This article was downloaded by: [University of Nebraska, Lincoln] On: 08 October 2014, At: 10:19 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of the American College of Nutrition Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uacn20
Associations Between Eating Occasion Characteristics and Age, Gender, Presence of Children and BMI Among U.S. Adults a
c
b
Marla Reicks PhD , Dennis Degeneffe MBA , Aaron Rendahl PhD , Marianne Smith Edge MS d
e
f
f
RD , Katie Burns BA , Brian O’Meara MS & Greg Blevins MBA a
Department of Food Science and Nutrition, University of Minnesota, St. Paul, Minnesota
b
School of Statistics, University of Minnesota, St. Paul, Minnesota
c
Consumer Centric Solutions, LLC, St. Paul, Minnesota
d
International Food Information Council Foundation, Washington, D.C.
e
APCO Worldwide, Washington, D.C.
f
Blevins-O'Meara Insights, Bloomington, Minnesota Published online: 20 Aug 2014.
To cite this article: Marla Reicks PhD, Dennis Degeneffe MBA, Aaron Rendahl PhD, Marianne Smith Edge MS RD, Katie Burns BA, Brian O’Meara MS & Greg Blevins MBA (2014) Associations Between Eating Occasion Characteristics and Age, Gender, Presence of Children and BMI Among U.S. Adults, Journal of the American College of Nutrition, 33:4, 315-327, DOI: 10.1080/07315724.2014.887485 To link to this article: http://dx.doi.org/10.1080/07315724.2014.887485
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Original Research
Associations Between Eating Occasion Characteristics and Age, Gender, Presence of Children and BMI Among U.S. Adults
Downloaded by [University of Nebraska, Lincoln] at 10:20 08 October 2014
Marla Reicks, PhD, Dennis Degeneffe, MBA, Aaron Rendahl, PhD, Marianne Smith Edge, MS, RD, Katie Burns, BA, Brian O’Meara, MS, Greg Blevins, MBA Department of Food Science and Nutrition (M.R.), and School of Statistics (A.R.), University of Minnesota, St. Paul, Minnesota; Consumer Centric Solutions, LLC, St. Paul, Minnesota (D.D.); International Food Information Council Foundation, Washington, D.C. (M.S.E.); APCO Worldwide, Washington, D.C. (K.B.); Blevins-O’Meara Insights, Bloomington, Minnesota (B.O., G.B.) Key words: diets, obesity, general nutrition, pediatric and geriatric considerations, preventative nutrition and chronic disease Objective: To describe how frequency and characteristics of traditional meal and non-meal occasions vary by age, gender, presence of children, and body mass index (BMI). Design: A cross-sectional survey was administered to a national demographically balanced sample of adults via an online market research panel. Setting: Online survey. Subjects: Survey respondents were in the 18- to 80-year-old age range and had consumed any food or beverage at home or away from home the previous day. The sample included 2702 adults reporting on 6689 eating/drinking occasions. Most (80.3%) had no children at home; 43.5% were male and about two thirds were overweight/obese. Measures of outcome: Eating occasion characteristics and goals by age, gender, presence of children, and BMI. Results: Older respondents were more likely to report planning traditional meal occasions and report on a breakfast occasion than younger respondents. Two prominent reasons that triggered consumption occasions were habit and hunger/thirst with one dominant benefit of satisfying hunger or thirst. Habit and nutrition played a larger role as a goal for eating occasions for older compared to younger respondents. When children were present in the household, respondents had a goal of connecting with “family, friends, or colleagues” at dinner compared to those without children. Few gender differences were noted; however, women more often reported goals of satisfying hunger/thirst and taste at lunch than men. BMI levels were related to a range of triggers, goals, and behaviors but not as prominently as the relationships observed with age. Those with BMI ≥ 30 were less health conscious regarding dinner and breakfast consumption compared to those with a lower BMI. Conclusions: Among demographic variables, age differences were noted in relation to eating occasion characteristics more often than other demographic characteristics or BMI. Understanding these differences can be beneficial in tailoring promotion of healthful intake at specific eating occasions for particular subgroups.
