- Email: [email protected]
College Students Must Overcome Barriers to Use Calorie Labels in Fast-Food Restaurants
Research Article College Students Must Overcome Barriers to Use Calorie Labels in Fast-Food Restaurants Kimberly A. Stran, PhD, RD1; Linda L. Knol, PhD, RD1; Lori W. Turner, PhD, RD2; Kimberly Severt, PhD1; Debra M. McCallum, PhD3; Jeannine C. Lawrence, PhD, RD1 ABSTRACT Objective: To explore predictors of intention of college students to use calorie labels on fast-food menus and differences in calories ordered after viewing calorie information. Design: Quasi-experimental design. Participants selected a meal from a menu without calorie labels, selected a meal from the same menu with calorie labels, and completed a survey that assessed demographics, dietary habits, Theory of Planned Behavior constructs, and potential barriers to use of calorie labeling. Setting: A southern university. Participants: Undergraduate university students (n ¼ 97). Main Outcome Measures: Predictors of intention to use calorie labels and whether calories selected from the nonlabeled menu differed from the labeled menu. Analysis: Confirmatory factor analysis, exploratory factor analysis, multiple regression, and paired t tests. Results: Participants ordered significantly fewer calories (P ¼ .02) when selecting from the labeled menu vs the menu without labels. Attitudes (P ¼ .006), subjective norms (P < .001), and perceived behavioral control (P ¼ .01) predicted intention to use calorie information but did not predict a difference in the calories ordered. Hunger (P ¼ .03) and cost (P ¼ .04) were barriers to using the calorie information. Conclusions and Implications: If students can overcome barriers, calorie labeling could provide information that college students need to select lower-calorie items at fast-food restaurants. Key Words: nutrition labeling, consumer health information, restaurants, health policy (J Nutr Educ Behav. 2015;-:1-9.) Accepted September 22, 2015.
INTRODUCTION The National Restaurant Association reports that Americans will spend $709.2 billion in restaurants in 2015, an 87% increase from 2000.1 Foods served in restaurants and fast-food establishments are typically high in total fat and calories and lacking important nutrients such as fiber and calcium.2,3 Therefore, consumption of frequent fast-food meals may lead to weight
gain and obesity.4 This issue is compounded by the fact that portion sizes at many fast-food restaurants continue to increase.5 The Restaurant Nutrition Menu Labeling Requirement within the Affordable Care Act requires chain restaurants with $ 20 locations to provide calorie information on their menus, menu boards, and drive-through menus.6 In addition, restaurants are required to post a statement about the suggested daily
1 Department of Human Nutrition and Hospitality Management, University of Alabama, Tuscaloosa, AL 2 Department of Health Science, University of Alabama, Tuscaloosa, AL 3 Institute for Social Science Research, University of Alabama, Tuscaloosa, AL Conflict of Interest Disclosure: The authors’ conflict of interest disclosures can be found online with this article on www.jneb.org. Address for correspondence: Kimberly A. Stran, PhD, RD, Department of Human Nutrition and Hospitality Management, 403 Russell Hall, Box 870311, University of Alabama, Tuscaloosa, AL 35487; Phone: (205) 737-5935; Fax: (205) 348-2982; E-mail: [email protected] Ó2015 Society for Nutrition Education and Behavior. Published by Elsevier, Inc. All rights reserved. http://dx.doi.org/10.1016/j.jneb.2015.09.009
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
caloric intake for the majority of American adults, thereby giving consumers a point of reference.6 The college environment can create changes in the dietary habits of students, especially for those who are living on their own for the first time. Many college students eat often in fast-food restaurants, which can lead to weight gain over time.7 Although the effects of calorie labeling in fastfood restaurants among adults have been widely studied,8-15 to date, only 2 studies have explored college students' use of posted calorie information in fast-food restaurants.16,17 Both studies found that women were more likely than men to use the calorie information to make a meal decision. The posting of calorie information did not have a significant effect on the calories ordered by men. In contrast, women who made meal selections from menus with calorie information ordered fewer calories than women with a standard fast-food menu.17 Reports in the existing literature lacked a theoretical framework to
1
2 Stran et al guide the choice of potential determinants of college students' use of calorie information in a fast-food restaurant. The researchers used the Theory of Planned Behavior (TPB)18 as a framework in this study to help explain the relationship between the constructs of the model (attitudes, subjective norms, and perceived behavioral control) and intention to use calorie labels. The theory posits that people are more likely to engage in the target behavior if they have an intention to participate in the behavior. Attitudes toward the behavior, subjective norms, and perceived behavioral control of the behavior each predict intention to perform the behavior. Therefore, the objectives of this study were to (1) confirm the factor structure of the TPB model through confirmatory factor analysis; (2) further investigate which sociodemographic information, current health status indicators, and TPB constructs were associated with intention of college students to use calorie labels in fast-food restaurants; and (3) determine whether college students changed their meal choice after viewing calorie information and describe the groups of college students who were more likely to change.
METHODS Participant Recruitment The authors completed recruitment for this study in conjunction with a similar study examining college students' use of calorie information in full-service restaurants. A convenience sample of 65 professors at a southeastern US university was contacted and 22 professors (33.8%) agreed to allow class time for participant recruitment. Among the 1,595 students present during classroom recruitment, 525 students (32.9%) asked for further information and 200 students (12.5%) participated in the overall study, with 100 students randomly assigned to this portion of the study. Power calculations were completed to determine adequate sample size. Based on the study of Pawlak et al19 of 53 students showing a change of 263 cal after provision of nutrition information, 90% power could be achieved with 87 students. Therefore, 100 students were randomly assigned to participate in the fast-food portion of this study.
Journal of Nutrition Education and Behavior Volume -, Number -, 2015 Students were offered a $5 cash incentive for participation. Students were excluded from the study if they majored in nutrition, were aged < 19 years, or followed a restricted diet owing to food allergy or intolerance. This study was approved by the University of Alabama's Institutional Review Board.
Data Collection Data collection took place over 5 weeks in March and April, 2013. Participants were seated in a mock dining room, asked to rate their current hunger status on a scale of 1 (ravenous) to 9 (stuffed), and then make a meal and beverage selection from a menu without calorie labels. Next, participants completed 10 math problems as a distractor, and then selected from a menu with calorie labels. After making the second meal choice, participants completed a 41-item survey. Finally, participants' height and weight were measured and recorded by trained research assistants. A stadiometer (SECA model 240, SECA Corp, Hamburg, Germany) was used to measure each participant's height to the nearest centimeter and a digital scale (Tanita model BF 350, Tanita Corporation of America, Arlington Heights, IL) was used to measure each participant's weight to the nearest 0.1 kg.
Instrument The researchers developed a survey to address sociodemographic information, general health status, potential barriers to healthful eating, and constructs in the TPB. The demographic section of the survey inquired about gender, age, rank in school, race/ ethnicity, and housing location during the school year. Previously validated questions addressed the participant's current health status, including weight perceptions, diet quality, current weight control practices, and concerns about weight.20,21 Finally, participants were asked to indicate the frequency with which they dine in fast-food restaurants. Previous research identified barriers to healthful eating among college students22-26; therefore, questions regarding 4 barriers (time, hunger, cost, and ease of use) were included in the survey as potential modifiers.
Questions used to assess attitudes, subjective norms, perceived behavioral control, and intention constructs of the TPB were developed in accordance with a manual for health researchers (Table 1).27 A set of standardized questions was used to assess attitudes (7 items), subjective norms (3 items), perceived behavioral control (1 item), and intention (4 items) to use calorie information to make a meal choice. To use these questions, a researcher inserts the behavior of interest into each question. Three surveys were excluded from the final analysis (n ¼ 97) owing to incomplete answers to questions that addressed the TPB constructs. Two fast-food menus were developed for this study to mimic menus of the top 3 fast-food chain restaurants in sales in the US. A menu from a nationally recognized restaurant was not used to broaden the scope of the study, so that results are applicable to more > 1 specific fastfood restaurant. The menus were designed to be generic to reduce selection of preferred items at a specific restaurant. The menus contained 32 typical items found in a fast-food restaurant, including hamburgers, chicken and fish sandwiches, salads, side items such as french fries and fruit cups, and beverages. The second fast-food menu was identical to the first menu, but contained calorie information next to the name of each food or beverage item. The second menu also had a different restaurant name, order of items, and color scheme to make it appear to be a different restaurant, because the time between each menu presentation was short. Calorie information was obtained from the Web sites of the top 3 fast-food restaurants and mean calorie level for each menu item was computed so that the information was realistic and similar to what would be posted on a restaurant menu board. Cost was not included on the menus so as to understand the direct result of caloric information on meal selections.
