is their passion and sacred to them. These factors are much more relevant to primary shoppers than diet or comfort foods that are associated with the middle tier, as well as the most economical and basic food choices associated with the bottom tier. For companies appealing to lower income levels, the top priorities for those individuals is price and safety, and they especially do not like what is considered to be a comfort or diet food. For fast food restaurants, their consumers are more likely to agree with the middle tier or prefer comfort food as well as healthier options, but they also agree with top tier personality traits, thus their choice of food or even restaurant tells something about their identity. With a high rating of food is my passion, their consumers may prefer variety. The sacred trait may lead them to prefer seasonal food items. For agribusinesses targeting consumers who want comfort food and diet foods, also targeting low income areas may not be successful. Over 77% of respondents claimed to be the primary shopper in their household, and this group is especially important not to overlook for food retail markets. These individuals are choosing and purchasing food items that satisfy their luxury tastes. Therefore, for these shoppers to buy something for their family that does not reflect their status or needs in the upper tiers, it would take direct requests from their loved ones for the specific item or be a frequent, expected purchase. Thus, appealing to the primary shopper as well must be taken into account when targeting other groups. Considering that females account for 57% of the primary shopper group in the survey, items for men and children need to have those 10 attributes that women would be willing to purchase. One thing is certain about all primary shoppers though; they describe food as their passion significantly more than non-primary shoppers. This is the one variable we found to be significantly different between the two groups, so food items that make the shopper feel like a food connoisseur would be a step in the right direction. With the increasing number of options that consumers have at the supermarket and in restaurant selection, types of consumers are branching off as well. Further research could look at the personality traits for each food industry entity closer to identify types of restaurants that appeal to the different personality traits. Also, including more possible personality variables could be insightful in breaking down food personality classification further, and including more specific explanation for each could be important to ensure interpretation is the same across individuals. Food personality types are becoming more apparent and will continue to evolve over time. 11 References Andrews, G. 2008. The Slow Food Story: Politics and Pleasure: McGill Queens University Press. Antle, John M. 1999. "The New Economics of Agriculture." American Journal of Agricultural Economics 81 (5):993-1010. Bell, A Chris, A. M. Stewart, A. J. Radford, and P. T. Cairney. 1981. "A Method for Describing Food Beliefs Which May Predict Personal Food Choice." Journal of Nutrition Education 13 (1):22-26. Belonax, Joseph J. 1997. Food Marketing. Needham Heights, Massachuesetts: Simon & Schuster Custom Publishing. Changing the Way We Eat. February 16 2013. Available from http://tedxmanhattan.org/change-the-wayyou-eat/. Demeritt, L. 2002. All Things Organic 2002: A Look at the Organic Consumer. Bellevue, WA: The Hartman Group. Glanz, Karen, M. Basil, E. Maibach, J. Goldberg, and D. Snyder. 