INTRODUCTION
sions per day from 1977 (5.9 eating occasions) to 2006 (6.6 eating occasions) and a decrease in the time between eating occasions by about an hour [3]. Nationally representative data (1988–1994) also showed that the most common eating patterns for U.S. adults were based on a traditional breakfast, lunch, and dinner and two snacks (32%) and breakfast, lunch, and dinner and one snack (15%) [4]. Current studies regarding the association between eating frequency and snacking frequency and obesity have produced conflicting results. In various demographic subgroups, studies have found positive [5,6], inverse [7,8], or
Eating patterns of individuals in the United States based on frequency of traditional meal and non-meal occasions have been implicated in the increasing prevalence of obesity [1]. Data from nationally representative samples of U.S. adults over the past 30 years have shown that total energy intake from snacks increased from 18% in 1977–1978 to 24% in 2003–2006 based on an increased frequency [2]. The increased frequency of snacking resulted in an increase in the total number of eating occa-
Address correspondence to: Marla Reicks, University of Minnesota, 1334 Eckles Avenue, St. Paul, MN 55108. E-mail: [email protected]
Journal of the American College of Nutrition, Vol. 33, No. 4, 315–327 (2014) C American College of Nutrition Published by Taylor & Francis Group, LLC 315
Downloaded by [University of Nebraska, Lincoln] at 10:20 08 October 2014
Eating Occasions by Demographic Characteristics no associations [9,10]. A recent study that examined relationships between eating frequency and snacking frequency and dietary intake showed that frequency was related to energy intake and not body mass index (BMI) in midlife women, with obese women having greater energy intake compared to normal-weight women who consumed the same number of meals and snacks [11]. Therefore, the types of foods chosen and the amounts consumed at any occasion may be more important than the number of occasions. Within eating occasions, several factors have been associated with the choice of foods and beverages. Previous studies have indicated that planning food intake in advance is related to better self-control, consumption of healthful foods, and advanced stage of change for healthful eating [12–14]. Implementation intentions have been defined as the generation of plans for when, where, and how one can meet dietary goals [15]. Results from a review indicated a positive relationship between implementation intentions for adults and consumption of healthful foods [15]. Previous studies have shown that primary factors in the initiation of daily food and beverage consumption occasions include mood, hunger, distraction [16], visual and olfactory cues [17], and habit [18]. These relationships have typically been studied with respect to eating behaviors characterized by excessive restraint [16] or disordered eating [19] or in laboratory settings; however, fewer studies have focused on triggers to eating occasions in everyday situations or with nonclinical samples. Previous studies have also focused on a single or very few factors that might initiate eating occasions as opposed to a more comprehensive approach. Several approaches have been used to explain need states or goals that an individual may have at specific eating occasions. One approach involved segmentation of eating occasions by situation-based needs or need states among midlife women [20]. Primary need states included indulgence, comfort, nutrition and health, nurturing others, and convenience, with variable intakes of energy, fat, sweets, fruits, and vegetables by eating occasion need segment [20]. Situational context was the basis for the differences in intake of sweetened beverages, cereals, and entr´ees among young adults who recorded eating occasion characteristics with personal digital assistants [13]. Qualitative 24-hour recalls were used in another study to characterize demands within eating occasions [21]. Results showed that adults followed purposeful routines based on daily schedules to meet demands and values. Means-end theory and practice [22] is another approach used to explain needs related to food choice. This theory would indicate that food choice is based on the perceived benefits that the food can provide to the individual. These benefits may be organized at several levels from lower order functional needs to higher level psychosocial needs. Studies of eating occasion characteristics have examined not only stimuli that initiate consumption or goals that impact food choice within the occasion but also motives for ending the occasion [23]. These motives have typically included internal cues,
316
such as satiation, or external cues, such as when others were finished eating, with a relationship observed between responsiveness to type of cue and weight status. Eating occasion characteristics and goals are likely to be dependent on the characteristics of the individual and his or her environment. For example, young adults may be less likely to have routine eating occasions based on instability in living arrangements, household composition, and employment compared to older adults. A large number of eating occasions of young adults were characterized as occurring without structure and in less traditional settings [13], which could contribute to less healthful food choices. In another study, eating frequency among adults varied by age group, education, income, and ethnicity but not by gender [4]. However, the type of foods selected for snacks varied by gender in a sample of Swiss adults, with women choosing more healthful options than men [24]. Using longitudinal data over 7 years from 3 urban centers, Laroche et al. [25] found that becoming a parent did not have major effects on food or nutrient intake; however, intake of saturated fat decreased compared to nonparents. Parents also experienced a nonsignificant increase in energy intake of 79 calories/day compared to nonparents (p = 0.058) and mothers gained more weight than women who did not become a mother over the same period [26]. Eating occasion characteristics and needs may also differ based on a tendency for emotional eating, level of responsivity to environmental cues, and extent of disinhibition, which may vary with weight status [27]. A better understanding of demographic characteristics and BMI on eating occasion situational factors and goals will allow for effective tailoring of interventions to improve eating patterns and the type and amounts of foods consumed at specific eating occasions. To date, many studies have examined the influence of triggers, goals, and planning or the influence of demographic characteristics and BMI on eating patterns and intake, but these studies have typically not also focused on differences by eating occasion type [4,15–18,25–27]. Therefore, the purpose of this study was to describe how frequency and characteristics of traditional meal and non-meal occasions, including triggers and goals (based on rational, physiological, and emotional dimensions), vary by demographic and physical characteristics including age, gender, presence of children, and BMI.