Data Analyses According to Schumacker and Lomax,28 confirmatory factor analysis (CFA) allows the researcher to specify
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
Stran et al 3
Table 1. Theory of Planned Behavior Construct Questions and Ranges (1–7) of Participant Responses (n ¼ 97)
approximation takes into account the error of approximation and indicates an acceptable model fit with values < 0.08 and a good model fit with values < 0.05.28,29,32 Based on these guidelines, the measurements did not indicate a good model fit for the model (c2 ¼ 368.091; degrees of freedom, ¼ 183; P < .001; CFI ¼ 0.806; NFI ¼ 0.684; Goodness of Fit Index ¼ 0.749, and root mean square error of approximation ¼ 0.101). This model also had 8 items with a factor loading < 0.40 (using calorie information to make a meal decision is worthless/useful, 0.195; using calorie information to make a meal decision is convenient/inconvenient, 0.385; use of calorie information is based on amount of time I have, 0.332; use of calorie information is based on how hungry I am, 0.239; use of calorie information is based on cost, 0.390; I expect to use calorie information to make a meal decision, 0.359; my parents think I should not/should use calorie information, 0.390; and my friends think I should not/should use calorie information, 0.026). The results were somewhat unexpected because the survey items were developed using standardized TPB questions. As suggested by Suhr,33 if CFA does not result in an appropriate model fit, EFA should be used. The researchers used EFA to explore the factor structure of the TPB questions because the CFA model did not fit. Exploratory factor analysis with varimax rotation was completed using Statistical Analysis Software (version 9.3, SAS Institute Inc, Cary, NC, 2012) to test each question for 3 of the 4 constructs from TPB: attitudes, subjective norms, and
Construct and Variables Attitudes (7 items) Using posted calorie information to make a meal decision is: Harmful (1)/beneficial (7) Worthless (1)/useful (7) Difficult (1)/easy (7) Inconvenient (1)/convenient (7) Having calorie information posted at fast-food restaurants is: Harmful (1)/beneficial (7) Bad (1)/good (7) Worthless (1)/useful (7) Subjective norms (3 items) My parents think I should not (1)/should (7) use posted calorie information My friends think I should not (1)/should (7) use posted calorie information My close friends think I should not (1)/should (7) use posted calorie information Perceived behavioral control (1 item) I am confident that I could use calorie information in a fast-food restaurant to make a healthy choice: strongly disagree (1)/strongly agree (7) Intention (4 items) I expect to use posted calorie information to make a meal decision: strongly disagree (1)/strongly agree (7) I want to use posted calorie information to make a meal decision: strongly disagree (1)/strongly agree (7) I intend to use posted calorie information to make a meal decision: strongly disagree (1)/strongly agree (7) If calorie information were available in a fast-food restaurant, with what frequency would you use it: never (1)/often (7)
a certain number of factors based on a theoretical model, unlike exploratory factor analysis (EFA), in which the numbers of factors is unknown. Hair et al29 went on to say that objectives of CFA are ‘‘1) to verify the proposed factor structure and 2) to explore if any significant modifications are needed.’’ Because one of the objectives in the study was to verify the factor structure with the TPB model, CFA was employed. Methodology suggested in previous research30 was used to evaluate the validity of the TPB by assessing the fit of the factors of TPB (attitudes, subjective norms, perceived behavior, and intention). Data were analyzed using AMOS version 23 (IBM, Chicago, IL, 2015). In CFA, the measured variables were modeled as a function of each latent variable, as identified in the Figure. As suggested by Hair et al,29 the factor loadings were first considered. Standard errors were then checked along with modification indices, unstandardized and standardized factor loadings, and factor covariance. Along
with chi-square, several types of overall model fit were measured, including the Comparative Fit Index (CFI), Normed Fit Index (NFI), Goodness of Fit Index, and root mean square error of approximation. According to Bentler,31 a value > 0.90 is representative of a well-fit model for both NFI and CFI. Goodness of Fit Index values range from 0 to 1, with 1 being a perfect fit.29 Root mean square error of
Figure. Theory of Planned Behavior. Adapted from Ajzen I. The theory of planned behavior. Organ Behav Hum Dec. 1991;50:182.
4 Stran et al
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
Table 2. Modified Theory of Planned Behavior Constructs and Variables With Standardized Loadings and Reliability (n ¼ 97)
Construct and Variables Attitudes (5 items) Using posted calorie information to make a meal decision is: Harmful/beneficial Worthless/useful Having calorie information posted at fast-food restaurants is: Harmful/beneficial Bad/good Worthless/useful
Mean (SD) 6.15 (0.99)
Standardized Loadings a .89
6.09 (1.04) 5.83 (1.40)
.84 .83
6.37 (1.03) 6.27 (1.32) 6.21 (1.10)
.88 .76 .88
Subjective norms (3 items) My parents think I should not/should use posted calorie information My friends think I should not/should use posted calorie information My close friends think I should not/should use posted calorie information
3.98 (1.21) 4.67 (1.32)
.63
3.93 (1.20)
.88
4.02 (1.39)
.91
Perceived behavioral control (1 item)
5.86 (1.46)
Intention (4 items) I expect to use posted calorie information to make a meal decision (strongly disagree/ strongly agree) I want to use posted calorie information to make a meal decision (strongly disagree/ strongly agree) I intend to use posted calorie information to make a meal decision (strongly disagree/ strongly agree) If calorie information were available in a fastfood restaurant, with what frequency would you use it? (never/often)
4.68 (1.54) 4.14 (1.76)
.83
4.87 (1.88)
.93
4.57 (1.84)
.95
5.13 (1.59)
.77
intentions (Table 2). Perceived behavioral control was assessed using only 1 question and so was not included in the factor analysis. An eigenvalue of > 1 and scree plots were used to determine the number of factors and questions with factor loadings > 0.40 were included in the respective factors.34 Internal consistencies of responses for each construct were assessed using Cronbach alpha. Five of the 7 questions assessing attitudes loaded onto 1 factor (Cronbach a ¼ .89) (Table 2). Two questions did not load onto any factor and were removed from the analysis (using calorie information posted on fast-food menus to make a meal decision is easy/difficult and convenient/inconvenient). The attitudes scale was constructed by averaging the responses of the 5 attitudes questions. Positive attitudes were represented by higher numbers. The 3
.86
.89
subjective norms questions also loaded onto 1 factor with Cronbach a ¼ .74. Subjective norms scores were averaged among the 3 questions, in which a higher number represented greater feelings of social pressure to use calorie labels. The 4-question intention scale had Cronbach a ¼ .89. Higher numbers represented greater intention to use calorie information if posted on fast-food menus. Cronbach a ¼ .7 is considered acceptable35; therefore, the modified attitudes, subjective norms, and intention scales were deemed appropriate for use in this study.
Potential Modifiers To assess the current health status of the students, 2 health status indicators were computed. Body mass index was calculated for each participant
using measured height and weight. Results were categorized based on National Institutes of Health guidelines into underweight and normal weight (<25 kg/m2) and overweight and obese ($25 kg/m2).36 Responses to the question about overall diet perceptions were dichotomized into an excellent/very good/good category and a fair/poor category, which is consistent with other research and provides a clear positive or negative view of current dietary practices.37 Race, housing status, current weight perception, dieting habits, use of the Nutrition Facts Panel, and frequency of fast-food visits were used as potential modifiers of intention and a difference in calories ordered from the menus. Barriers questions were used individually in the analyses as potential modifiers. The authors used univariate analyses (t tests and ANOVA) to determine significant differences (P < .05) in intention scores and changes in calories ordered for each potential covariate. Variables with significant differences were used in the final multiple regression analyses, in addition to gender and hunger status variables. These 2 variables were included to adjust for observed differences in calories ordered. Although the use of calorie information on the Nutrition Facts Panel was significant in univariate analyses, this variable was not included in multiple regression because this is closely related to the study design and could account for a large amount of the variance if included in the model. Multiple regression was used to test whether the 3 TPB factors would predict intention to use calorie labels while adjusting for confounders, or the variables that were significant in univariate analysis. The difference between total calories ordered from each of the 2 fast-food menus was calculated for each participant using the calorie information posted on the second menu. A paired t test was conducted to determine whether a significant difference existed in the number of calories ordered without posted calorie information vs with it. Multiple regression was used to test whether the 4 TPB factors would predict a change in the caloric content of meals ordered while adjusting for confounders that were significant from univariate analyses. P < .05 was used for all tests.