1998. "Why Americans Eat What They Do: Taste, Nutrition, Cost, Convenience, and Weight Control Concerns as Influences on Food Consumption." Journal of the American Dietetic Association 98 (10):1118-1126. Izrael, David, David C Hoaglin, and Michael P Battaglia. 2000. A SAS Macro for Balancing a Weighted Sample. Paper presented at Twenty-Fifth Annual SAS Users Group International Conference. ———. 2004. To Rake or Not to Rake is Not the Question Anymore with the Enhanced Raking Macro. Paper presented at SUGI Conference, May 2004, at Montreal, Canada. Lin, Biing-Hwan, Elizabeth Frazao, and Joanne F Guthrie. 1999. Away-from-Home Foods Increasingly Important to Quality of American Diet. United States Department of Agriculture, Economic Research Service, Lindeman, M., and K. Stark. 1999. "Pleasure, Pursuit of Health or Negotiation of Identity? Personality Correlates of Food Choice Motives Among Young and Middle-Aged Women." Appetite 33 (1):141-161. 12 Maslow, A.H., and K.J. Lewis. 1987. Maslow's Hierarchy of Needs: Salenger Incorporated. Miller, L. 2010. Divided We Eat. Newsweek, November 29, 43-48. Senauer, B. 2001. "The Food Consumer in the 21st Century: New Research Perspectives." Working Paper: Retail Food Industry Center, University of Minnesota. Stewart, H., N. Blisard, S. Bhuyan, and R. M. Nayga Jr. 2004. The Demand for Food Away from Home. In Agricultural Economic Report, edited by US Department of Agriculture Economic Research Service. Stewart, H., J.L. Lusk, J. Roosen, and J.F. Shogren. 2011. "Food Away from Home." The Oxford Handbook of the Economics of Food Consumption and Policy 1: 647-666. 13 Table 1. Demographics of Survey Sample Compared to the United States (US Census Bureau 2010) Category U.S. Census Sample Age 18-34 years old 30.71% 30.36% 35-44 years old 17.64% 18.86% 45-54 years old 19.36% 20.54% 55-64 years old 15.44% 15.37% 65 years or older 16.84% 14.86% Education Some college, high school degree or less Bachelor's degree Graduate degree (MS, MA, MBA, PhD, MD, or JD) 79.44% 13.58% 6.99% 62.98% 21.59% 15.42% Gender Female Male 49.13% 50.87% 51.16% 48.84% Ethnicity White Black or African American Hispanic Other 64.89% 12.14% 16.07% 6.90% 84.29% 6.71% 3.43% 5.57% Location Northeast U.S Census region Midwest U.S Census region South U.S Census region West U.S Census region 17.96% 21.72% 36.91% 23.41% 19.69% 21.54% 32.38% 26.39% Annual household income Less than $20,000 $20,000 to $39,999 $40,000 to $59,999 $60,000 to $79,999 $80,000 to $99,999 $100,000 or more 14.06% 19.72% 17.05% 14.00% 10.17% 25.01% 14.47% 26.74% 19.25% 14.73% 10.34% 14.47% Number of times exercising per week None 1-2 hours per week 3-4 hours per week 5-6 hours per week 7 or more hours per week N/A N/A N/A N/A N/A 22.11% 32.12% 23.54% 13.26% 8.97% Number of times eating fast food per week None 1-2 meals per week 3-4 meals per week 5 or more meals per week N/A N/A N/A N/A 26.27% 60.26% 11.11% 2.35% Primary shopper of the household Yes No Equally shared N/A N/A N/A 77.75% 5.63% 16.62% 14 Table 2. Summary Statistics on Food Attribute and Food Personality Variables Variable N Mean Std Dev Food is a… Necessity 774 