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
Stran et al 5
Table 3. Sociodemographic Characteristics, Health Status Indicators, Dietary Habits, Mean Intention Scores, and Mean Differences in Calories Ordered by College Students
Characteristics Gender (n ¼ 100) Male Female
n
%
Mean Intention Score (SD)a
37 63
37 63
4.37 (1.54) 4.86 (1.53)
Race (n ¼ 100) Non-Hispanic white Non-Hispanic black Other
74 18 8
74 18 8
4.69 (1.53) 4.38 (1.80) 5.22 (0.91)
Housing (n ¼ 100) On campus Off campus
49 51
49 51
4.64 (1.62) 4.88 (1.45)
Perceived diet quality (n ¼ 100) Excellent, very good, or good Fair or poor
59 41
59 41
4.93 (1.60) 4.32 (1.40)
Weight perception (n ¼ 97) Underweight About the right weight Overweight
6 72 19
6 74 20
3.08 (1.78) 4.55 (1.54) 5.41 (0.87)
Body mass index (n ¼ 100) < 25 kg/m2 $ 25 kg/m2
62 38
62 38
4.58 (1.70) 4.84 (1.25)
Dieting (n ¼ 100) Yes No
16 84
16 84
5.22 (1.64) 4.57 (1.51)
Worried about weight (n ¼ 100) Yes No
48 52
48 52
5.13 (1.35) 4.26 (1.61)
Typical fast-food visits (n ¼ 100) # 3 times/wk $ 4 times/wk
84 16
84 16
4.93 (1.49) 3.38 (1.14)
65
65
5.25 (1.16)
35
35
3.61 (1.61)
Nutrition Facts Panel use (n ¼ 100) Always, most of the time, sometimes Rarely, never
P .13
Mean Difference in Calories Ordered (SD)a
P .04
3.4 (296.5) 114.0 (291.9) .44
.18 80.9 (301.3) 28.1 (259.7) 196.3 (313.1)
.18
.36 42.4 (291.2) 97.5 (304.1)
.04
.59 84.0 (268.3) 51.2 (337.9)
.003
.33 55.8 (363.7) 51.2 (320.0) 126.8 (195.3)
.43
.31 94.4 (308.6) 31.6 (278.3)
.13
.80 53.1 (337.9) 73.9 (291.4)
.005
.26 105.5 (261.4) 38.3 (326.7)
< .001
< .001
.54 62.5 (299.5) 112.8 (293.4) .006 129.5 (291.3) 38.9 (281.3)
< .001
.44
Use of calorie information on Nutrition Facts Panel (n ¼ 100) Always, most of the time, sometimes Rarely, never
79
79
5.08 (1.35)
82.4 (292.3)
21
21
3.14 (1.23)
26.0 (320.1)
Calories needed in 1 d (n ¼ 94) # 1,500 1,501–2,000 $ 2,001
28 31 35
30 33 37
4.81 (1.40) 4.85 (1.62) 4.58 (1.38)
.72
.32 89.3 (270.2) 121.6 (312.9) 11.4 (319.8)
t tests and ANOVAs were used to determine differences in mean intention scores and mean differences in calories ordered. Note: Bonferroni correction was used for multiple comparisons (P < .016). a
RESULTS Demographics The majority of participants were women (63%), non-Hispanic white (74%), and freshmen or sophomores
(67%) (Table 3). Mean age was 20.0 years (SD, 1.27 years; range, 19–26 years). Whereas nearly half of participants reported that they were worried about their weight (48%), only 16% were dieting and 74% perceived
themselves to be about the right weight. A total of 79% of participants reported using the calorie information on a label at least sometimes whereas 65% reported using the Nutrition Facts Panel at least
6 Stran et al
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
Table 4. Predictors of Intention to Use Posted Calorie Information in Fast-Food Restaurants (n ¼ 97)* Variable
b
Gender
0.13
0.26
0.50
.62
Attitudes
0.37
0.13
2.84
.006
SE
t
P
Subjective norms
0.42
0.11
3.77
< .001
Perceived behavioral control
0.45
0.18
2.51
.01
Perceived diet quality
0.56
0.25
2.23
.03
Perceived weight status
0.01
0.16
0.05
.96
Worried about weight
0.51
0.27
1.93
.06
Frequency of dining at fast-food restaurants
0.67
0.35
1.92
.06
Time barrier
0.01
0.08
0.16
.87
Ease barrier
0.10
0.14
0.70
.48
Cost barrier
0.02
0.08
0.25
.80
0.03
0.07
0.43
.67
0.06
0.08
0.71
.48
Hunger barrier Hunger status during participation
*Multivariate coefficient ¼ 0.52; F ¼ 6.92; P < .001.
sometimes when deciding to buy a food product. Mean number of times respondents visited fast-food restaurants was 2.61 times/wk.
Intention Table 3 depicts mean intention scores by sociodemographic characteristics, health status indicators, and dietary habits. In univariate analysis, significant differences in mean intention scores were found for perceived diet quality (t ¼ 1.98; P < .05), perceived weight status (t ¼ 6.24; P ¼ .003),
worry about weight (t ¼ 2.89; P ¼ .005), and frequency of fast-food restaurant visits (t ¼ 3.95; P < .001). In multivariate analysis, attitudes, subjective norms, and perceived behavioral control predicted higher intention to use calorie information in a fast-food restaurant. Students reporting fair/poor diet quality (t ¼ 2.23; P ¼ .03) had significantly lower intention scores than their counterparts (Table 4). These variables explained 52% of the variance in intention scores (multivariate coefficient, 0.52; F ¼ 6.92; P < .001).
Table 5. Predictors of a Difference in Calorie Content of Meals Ordered From a Fast-Food Restaurant Menu (n ¼ 97)* Variable Gender
b 86.3
SE 63.1
t 1.37
P .18
Attitudes
23.6
33.7
0.70
.49
5.4
27.1
0.20
.84
Perceived behavioral control
31.6
45.5
0.69
.49
Intention
Subjective norms
31.7
24.3
1.30
.18
Time barrier
27.0
18.3
1.48
.14
Ease barrier
43.9
35.1
1.25
.21
Cost barrier
37.4
18.3
2.04
.04
Hunger barrier
36.7
16.4
2.23
.03
Hunger status during participation
5.1
19.3
0.27
.79
*Multivariate coefficient ¼ 0.18; F ¼ 1.92; P ¼ .05.
Difference in Calories Ordered Participants in this study ordered significantly fewer calories (mean ¼ 70.6; SD ¼ 298) when selecting from the menu with labels vs the one without labels (t ¼ 2.37; P ¼ .02). Mean caloric content for an order without labels was 908.9 cal (SD ¼ 290.1) vs 838.3 cal (SD, 341.5) when menu labels were present. In univariate testing, significant changes in the difference in calories ordered were found for gender (t ¼ 1.93; P ¼ .05) and hunger status (F ¼ 2.13; P ¼ .04) (Table 3). In multivariate analysis, significant predictors of a difference in the calorie content of meals ordered were the hunger barrier and cost barrier (Table 5). These variables explained 18% of the variance in differences in calorie content of meals ordered (multivariate coefficient, 0.18; F ¼ 1.92; P ¼ .05). As scores for agreement with the statement ‘‘The decision to use calorie information in a fast-food restaurant depends on how hungry I am’’ increased by 1 unit, participants ordered 36.7 more calories from the labeled menu than when the menu was not labeled. As scores for agreement with the statement ‘‘The decision to use calorie information in a fast-food restaurant is based on the cost of the menu items’’ increased by 1 unit, participants ordered 37.4 fewer calories from the labeled menu vs the menu that was not labeled. Gender (t ¼ -1.37; P ¼ .18) and hunger status at the time of participation (t ¼ 0.27; P ¼ .79) were not significant predictors in the difference in calories ordered among participants.
DISCUSSION The TPB model suggests that attitudes, social norms, and perceived behavioral control will predict intention to change a behavior, which is the strongest predictor of behavior. The TPB model predicted intention to use calorie information in a fast-food restaurant but was not associated with a difference in the caloric content of the meal ordered when menus were labeled. As predicted, attitudes, subjective norms, and perceived behavioral control were significantly related
Journal of Nutrition Education and Behavior Volume -, Number -, 2015 to intention to use a menu label in this small sample of college students. However, these variables were not related to a change in calories ordered when the barriers of hunger and cost were added to the model. Barriers to engaging in the behavior (time, ease, cost, and hunger) did not affect intention to engage in the behavior, but hunger and cost were significantly related to differences in the caloric content of the orders when menus were labeled. Although a previous study16 also found that positive attitudes toward this behavior were associated with greater intention to use posted information, this study assessed all TPB constructs using a quasi-experimental design. In this study, hunger is a commonly perceived barrier to using menu labels in a fast-food restaurant. Participants who strongly agreed that the decision to use calorie information in a fastfood restaurant is based on hunger ordered more calories than did those who did not agree with this statement. Current hunger status was not predictive of a difference in calories ordered between nonlabeled and labeled menus. In a study of 324 college students, Horacek and Betts38 found that students who perceived their dietary choices to be strongly influenced by hunger and other cues consumed significantly more calories than those whose dietary choices were selected based on the nutritional aspects of the foods. The current findings also indicated that when participants perceived cost to be a barrier, they ordered fewer calories at a fast-food restaurant. Students concerned about their finances may have developed behaviors that simultaneously cut costs and calories. The authors did not measure food insecurity and socioeconomic status in this study. Future studies should assess these measures in light of recent research that suggests higher than expected rates of food insecurity among college students.39 Previous studies have also indicated that cost is a potential barrier to healthful eating practices by university students.23,40 At the time of this study, the menu labeling law was not fully implemented. Thus, students may not have the opportunity to practice and develop skills in ordering menu
items while considering the caloric content of the items against other known attributes of the food such as cost or personal preferences. Neither college students' weight status nor weight concerns were predictive of changes in calories ordered when menus were labeled in a fast-food restaurant. However, those who perceive their overall diet to be poor or fair had significantly lower intention to use the menu labels compared with their counterparts who perceive their diets to be excellent, very good, or good. In addition, frequency of fast-food consumption was not a significant predictor of a difference in the calories ordered between menus. Because this is a cross-sectional study, it cannot be inferred whether students who limit fast-food visits would value nutrition. The scope and interpretation of this study have limitations. Participants in this study examined a generic fast-food restaurant menu and ordered items they would most likely consume. An actual meal was not served or consumed; thus, an accurate calorie count of the meal was not assessed. Only 1 question was used to assess students' perceived behavioral control; therefore, it was not able to be used in the factor analysis. Although cost can be a barrier to healthful eating for college students,23,40 prices were not included on the menus used in this study. The researchers decided to leave price off the menu in an effort to understand the direct result of caloric information in the selections. In addition, although the menus used had a limited number of options, as in a fast-food restaurant, and the researchers asked participants to look over the menu quickly, the researchers could not mimic the fastpaced ordering procedures that are associated with visiting a fast-food counter or drive-through. Dietary restraint has a role in food selection. This study measured only self-report of dieting, and a more thorough means of assessing dietary restraint is needed in future studies. Because this study was conducted at 1 southeastern university, the results may not be applicable to students with a more diverse background or to nonstudents.
Stran et al 7
IMPLICATIONS FOR RESEARCH AND PRACTICE Results from this study suggest that students weigh cost and hunger into their decisions, and these barriers to healthy eating habits are also barriers to the use of menu labels in fast-food restaurants. The TPB predicted intention to use menu labels but did not predict changes in calories ordered between menus. After viewing the menu with calorie labels, participants in this small study were able to select a lower-calorie meal. Educational programs are needed aimed at decreasing perceived barriers, such as cost and mindful eating practices. Future research should investigate the influence of students' barriers in a fastfood restaurant setting. In addition, because this study did not measure participants' intake of the meals ordered, future studies should investigate whether the TPB is predictive of both intention and a change in calories ordered and consumed in a fast-food restaurant setting. Restaurant menu labeling in fast-food restaurants could provide the information college students need to select lowercalorie items.
ACKNOWLEDGMENTS The authors would like to thank the University of Alabama Coordinated Program students for their assistance with subject recruitment and data collection.
REFERENCES 1. National Restaurant Association. 2015 Restaurant Industry Pocket Factbook. 2015. https://www.restaurant.org/Downloads/ PDFs/News-Research/research/Factbook 2015_LetterSize-FINAL.pdf. Accessed July 23, 2015. 2. Lin BH, Guthrie J, Fraz~ao E. Nutrient contribution of food away from home. In: Fraz~ao E, ed. American’s Eating Habits: Changes and Consequences. Washington, DC: US Dept of Agriculture; 1999:213-242. 3. Scourboutakous MJ, L’Abbe MR. Restaurant menus: calories, calorie density, and serving size. Am J Prev Med. 2012;43:249-255.
8 Stran et al 4. Burton S, Creyer EH, Kees J, Huggins K. Attacking the obesity epidemic: the potential health benefits of providing nutrition information in restaurants. Am J Public Health. 2006;96:1669-1675. 5. Young LR, Nestle M. Reducing portion sizes to prevent obesity: a call to action. Am J Prev Med. 2012;43:565-568. 6. United States Food and Drug Administration. Patient Protection and Affordable Care Act, Public Law 111–148, Sec. 4205–Nutrition labeling of standard menu items at chain restaurants. March 23, 2010. http://www.gpo.gov/ fdsys/pkg/FR-2010-08-25/pdf/2010-21065. pdf. Accessed July 23, 2015. 7. Racette SB, Deusinger SS, Strube MJ, Highstein GR, Deusinger RH. Weight changes, exercise, and dietary patterns during freshman and sophomore years of college. J Am Coll Health. 2005;53: 245-251. 8. Bassett MT, Dumanovsky T, Huang C, et al. Purchasing behavior and calorie information at fast-food chains in New York City, 2007. Am J Public Health. 2008;98:1457-1459. 9. Dumanovsky T, Huang CY, Nonas CA, Matte TD, Bassett MT, Silver LD. Changes in energy content of lunchtime purchases from fast food restaurants after introduction of calorie labelling: cross sectional customer surveys. BMJ. 2011;343:d4464. 10. Farley TA, Caffarelli A, Bassett MT, Silver L, Frieden TR. New York City’s fight over calorie labeling: a two year struggle ultimately proves that innovation in food regulation is entirely possible at the local level. Health Affairs. 2009;28:w1098-w1109. 11. Finkelstein EA, Strombotne KL, Chan NL, Krieger J. Mandatory menu labeling in one fast-food chain in King County, Washington. Am J Prev Med. 2011;40:122-127. 12. Lando AM, Labiner-Wolfe J. Helping consumers make more healthful food choices: consumer views on modifying food labels and providing point-ofpurchase nutrition information at quick-service restaurants. J Nutr Educ Behav. 2007;39:157-163. 13. Piron J, Smith LV, Simon P, Cummings PL, Kuo T. Knowledge, attitudes and potential response to menu labelling in an urban public health clinic population. Public Health Nutr. 2009;13: 550-555. 14. Vadiveloo MK, Dixon LB, Elbel B. Consumer purchasing patterns in
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
response to calorie labeling legislation in New York City. Int J Behav Nutr Phys Act. 2011;8:51. http://dx.doi.org/10.1186 /1479-5868-8-51. Wisdom J, Downs J, Loewenstein G. Promoting healthy choices: information vs. convenience. Am Econ J Appl Econ. 2010;2:164-178. Bates K, Burton S, Howlett E, Huggins K. The roles of gender and motivation as moderators of the effects of calorie and nutrition information provision on away-from-home foods. J Consum Aff. 2009;43:249-273. Gerend MA. Does calorie information promote lower calorie fast food choices among college students? J Adolesc Health. 2009;44:84-86. Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process. 1991;50:182. Pawlak R, Cerutti CS, Quinton R. Taking an undergraduate nutrition course results in favorable attitudes toward a healthful diet and improved intake of several key nutrients. Fam Consum Sci Res J. 2009;38:3-10. Centers for Disease Control and Prevention. National Center for Health Statistics (NCHS). National Health and Nutrition Examination Survey Questionnaire. Hyattsville, MD: US Dept of Health and Human Services, Centers for Disease Control and Prevention, 2007-2008. http://wwwn.cdc.gov/nchs/ nhanes/search/nhanes07_08.aspx. Accessed July 24, 2015. Centers for Disease Control and Prevention. National Center for Health Statistics (NCHS). National Health and Nutrition Examination Survey Data. Hyattsville, MD: US Dept of Health and Human Services, Centers for Disease Control and Prevention, 2009-2010. http://wwwn.cdc. gov/nchs/nhanes/search/nhanes09_10. aspx. Accessed July 24, 2015. King KA, Mohl K, Bernard AL, Vidourek RA. Does involvement in healthy eating among university students differ based on exercise status and reasons for exercise? Calif J Health Promot. 2007;5(3):106-119. LaCaille LJ, Dauner KN, Krambeer RJ, Pedersen J. Psychosocial and environmental determinants of eating behaviors, physical activity, and weight change among college students: a qualitative analysis. J Am Coll Health. 2011; 59:531-538. Silliman K, Rodas-Fortier K, Neyman M. A survey of dietary and
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
exercise habits and perceived barriers to following a healthy lifestyle in a college population. Californian J Health Promot. 2004;2:10-19. Rydell SA, Harnack LJ, Oakes JM, Story M, Jeffrey RW, French SA. Why eat at fast food restaurants: reported reasons among frequent customers. J Am Diet Assoc. 2008;108: 2066-2070. Driskell JA, Meckna BR, Scales NE. Differences exist in the eating habits of university men and women at fast food restaurants. Nutr Res. 2006;26: 524-530. Francis JJ, Eccles MP, Johnston M, et al. Constructing Questionnaires Based on the Theory of Planned Behavior: A Manual for Health Services Researchers. Newcastle-upon-Tyne, UK: University of Newcastle; 2004. Schumacker RE, Lomax RG. A Beginner’s Guide to Structural Equation Modeling. 2nd ed. Mahwah, NJ: Lawrence Erlbaum Associates; 2004. Hair JF, Anderson RE, Tatham RL, Black WC. Multivariate Data Analysis. 5th ed. Upper Saddle River, NJ: Prentice Hall; 1998. Birch LL, Fisher JO, GrimmThomas K, Markey CN, Sawyer R, Johnson SL. Confirmatory factor analysis of the Child Feeding Questionnaire: a measure of parental attitudes, beliefs, and practices about child feeding and obesity proneness. Appetite. 2001;36:201-210. Bentler PM. Comparative fit indexes in structural equation models. Psychol Bull. 1990;88:588-606. Browne MW, Cudeck R. Alternative ways of assessing model fit. In: Bollen K, Long J, eds. Testing Structural Equation Models. Newbury Park, CA: Sage; 1993:136-192. Suhr DD. Exploratory or confirmatory factor analysis? Proceedings of the Thirty-first Annual SAS Users Group International Conference. Cary, NC: SAS Institute; 2006. O’Rourke N, Hatcher L, Stepanski EJ. A Step-by-Step Approach to Using SAS for Univariate and Multivariate Statistics. 2nd ed. Cary, NC: SAS Institute, Inc; 2005. Nunnally J, Bernstein L. Psychometric Theory. New York, NY: McGrawHill Higher, Inc; 1994. National Institutes of Health. The Practical Guide: Identification, Evaluation, and Treatment of Overweight and Obesity in Adults (NIH Pub No. 00–4084).