2.970 1.143 Fuel 774 3.337 1.054 Symbol 774 2.840 1.149 Comforting 772 3.272 1.091 Enemy 770 2.299 1.163 Passion 774 3.124 1.109 Social 772 3.299 1.116 Sacred 771 2.754 1.064 Food Personality Factors Bottom (necessity, fuel) Middle (comforting, enemy) Top (symbol, passion, social, sacred) 774 768 769 Food Attributes (1=Least Important, 3=Most Important) Price 774 Naturalness 774 Nutrition 774 Grown 774 Safety 774 Environmental impact 774 Appearance 774 Convenience 774 Taste 774 Minimum Maximum 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 5.00 5.00 5.00 5.00 5.00 5.00 5.00 5.00 3.154 2.788 3.005 0.945 0.907 0.801 1.00 1.00 1.00 5.00 5.00 5.00 2.530 1.797 2.545 1.598 2.202 1.399 1.704 1.640 2.585 0.691 0.786 0.606 0.694 0.720 0.595 0.700 0.709 0.675 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 Food Attribute Factors Necessity (price, safety) Value-added (nutrition, convenience, appearance) Luxury (natural, grown, environment, taste) 774 774 774 -0.164 -0.266 0.552 0.561 0.470 0.469 -1.00 -1.67 0.00 1.00 0.33 2.00 Food stamp participation (1 = Yes) Education Indicator (1 = College degree) Weight 771 774 774 0.292 0.367 1.000 0.455 0.482 0.873 0.00 0.00 0.14 1.00 1.00 7.26 15 Table 3. Factor Analysis for Food Personality Attributes Factor 1 – Top Tier Factor 2 - Bottom Tier Necessity -4 60 * Fuel 6 61 * Symbol 41 * 10 Comforting 18 0 Enemy -7 2 Passion 68 * -8 Social 57 * -9 Sacred 59 * 16 Factor 3 – Middle Tier 3 0 19 50 * 48 * -5 18 -11 Printed values are multiplied by 100 and rounded to the nearest integer. Values greater than 0.4 are flagged by an '*'. Table 4. Factor Analysis for Food Attributes Attributes Factor 1 - Luxury Price -35 Naturalness 66 * Nutrition -28 Grown 69 * Safety -13 Environment 69 * Convenience -26 Taste -67 * Appearance -36 Factor 2 - Value Added -2 27 92 * -3 3 -9 -57 * 5 -50 * Factor 3 – Necessity -47 * -18 -4 -6 95 * 13 -27 -6 1 Printed values are multiplied by 100 and rounded to the nearest integer. Values greater than 0.4 are flagged by an '*'. Table 5. Food Attribute Average Rankings Rank 1. 2. 3. 4. 5. 6. 7. 8. 9. Attribute Taste Nutrition Price Safe Naturalness Appearance Convenience Grown Environmental Average 2.59 2.55 2.53 2.20 1.80 1.70 1.64 1.60 1.40 Family not in Agriculture Taste 2.62 Nutrition 2.54 Price 2.53 Safety 2.21 Naturalness 1.77 Appearance 1.71 Convenience 1.65 Grown 1.57 Environment 1.39 16 Family in Agriculture Nutrition Price Taste Safety Naturalness Grown Appearance Convenience Environment 2.56 2.49 2.31 2.17 2.01 1.83 1.63 1.53 1.47 Table 6. Food Personality Traits Average Rating of Importance Trait Social Fuel Comforting My Passion A Necessity A Symbol Sacred My Enemy Average 3.31 3.30 3.28 3.11 2.92 2.81 2.73 2.34 Table 7. Difference in the Mean Rank of Food Attributes and Level of Agreement with Food Personality Statements Not on Food Stamps (N=546) On Food Stamps (N=225) Mean Standard Mean Standard Deviation Deviation Variable Difference p-value Price 2.5183 0.7095 2.5644 0.6385 0.0461 0.3788 Naturalness 1.7857 0.7757 1.8178 0.8060 0.0321 0.6100 Nutrition 2.5421 0.6020 2.5556 0.6180 0.0134 0.7794 Grown 1.5678 0.6880 1.6533 0.6910 0.0856 0.1164 Safety 2.1722 0.7114 2.2711 0.7394 0.0989 0.0872 Environment 1.3993 0.6102 1.4000 0.5590 0.0007 0.9920 1.6722 0.7223 1.5644 0.6725 -0.1077 Convenience 0.0478 2.6355 0.6331 2.4711 0.7502 -0.1644 Taste 0.0038 Appearance 1.7070 0.6948 1.7022 0.7167 -0.0047 0.9362 2.9029 1.1037 3.1333 1.2247 0.2304 Necessity 0.0146 Fuel 3.2985 1.0337 3.4267 1.1041 0.1281 0.1362 Symbol 2.8315 1.1337 2.8489 1.1894 0.0174 0.8494 Comforting 3.3088 1.0755 3.1867 1.1263 -0.1222 0.1646 Enemy 2.2813 1.1338 2.3318 1.2329 0.0506 0.5962 Passion 3.0861 1.0895 3.2133 1.1568 0.1273 0.1586 Social 3.3051 1.0902 3.2889 1.1806 -0.0163 0.8572 2.6943 1.0338 2.8889 1.1264 0.1946 Sacred 0.0258 17 Table 8. Regression Results for Each Food Personality Factor Level as Dependent Variable Bottom Level Middle Level Top Level Parameter Parameter Parameter Variable Estimate p-value Estimate p-value Estimate p-value Intercept <.0001 <.0001 <.0001 2.5999 2.5979 2.5221 Necessity 0.0219 0.0377 -0.0254 0.6339 0.1481 0.1295 Value-added -0.1490 0.0694 -0.1354 0.0875 0.0023 -0.2069 Luxury 0.0022 -0.0058 0.9400 -0.1264 0.0565 -0.2457 Primary shopper 0.2162 0.0856 0.1925 0.1130 <.0001 0.5176 Fast food 0.0004 0.9891 <.0001 0.0007 0.1391 0.0907 Exercise -0.0002 0.9907 -0.0137 0.3720 0.0008 0.0441 Food stamp 0.0475 0.5588 0.0209 0.7901 0.0981 0.1443 female 0.0004 0.0130 0.8542 -0.0947 0.1183 -0.2592 Less than $20,000 0.0003 0.0170 -0.1892 0.0708 0.4556 -0.2923 $20,000 to $39,999 <.0001 0.0296 0.0013 0.4744 -0.2351 -0.2976 $40,000 to $59,999 0.0054 0.0012 -0.1585 0.0883 0.3135 -0.3523 $60,000 to $79,999 0.1396 0.2403 -0.2151 0.0611 -0.1878 0.0560 $80,000 to $99,999 0.1868 0.1466 -0.0258 0.8353 0.0153 -0.2580 Education indicator 0.1472 0.0834 0.0601 0.4641 0.0852 0.2246 **Coefficients in bold are significant at the 5% level. Table 9. Difference in the Mean Rank of Food Attributes and Level of Agreement with Food Personality Statements Not Primary Equally Shared Primary Shopper Shopper (N=43) Shopper (N=127) (N=595) Variable Mean Std Dev Mean Std Dev Mean Std Dev Differencea p-value Price 2.604651 0.583076 2.582677 0.659978 2.509244 0.706749 -0.0954 0.3078 Naturalness 1.651163 0.719911 1.692913 0.771638 1.836975 0.792629 0.1858 0.1052 Nutrition 2.511628 0.592497 2.582677 0.569613 2.539496 0.616581 0.0279 0.7642 Grown 1.627907 0.690868 1.496063 0.653125 1.620168 0.703716 -0.0077 0.9442 Safety 2.279070 0.766115 2.259843 0.703874 2.179832 0.722134 -0.0992 0.4122 Environment 1.441861 0.665558 1.354331 0.556742 1.408403 0.599885 -0.0335 0.7490 Convenience 1.604651 0.694863 1.677165 0.677084 1.630252 0.713807 0.0256 0.8180 Taste 2.511628 0.767559 2.661417 0.632712 2.574790 0.675973 0.0632 0.5962 Appearance 1.767442 0.750784 1.692913 0.660829 1.700840 0.704423 -0.0666 0.5754 Necessity 2.976744 1.034831 2.858268 1.059533 2.973109 1.163104 -0.0036 0.9840 Fuel 3.209302 1.013202 3.299213 1.018207 3.349580 1.067808 0.1403 0.3844 Symbol 2.581395 1.199852 2.700787 1.018207 2.884034 1.172782 0.3026 0.1118 Comforting 3.093023 0.995561 3.212598 1.081078 3.295110 1.100795 0.2021 0.2006 Enemy 2.142857 1.049307 2.373016 1.078771 2.293423 1.186171 0.1506 0.3682 2.697674 1.145070 3.062992 1.036998 3.171429 1.115029 0.4738 0.0086 Passion Social 3.046512 1.068009 3.228347 1.017225 3.332209 1.137175 0.2857 0.0910 Sacred 2.581395 0.851681 2.669291 0.976348 2.788851 1.093440 0.2075 0.1310 a. Difference is the difference between Primary Shopper and Not Primary Shopper 18 Figure 1. Maslow’s Hierarchy of Needs in Relation to Food (Belonax 1997) Figure 2. Food Personality Question 19