Journal of Nutrition Education and Behavior Volume -, Number -, 2015 October 2000. http://www.nhlbi.nih. gov/guidelines/obesity/prctgd_c.pdf. Accessed July 24, 2015. 37. Goodwin DK, Knol LK, Eddy JM, Fitzhugh EC, Kendrick OW, Donahue RE. The relationship between self-rated health status and the overall quality of dietary intake of US
adolescents. J Am Diet Assoc. 2006;106: 1450-1453. 38. Horacek TM, Betts NM. Students cluster into 4 groups according to the factors influencing their dietary intake. J Am Diet Assoc. 1998;98:1464-1467. 39. Gaines A, Robb CA, Knol LL, Sickler S. Examining the role of financial factors,
Stran et al 9 resources and skills in predicting food security status among college students. Int J Consumer Sci. 2014;38:374-384. 40. Garcia AC, Sykes L, Matthews J, Martin N, Leipert B. Perceived facilitators of and barriers to healthful eating among university students. Can J Diet Pract Res. 2010;71:e28-e33.
9.e1 Stran et al
CONFLICT OF INTEREST The authors have not stated any conflicts of interest.
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
INTRODUCTION The National Restaurant Association reports that Americans will spend $709.2 billion in restaurants in 2015, an 87% increase from 2000.1 Foods served in restaurants and fast-food establishments are typically high in total fat and calories and lacking important nutrients such as fiber and calcium.2,3 Therefore, consumption of frequent fast-food meals may lead to weight
gain and obesity.4 This issue is compounded by the fact that portion sizes at many fast-food restaurants continue to increase.5 The Restaurant Nutrition Menu Labeling Requirement within the Affordable Care Act requires chain restaurants with $ 20 locations to provide calorie information on their menus, menu boards, and drive-through menus.6 In addition, restaurants are required to post a statement about the suggested daily
1 Department of Human Nutrition and Hospitality Management, University of Alabama, Tuscaloosa, AL 2 Department of Health Science, University of Alabama, Tuscaloosa, AL 3 Institute for Social Science Research, University of Alabama, Tuscaloosa, AL Conflict of Interest Disclosure: The authors’ conflict of interest disclosures can be found online with this article on www.jneb.org. Address for correspondence: Kimberly A. Stran, PhD, RD, Department of Human Nutrition and Hospitality Management, 403 Russell Hall, Box 870311, University of Alabama, Tuscaloosa, AL 35487; Phone: (205) 737-5935; Fax: (205) 348-2982; E-mail: [email protected] Ó2015 Society for Nutrition Education and Behavior. Published by Elsevier, Inc. All rights reserved. http://dx.doi.org/10.1016/j.jneb.2015.09.009
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
caloric intake for the majority of American adults, thereby giving consumers a point of reference.6 The college environment can create changes in the dietary habits of students, especially for those who are living on their own for the first time. Many college students eat often in fast-food restaurants, which can lead to weight gain over time.7 Although the effects of calorie labeling in fastfood restaurants among adults have been widely studied,8-15 to date, only 2 studies have explored college students' use of posted calorie information in fast-food restaurants.16,17 Both studies found that women were more likely than men to use the calorie information to make a meal decision. The posting of calorie information did not have a significant effect on the calories ordered by men. In contrast, women who made meal selections from menus with calorie information ordered fewer calories than women with a standard fast-food menu.17 Reports in the existing literature lacked a theoretical framework to
1
2 Stran et al guide the choice of potential determinants of college students' use of calorie information in a fast-food restaurant. The researchers used the Theory of Planned Behavior (TPB)18 as a framework in this study to help explain the relationship between the constructs of the model (attitudes, subjective norms, and perceived behavioral control) and intention to use calorie labels. The theory posits that people are more likely to engage in the target behavior if they have an intention to participate in the behavior. Attitudes toward the behavior, subjective norms, and perceived behavioral control of the behavior each predict intention to perform the behavior. Therefore, the objectives of this study were to (1) confirm the factor structure of the TPB model through confirmatory factor analysis; (2) further investigate which sociodemographic information, current health status indicators, and TPB constructs were associated with intention of college students to use calorie labels in fast-food restaurants; and (3) determine whether college students changed their meal choice after viewing calorie information and describe the groups of college students who were more likely to change.
METHODS Participant Recruitment The authors completed recruitment for this study in conjunction with a similar study examining college students' use of calorie information in full-service restaurants. A convenience sample of 65 professors at a southeastern US university was contacted and 22 professors (33.8%) agreed to allow class time for participant recruitment. Among the 1,595 students present during classroom recruitment, 525 students (32.9%) asked for further information and 200 students (12.5%) participated in the overall study, with 100 students randomly assigned to this portion of the study. Power calculations were completed to determine adequate sample size. Based on the study of Pawlak et al19 of 53 students showing a change of 263 cal after provision of nutrition information, 90% power could be achieved with 87 students. Therefore, 100 students were randomly assigned to participate in the fast-food portion of this study.
Journal of Nutrition Education and Behavior Volume -, Number -, 2015 Students were offered a $5 cash incentive for participation. Students were excluded from the study if they majored in nutrition, were aged < 19 years, or followed a restricted diet owing to food allergy or intolerance. This study was approved by the University of Alabama's Institutional Review Board.
Data Collection Data collection took place over 5 weeks in March and April, 2013. Participants were seated in a mock dining room, asked to rate their current hunger status on a scale of 1 (ravenous) to 9 (stuffed), and then make a meal and beverage selection from a menu without calorie labels. Next, participants completed 10 math problems as a distractor, and then selected from a menu with calorie labels. After making the second meal choice, participants completed a 41-item survey. Finally, participants' height and weight were measured and recorded by trained research assistants. A stadiometer (SECA model 240, SECA Corp, Hamburg, Germany) was used to measure each participant's height to the nearest centimeter and a digital scale (Tanita model BF 350, Tanita Corporation of America, Arlington Heights, IL) was used to measure each participant's weight to the nearest 0.1 kg.
Instrument The researchers developed a survey to address sociodemographic information, general health status, potential barriers to healthful eating, and constructs in the TPB. The demographic section of the survey inquired about gender, age, rank in school, race/ ethnicity, and housing location during the school year. Previously validated questions addressed the participant's current health status, including weight perceptions, diet quality, current weight control practices, and concerns about weight.20,21 Finally, participants were asked to indicate the frequency with which they dine in fast-food restaurants. Previous research identified barriers to healthful eating among college students22-26; therefore, questions regarding 4 barriers (time, hunger, cost, and ease of use) were included in the survey as potential modifiers.
Questions used to assess attitudes, subjective norms, perceived behavioral control, and intention constructs of the TPB were developed in accordance with a manual for health researchers (Table 1).27 A set of standardized questions was used to assess attitudes (7 items), subjective norms (3 items), perceived behavioral control (1 item), and intention (4 items) to use calorie information to make a meal choice. To use these questions, a researcher inserts the behavior of interest into each question. Three surveys were excluded from the final analysis (n ¼ 97) owing to incomplete answers to questions that addressed the TPB constructs. Two fast-food menus were developed for this study to mimic menus of the top 3 fast-food chain restaurants in sales in the US. A menu from a nationally recognized restaurant was not used to broaden the scope of the study, so that results are applicable to more > 1 specific fastfood restaurant. The menus were designed to be generic to reduce selection of preferred items at a specific restaurant. The menus contained 32 typical items found in a fast-food restaurant, including hamburgers, chicken and fish sandwiches, salads, side items such as french fries and fruit cups, and beverages. The second fast-food menu was identical to the first menu, but contained calorie information next to the name of each food or beverage item. The second menu also had a different restaurant name, order of items, and color scheme to make it appear to be a different restaurant, because the time between each menu presentation was short. Calorie information was obtained from the Web sites of the top 3 fast-food restaurants and mean calorie level for each menu item was computed so that the information was realistic and similar to what would be posted on a restaurant menu board. Cost was not included on the menus so as to understand the direct result of caloric information on meal selections.
Data Analyses According to Schumacker and Lomax,28 confirmatory factor analysis (CFA) allows the researcher to specify
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
Stran et al 3
Table 1. Theory of Planned Behavior Construct Questions and Ranges (1–7) of Participant Responses (n ¼ 97)
approximation takes into account the error of approximation and indicates an acceptable model fit with values < 0.08 and a good model fit with values < 0.05.28,29,32 Based on these guidelines, the measurements did not indicate a good model fit for the model (c2 ¼ 368.091; degrees of freedom, ¼ 183; P < .001; CFI ¼ 0.806; NFI ¼ 0.684; Goodness of Fit Index ¼ 0.749, and root mean square error of approximation ¼ 0.101). This model also had 8 items with a factor loading < 0.40 (using calorie information to make a meal decision is worthless/useful, 0.195; using calorie information to make a meal decision is convenient/inconvenient, 0.385; use of calorie information is based on amount of time I have, 0.332; use of calorie information is based on how hungry I am, 0.239; use of calorie information is based on cost, 0.390; I expect to use calorie information to make a meal decision, 0.359; my parents think I should not/should use calorie information, 0.390; and my friends think I should not/should use calorie information, 0.026). The results were somewhat unexpected because the survey items were developed using standardized TPB questions. As suggested by Suhr,33 if CFA does not result in an appropriate model fit, EFA should be used. The researchers used EFA to explore the factor structure of the TPB questions because the CFA model did not fit. Exploratory factor analysis with varimax rotation was completed using Statistical Analysis Software (version 9.3, SAS Institute Inc, Cary, NC, 2012) to test each question for 3 of the 4 constructs from TPB: attitudes, subjective norms, and
Construct and Variables Attitudes (7 items) Using posted calorie information to make a meal decision is: Harmful (1)/beneficial (7) Worthless (1)/useful (7) Difficult (1)/easy (7) Inconvenient (1)/convenient (7) Having calorie information posted at fast-food restaurants is: Harmful (1)/beneficial (7) Bad (1)/good (7) Worthless (1)/useful (7) Subjective norms (3 items) My parents think I should not (1)/should (7) use posted calorie information My friends think I should not (1)/should (7) use posted calorie information My close friends think I should not (1)/should (7) use posted calorie information Perceived behavioral control (1 item) I am confident that I could use calorie information in a fast-food restaurant to make a healthy choice: strongly disagree (1)/strongly agree (7) Intention (4 items) I expect to use posted calorie information to make a meal decision: strongly disagree (1)/strongly agree (7) I want to use posted calorie information to make a meal decision: strongly disagree (1)/strongly agree (7) I intend to use posted calorie information to make a meal decision: strongly disagree (1)/strongly agree (7) If calorie information were available in a fast-food restaurant, with what frequency would you use it: never (1)/often (7)
a certain number of factors based on a theoretical model, unlike exploratory factor analysis (EFA), in which the numbers of factors is unknown. Hair et al29 went on to say that objectives of CFA are ‘‘1) to verify the proposed factor structure and 2) to explore if any significant modifications are needed.’’ Because one of the objectives in the study was to verify the factor structure with the TPB model, CFA was employed. Methodology suggested in previous research30 was used to evaluate the validity of the TPB by assessing the fit of the factors of TPB (attitudes, subjective norms, perceived behavior, and intention). Data were analyzed using AMOS version 23 (IBM, Chicago, IL, 2015). In CFA, the measured variables were modeled as a function of each latent variable, as identified in the Figure. As suggested by Hair et al,29 the factor loadings were first considered. Standard errors were then checked along with modification indices, unstandardized and standardized factor loadings, and factor covariance. Along
with chi-square, several types of overall model fit were measured, including the Comparative Fit Index (CFI), Normed Fit Index (NFI), Goodness of Fit Index, and root mean square error of approximation. According to Bentler,31 a value > 0.90 is representative of a well-fit model for both NFI and CFI. Goodness of Fit Index values range from 0 to 1, with 1 being a perfect fit.29 Root mean square error of
Figure. Theory of Planned Behavior. Adapted from Ajzen I. The theory of planned behavior. Organ Behav Hum Dec. 1991;50:182.
4 Stran et al
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
Table 2. Modified Theory of Planned Behavior Constructs and Variables With Standardized Loadings and Reliability (n ¼ 97)
Construct and Variables Attitudes (5 items) Using posted calorie information to make a meal decision is: Harmful/beneficial Worthless/useful Having calorie information posted at fast-food restaurants is: Harmful/beneficial Bad/good Worthless/useful
Mean (SD) 6.15 (0.99)
Standardized Loadings a .89
6.09 (1.04) 5.83 (1.40)
.84 .83
6.37 (1.03) 6.27 (1.32) 6.21 (1.10)
.88 .76 .88
Subjective norms (3 items) My parents think I should not/should use posted calorie information My friends think I should not/should use posted calorie information My close friends think I should not/should use posted calorie information
3.98 (1.21) 4.67 (1.32)
.63
3.93 (1.20)
.88
4.02 (1.39)
.91
Perceived behavioral control (1 item)
5.86 (1.46)
Intention (4 items) I expect to use posted calorie information to make a meal decision (strongly disagree/ strongly agree) I want to use posted calorie information to make a meal decision (strongly disagree/ strongly agree) I intend to use posted calorie information to make a meal decision (strongly disagree/ strongly agree) If calorie information were available in a fastfood restaurant, with what frequency would you use it? (never/often)
4.68 (1.54) 4.14 (1.76)
.83
4.87 (1.88)
.93
4.57 (1.84)
.95
5.13 (1.59)
.77
intentions (Table 2). Perceived behavioral control was assessed using only 1 question and so was not included in the factor analysis. An eigenvalue of > 1 and scree plots were used to determine the number of factors and questions with factor loadings > 0.40 were included in the respective factors.34 Internal consistencies of responses for each construct were assessed using Cronbach alpha. Five of the 7 questions assessing attitudes loaded onto 1 factor (Cronbach a ¼ .89) (Table 2). Two questions did not load onto any factor and were removed from the analysis (using calorie information posted on fast-food menus to make a meal decision is easy/difficult and convenient/inconvenient). The attitudes scale was constructed by averaging the responses of the 5 attitudes questions. Positive attitudes were represented by higher numbers. The 3
.86
.89
subjective norms questions also loaded onto 1 factor with Cronbach a ¼ .74. Subjective norms scores were averaged among the 3 questions, in which a higher number represented greater feelings of social pressure to use calorie labels. The 4-question intention scale had Cronbach a ¼ .89. Higher numbers represented greater intention to use calorie information if posted on fast-food menus. Cronbach a ¼ .7 is considered acceptable35; therefore, the modified attitudes, subjective norms, and intention scales were deemed appropriate for use in this study.
Potential Modifiers To assess the current health status of the students, 2 health status indicators were computed. Body mass index was calculated for each participant
using measured height and weight. Results were categorized based on National Institutes of Health guidelines into underweight and normal weight (<25 kg/m2) and overweight and obese ($25 kg/m2).36 Responses to the question about overall diet perceptions were dichotomized into an excellent/very good/good category and a fair/poor category, which is consistent with other research and provides a clear positive or negative view of current dietary practices.37 Race, housing status, current weight perception, dieting habits, use of the Nutrition Facts Panel, and frequency of fast-food visits were used as potential modifiers of intention and a difference in calories ordered from the menus. Barriers questions were used individually in the analyses as potential modifiers. The authors used univariate analyses (t tests and ANOVA) to determine significant differences (P < .05) in intention scores and changes in calories ordered for each potential covariate. Variables with significant differences were used in the final multiple regression analyses, in addition to gender and hunger status variables. These 2 variables were included to adjust for observed differences in calories ordered. Although the use of calorie information on the Nutrition Facts Panel was significant in univariate analyses, this variable was not included in multiple regression because this is closely related to the study design and could account for a large amount of the variance if included in the model. Multiple regression was used to test whether the 3 TPB factors would predict intention to use calorie labels while adjusting for confounders, or the variables that were significant in univariate analysis. The difference between total calories ordered from each of the 2 fast-food menus was calculated for each participant using the calorie information posted on the second menu. A paired t test was conducted to determine whether a significant difference existed in the number of calories ordered without posted calorie information vs with it. Multiple regression was used to test whether the 4 TPB factors would predict a change in the caloric content of meals ordered while adjusting for confounders that were significant from univariate analyses. P < .05 was used for all tests.
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
Stran et al 5
Table 3. Sociodemographic Characteristics, Health Status Indicators, Dietary Habits, Mean Intention Scores, and Mean Differences in Calories Ordered by College Students
Characteristics Gender (n ¼ 100) Male Female
n
%
Mean Intention Score (SD)a
37 63
37 63
4.37 (1.54) 4.86 (1.53)
Race (n ¼ 100) Non-Hispanic white Non-Hispanic black Other
74 18 8
74 18 8
4.69 (1.53) 4.38 (1.80) 5.22 (0.91)
Housing (n ¼ 100) On campus Off campus
49 51
49 51
4.64 (1.62) 4.88 (1.45)
Perceived diet quality (n ¼ 100) Excellent, very good, or good Fair or poor
59 41
59 41
4.93 (1.60) 4.32 (1.40)
Weight perception (n ¼ 97) Underweight About the right weight Overweight
6 72 19
6 74 20
3.08 (1.78) 4.55 (1.54) 5.41 (0.87)
Body mass index (n ¼ 100) < 25 kg/m2 $ 25 kg/m2
62 38
62 38
4.58 (1.70) 4.84 (1.25)
Dieting (n ¼ 100) Yes No
16 84
16 84
5.22 (1.64) 4.57 (1.51)
Worried about weight (n ¼ 100) Yes No
48 52
48 52
5.13 (1.35) 4.26 (1.61)
Typical fast-food visits (n ¼ 100) # 3 times/wk $ 4 times/wk
84 16
84 16
4.93 (1.49) 3.38 (1.14)
65
65
5.25 (1.16)
35
35
3.61 (1.61)
Nutrition Facts Panel use (n ¼ 100) Always, most of the time, sometimes Rarely, never
P .13
Mean Difference in Calories Ordered (SD)a
P .04
3.4 (296.5) 114.0 (291.9) .44
.18 80.9 (301.3) 28.1 (259.7) 196.3 (313.1)
.18
.36 42.4 (291.2) 97.5 (304.1)
.04
.59 84.0 (268.3) 51.2 (337.9)
.003
.33 55.8 (363.7) 51.2 (320.0) 126.8 (195.3)
.43
.31 94.4 (308.6) 31.6 (278.3)
.13
.80 53.1 (337.9) 73.9 (291.4)
.005
.26 105.5 (261.4) 38.3 (326.7)
< .001
< .001
.54 62.5 (299.5) 112.8 (293.4) .006 129.5 (291.3) 38.9 (281.3)
< .001
.44
Use of calorie information on Nutrition Facts Panel (n ¼ 100) Always, most of the time, sometimes Rarely, never
79
79
5.08 (1.35)
82.4 (292.3)
21
21
3.14 (1.23)
26.0 (320.1)
Calories needed in 1 d (n ¼ 94) # 1,500 1,501–2,000 $ 2,001
28 31 35
30 33 37
4.81 (1.40) 4.85 (1.62) 4.58 (1.38)
.72
.32 89.3 (270.2) 121.6 (312.9) 11.4 (319.8)
t tests and ANOVAs were used to determine differences in mean intention scores and mean differences in calories ordered. Note: Bonferroni correction was used for multiple comparisons (P < .016). a
RESULTS Demographics The majority of participants were women (63%), non-Hispanic white (74%), and freshmen or sophomores
(67%) (Table 3). Mean age was 20.0 years (SD, 1.27 years; range, 19–26 years). Whereas nearly half of participants reported that they were worried about their weight (48%), only 16% were dieting and 74% perceived
themselves to be about the right weight. A total of 79% of participants reported using the calorie information on a label at least sometimes whereas 65% reported using the Nutrition Facts Panel at least
6 Stran et al
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
Table 4. Predictors of Intention to Use Posted Calorie Information in Fast-Food Restaurants (n ¼ 97)* Variable
b
Gender
0.13
0.26
0.50
.62
Attitudes
0.37
0.13
2.84
.006
SE
t
P
Subjective norms
0.42
0.11
3.77
< .001
Perceived behavioral control
0.45
0.18
2.51
.01
Perceived diet quality
0.56
0.25
2.23
.03
Perceived weight status
0.01
0.16
0.05
.96
Worried about weight
0.51
0.27
1.93
.06
Frequency of dining at fast-food restaurants
0.67
0.35
1.92
.06
Time barrier
0.01
0.08
0.16
.87
Ease barrier
0.10
0.14
0.70
.48
Cost barrier
0.02
0.08
0.25
.80
0.03
0.07
0.43
.67
0.06
0.08
0.71
.48
Hunger barrier Hunger status during participation
*Multivariate coefficient ¼ 0.52; F ¼ 6.92; P < .001.
sometimes when deciding to buy a food product. Mean number of times respondents visited fast-food restaurants was 2.61 times/wk.
Intention Table 3 depicts mean intention scores by sociodemographic characteristics, health status indicators, and dietary habits. In univariate analysis, significant differences in mean intention scores were found for perceived diet quality (t ¼ 1.98; P < .05), perceived weight status (t ¼ 6.24; P ¼ .003),
worry about weight (t ¼ 2.89; P ¼ .005), and frequency of fast-food restaurant visits (t ¼ 3.95; P < .001). In multivariate analysis, attitudes, subjective norms, and perceived behavioral control predicted higher intention to use calorie information in a fast-food restaurant. Students reporting fair/poor diet quality (t ¼ 2.23; P ¼ .03) had significantly lower intention scores than their counterparts (Table 4). These variables explained 52% of the variance in intention scores (multivariate coefficient, 0.52; F ¼ 6.92; P < .001).
Table 5. Predictors of a Difference in Calorie Content of Meals Ordered From a Fast-Food Restaurant Menu (n ¼ 97)* Variable Gender
b 86.3
SE 63.1
t 1.37
P .18
Attitudes
23.6
33.7
0.70
.49
5.4
27.1
0.20
.84
Perceived behavioral control
31.6
45.5
0.69
.49
Intention
Subjective norms
31.7
24.3
1.30
.18
Time barrier
27.0
18.3
1.48
.14
Ease barrier
43.9
35.1
1.25
.21
Cost barrier
37.4
18.3
2.04
.04
Hunger barrier
36.7
16.4
2.23
.03
Hunger status during participation
5.1
19.3
0.27
.79
*Multivariate coefficient ¼ 0.18; F ¼ 1.92; P ¼ .05.
Difference in Calories Ordered Participants in this study ordered significantly fewer calories (mean ¼ 70.6; SD ¼ 298) when selecting from the menu with labels vs the one without labels (t ¼ 2.37; P ¼ .02). Mean caloric content for an order without labels was 908.9 cal (SD ¼ 290.1) vs 838.3 cal (SD, 341.5) when menu labels were present. In univariate testing, significant changes in the difference in calories ordered were found for gender (t ¼ 1.93; P ¼ .05) and hunger status (F ¼ 2.13; P ¼ .04) (Table 3). In multivariate analysis, significant predictors of a difference in the calorie content of meals ordered were the hunger barrier and cost barrier (Table 5). These variables explained 18% of the variance in differences in calorie content of meals ordered (multivariate coefficient, 0.18; F ¼ 1.92; P ¼ .05). As scores for agreement with the statement ‘‘The decision to use calorie information in a fast-food restaurant depends on how hungry I am’’ increased by 1 unit, participants ordered 36.7 more calories from the labeled menu than when the menu was not labeled. As scores for agreement with the statement ‘‘The decision to use calorie information in a fast-food restaurant is based on the cost of the menu items’’ increased by 1 unit, participants ordered 37.4 fewer calories from the labeled menu vs the menu that was not labeled. Gender (t ¼ -1.37; P ¼ .18) and hunger status at the time of participation (t ¼ 0.27; P ¼ .79) were not significant predictors in the difference in calories ordered among participants.
DISCUSSION The TPB model suggests that attitudes, social norms, and perceived behavioral control will predict intention to change a behavior, which is the strongest predictor of behavior. The TPB model predicted intention to use calorie information in a fast-food restaurant but was not associated with a difference in the caloric content of the meal ordered when menus were labeled. As predicted, attitudes, subjective norms, and perceived behavioral control were significantly related
Journal of Nutrition Education and Behavior Volume -, Number -, 2015 to intention to use a menu label in this small sample of college students. However, these variables were not related to a change in calories ordered when the barriers of hunger and cost were added to the model. Barriers to engaging in the behavior (time, ease, cost, and hunger) did not affect intention to engage in the behavior, but hunger and cost were significantly related to differences in the caloric content of the orders when menus were labeled. Although a previous study16 also found that positive attitudes toward this behavior were associated with greater intention to use posted information, this study assessed all TPB constructs using a quasi-experimental design. In this study, hunger is a commonly perceived barrier to using menu labels in a fast-food restaurant. Participants who strongly agreed that the decision to use calorie information in a fastfood restaurant is based on hunger ordered more calories than did those who did not agree with this statement. Current hunger status was not predictive of a difference in calories ordered between nonlabeled and labeled menus. In a study of 324 college students, Horacek and Betts38 found that students who perceived their dietary choices to be strongly influenced by hunger and other cues consumed significantly more calories than those whose dietary choices were selected based on the nutritional aspects of the foods. The current findings also indicated that when participants perceived cost to be a barrier, they ordered fewer calories at a fast-food restaurant. Students concerned about their finances may have developed behaviors that simultaneously cut costs and calories. The authors did not measure food insecurity and socioeconomic status in this study. Future studies should assess these measures in light of recent research that suggests higher than expected rates of food insecurity among college students.39 Previous studies have also indicated that cost is a potential barrier to healthful eating practices by university students.23,40 At the time of this study, the menu labeling law was not fully implemented. Thus, students may not have the opportunity to practice and develop skills in ordering menu
items while considering the caloric content of the items against other known attributes of the food such as cost or personal preferences. Neither college students' weight status nor weight concerns were predictive of changes in calories ordered when menus were labeled in a fast-food restaurant. However, those who perceive their overall diet to be poor or fair had significantly lower intention to use the menu labels compared with their counterparts who perceive their diets to be excellent, very good, or good. In addition, frequency of fast-food consumption was not a significant predictor of a difference in the calories ordered between menus. Because this is a cross-sectional study, it cannot be inferred whether students who limit fast-food visits would value nutrition. The scope and interpretation of this study have limitations. Participants in this study examined a generic fast-food restaurant menu and ordered items they would most likely consume. An actual meal was not served or consumed; thus, an accurate calorie count of the meal was not assessed. Only 1 question was used to assess students' perceived behavioral control; therefore, it was not able to be used in the factor analysis. Although cost can be a barrier to healthful eating for college students,23,40 prices were not included on the menus used in this study. The researchers decided to leave price off the menu in an effort to understand the direct result of caloric information in the selections. In addition, although the menus used had a limited number of options, as in a fast-food restaurant, and the researchers asked participants to look over the menu quickly, the researchers could not mimic the fastpaced ordering procedures that are associated with visiting a fast-food counter or drive-through. Dietary restraint has a role in food selection. This study measured only self-report of dieting, and a more thorough means of assessing dietary restraint is needed in future studies. Because this study was conducted at 1 southeastern university, the results may not be applicable to students with a more diverse background or to nonstudents.
Stran et al 7
IMPLICATIONS FOR RESEARCH AND PRACTICE Results from this study suggest that students weigh cost and hunger into their decisions, and these barriers to healthy eating habits are also barriers to the use of menu labels in fast-food restaurants. The TPB predicted intention to use menu labels but did not predict changes in calories ordered between menus. After viewing the menu with calorie labels, participants in this small study were able to select a lower-calorie meal. Educational programs are needed aimed at decreasing perceived barriers, such as cost and mindful eating practices. Future research should investigate the influence of students' barriers in a fastfood restaurant setting. In addition, because this study did not measure participants' intake of the meals ordered, future studies should investigate whether the TPB is predictive of both intention and a change in calories ordered and consumed in a fast-food restaurant setting. Restaurant menu labeling in fast-food restaurants could provide the information college students need to select lowercalorie items.
ACKNOWLEDGMENTS The authors would like to thank the University of Alabama Coordinated Program students for their assistance with subject recruitment and data collection.
REFERENCES 1. National Restaurant Association. 2015 Restaurant Industry Pocket Factbook. 2015. https://www.restaurant.org/Downloads/ PDFs/News-Research/research/Factbook 2015_LetterSize-FINAL.pdf. Accessed July 23, 2015. 2. Lin BH, Guthrie J, Fraz~ao E. Nutrient contribution of food away from home. In: Fraz~ao E, ed. American’s Eating Habits: Changes and Consequences. Washington, DC: US Dept of Agriculture; 1999:213-242. 3. Scourboutakous MJ, L’Abbe MR. Restaurant menus: calories, calorie density, and serving size. Am J Prev Med. 2012;43:249-255.
8 Stran et al 4. Burton S, Creyer EH, Kees J, Huggins K. Attacking the obesity epidemic: the potential health benefits of providing nutrition information in restaurants. Am J Public Health. 2006;96:1669-1675. 5. Young LR, Nestle M. Reducing portion sizes to prevent obesity: a call to action. Am J Prev Med. 2012;43:565-568. 6. United States Food and Drug Administration. Patient Protection and Affordable Care Act, Public Law 111–148, Sec. 4205–Nutrition labeling of standard menu items at chain restaurants. March 23, 2010. http://www.gpo.gov/ fdsys/pkg/FR-2010-08-25/pdf/2010-21065. pdf. Accessed July 23, 2015. 7. Racette SB, Deusinger SS, Strube MJ, Highstein GR, Deusinger RH. Weight changes, exercise, and dietary patterns during freshman and sophomore years of college. J Am Coll Health. 2005;53: 245-251. 8. Bassett MT, Dumanovsky T, Huang C, et al. Purchasing behavior and calorie information at fast-food chains in New York City, 2007. Am J Public Health. 2008;98:1457-1459. 9. Dumanovsky T, Huang CY, Nonas CA, Matte TD, Bassett MT, Silver LD. Changes in energy content of lunchtime purchases from fast food restaurants after introduction of calorie labelling: cross sectional customer surveys. BMJ. 2011;343:d4464. 10. Farley TA, Caffarelli A, Bassett MT, Silver L, Frieden TR. New York City’s fight over calorie labeling: a two year struggle ultimately proves that innovation in food regulation is entirely possible at the local level. Health Affairs. 2009;28:w1098-w1109. 11. Finkelstein EA, Strombotne KL, Chan NL, Krieger J. Mandatory menu labeling in one fast-food chain in King County, Washington. Am J Prev Med. 2011;40:122-127. 12. Lando AM, Labiner-Wolfe J. Helping consumers make more healthful food choices: consumer views on modifying food labels and providing point-ofpurchase nutrition information at quick-service restaurants. J Nutr Educ Behav. 2007;39:157-163. 13. Piron J, Smith LV, Simon P, Cummings PL, Kuo T. Knowledge, attitudes and potential response to menu labelling in an urban public health clinic population. Public Health Nutr. 2009;13: 550-555. 14. Vadiveloo MK, Dixon LB, Elbel B. Consumer purchasing patterns in
Journal of Nutrition Education and Behavior Volume -, Number -, 2015
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
response to calorie labeling legislation in New York City. Int J Behav Nutr Phys Act. 2011;8:51. http://dx.doi.org/10.1186 /1479-5868-8-51. Wisdom J, Downs J, Loewenstein G. Promoting healthy choices: information vs. convenience. Am Econ J Appl Econ. 2010;2:164-178. Bates K, Burton S, Howlett E, Huggins K. The roles of gender and motivation as moderators of the effects of calorie and nutrition information provision on away-from-home foods. J Consum Aff. 2009;43:249-273. Gerend MA. Does calorie information promote lower calorie fast food choices among college students? J Adolesc Health. 2009;44:84-86. Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process. 1991;50:182. Pawlak R, Cerutti CS, Quinton R. Taking an undergraduate nutrition course results in favorable attitudes toward a healthful diet and improved intake of several key nutrients. Fam Consum Sci Res J. 2009;38:3-10. Centers for Disease Control and Prevention. National Center for Health Statistics (NCHS). National Health and Nutrition Examination Survey Questionnaire. Hyattsville, MD: US Dept of Health and Human Services, Centers for Disease Control and Prevention, 2007-2008. http://wwwn.cdc.gov/nchs/ nhanes/search/nhanes07_08.aspx. Accessed July 24, 2015. Centers for Disease Control and Prevention. National Center for Health Statistics (NCHS). National Health and Nutrition Examination Survey Data. Hyattsville, MD: US Dept of Health and Human Services, Centers for Disease Control and Prevention, 2009-2010. http://wwwn.cdc. gov/nchs/nhanes/search/nhanes09_10. aspx. Accessed July 24, 2015. King KA, Mohl K, Bernard AL, Vidourek RA. Does involvement in healthy eating among university students differ based on exercise status and reasons for exercise? Calif J Health Promot. 2007;5(3):106-119. LaCaille LJ, Dauner KN, Krambeer RJ, Pedersen J. Psychosocial and environmental determinants of eating behaviors, physical activity, and weight change among college students: a qualitative analysis. J Am Coll Health. 2011; 59:531-538. Silliman K, Rodas-Fortier K, Neyman M. A survey of dietary and
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
exercise habits and perceived barriers to following a healthy lifestyle in a college population. Californian J Health Promot. 2004;2:10-19. Rydell SA, Harnack LJ, Oakes JM, Story M, Jeffrey RW, French SA. Why eat at fast food restaurants: reported reasons among frequent customers. J Am Diet Assoc. 2008;108: 2066-2070. Driskell JA, Meckna BR, Scales NE. Differences exist in the eating habits of university men and women at fast food restaurants. Nutr Res. 2006;26: 524-530. Francis JJ, Eccles MP, Johnston M, et al. Constructing Questionnaires Based on the Theory of Planned Behavior: A Manual for Health Services Researchers. Newcastle-upon-Tyne, UK: University of Newcastle; 2004. Schumacker RE, Lomax RG. A Beginner’s Guide to Structural Equation Modeling. 2nd ed. Mahwah, NJ: Lawrence Erlbaum Associates; 2004. Hair JF, Anderson RE, Tatham RL, Black WC. Multivariate Data Analysis. 5th ed. Upper Saddle River, NJ: Prentice Hall; 1998. Birch LL, Fisher JO, GrimmThomas K, Markey CN, Sawyer R, Johnson SL. Confirmatory factor analysis of the Child Feeding Questionnaire: a measure of parental attitudes, beliefs, and practices about child feeding and obesity proneness. Appetite. 2001;36:201-210. Bentler PM. Comparative fit indexes in structural equation models. Psychol Bull. 1990;88:588-606. Browne MW, Cudeck R. Alternative ways of assessing model fit. In: Bollen K, Long J, eds. Testing Structural Equation Models. Newbury Park, CA: Sage; 1993:136-192. Suhr DD. Exploratory or confirmatory factor analysis? Proceedings of the Thirty-first Annual SAS Users Group International Conference. Cary, NC: SAS Institute; 2006. O’Rourke N, Hatcher L, Stepanski EJ. A Step-by-Step Approach to Using SAS for Univariate and Multivariate Statistics. 2nd ed. Cary, NC: SAS Institute, Inc; 2005. Nunnally J, Bernstein L. Psychometric Theory. New York, NY: McGrawHill Higher, Inc; 1994. National Institutes of Health. The Practical Guide: Identification, Evaluation, and Treatment of Overweight and Obesity in Adults (NIH Pub No. 00–4084).
Journal of Nutrition Education and Behavior Volume -, Number -, 2015 October 2000. http://www.nhlbi.nih. gov/guidelines/obesity/prctgd_c.pdf. Accessed July 24, 2015. 37. Goodwin DK, Knol LK, Eddy JM, Fitzhugh EC, Kendrick OW, Donahue RE. The relationship between self-rated health status and the overall quality of dietary intake of US
adolescents. J Am Diet Assoc. 2006;106: 1450-1453. 38. Horacek TM, Betts NM. Students cluster into 4 groups according to the factors influencing their dietary intake. J Am Diet Assoc. 1998;98:1464-1467. 39. Gaines A, Robb CA, Knol LL, Sickler S. Examining the role of financial factors,
Stran et al 9 resources and skills in predicting food security status among college students. Int J Consumer Sci. 2014;38:374-384. 40. Garcia AC, Sykes L, Matthews J, Martin N, Leipert B. Perceived facilitators of and barriers to healthful eating among university students. Can J Diet Pract Res. 2010;71:e28-e33.
9.e1 Stran et al
CONFLICT OF INTEREST The authors have not stated any conflicts of interest.
Journal of Nutrition Education and Behavior Volume -, Number -, 2015