MATERIALS AND METHODS Overview A 2-phase study was conducted by Blevins-O’Meara Insights (BOI), an independent marketing research firm, commissioned by the International Food Information Council (IFIC) Foundation. Preliminary focus group interviews in 3 regions of the United States identified potential patterns and differences with respect to situational context and characteristics and
VOL. 33, NO. 4
Eating Occasions by Demographic Characteristics
Downloaded by [University of Nebraska, Lincoln] at 10:20 08 October 2014
emotions and motivations for specific eating and drinking occasions. These focus group findings were used as the basis for the development of an online questionnaire administered to a nationally representative sample of adults 18–80 years of age to describe perceptions and approaches related to daily eating and drinking occasions including frequency, underlying motivations, and situational context. The study was intended to provide a broad description of the goals and motivations that drive food consumption in the United States. This article will focus only on selected measures to describe the relationship between age, gender, presence of children in the household, and BMI and eating occasion characteristics and needs/goals.
Subjects In the first formative phase, a series of 12 2-hour focus group interviews were conducted with a demographically reflective sample of the current U.S. population, including men and women aged 18–69 in 3 cities (Minneapolis, Minnesota [n = 15], Atlanta, Georgia [n = 15], and Phoenix, Arizona [n = 12]). The qualitative research was designed to generate a broad set of ideas, behaviors, and language to include in the quantitative survey. A diverse set of consumers was screened to include people from various demographic groups and provide representation across major U.S. Census regions. Several respondents from each of the following demographic groups were included: younger (18– 44 years) and older (45–69 years); male and female; Hispanic, African American, and white; married and single; having either no children, young children (0–3 years), or older (4 + years) children at home; and being employed or not employed. Because geographic dispersion was also considered valuable in addressing the broad objective of the study, 42 respondents were interviewed across these demographic variables in Atlanta, Minneapolis, and Phoenix. For the second phase, an online survey was conducted using a national online panel of consumers who had been recruited and maintained by Research Now, a private research company under the parent company e-Rewards, headquartered in Plano, Texas. Research Now uses a variety of recruiting practices to build and maintain their panel, including the recruiting of credit card holders, airline flyers, and magazine subscribers. This has resulted in a panel that includes about 6.5 million respondents. From the Research Now panel, an initial sample of individuals was selected so that it included a nationally representative geodemographically balanced cross section of consumers on the basis of age range, household income, gender, region, and race/ethnicity [28]. The selected panel members were sent email invitations to participate in a survey and were provided a Research Now web link. Upon entry to the website, respondents were asked to verify that they were in the 18- to 80-year-old age range and that they had consumed any food or beverage either at home or away from home the previous day. Those who qualified to participate in the survey by meeting these require-
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION
ments then continued with the remainder of the questionnaire. Respondents were informed of the remainder as follows: “You will be asked to recall how many different eating or drinking occasions you had yesterday and what you ate and/or drank on each occasion.” For their participation, respondents were paid $5.50 by Research Now in electronic currency redeemable for various consumer goods and services. Response rates varied by group, with lower responses from groups including those with low income, those who were ethnically diverse, young males, and those with lower education levels. Discrepancies were mitigated by first going through supplemental waves of invitations to those groups that were underrepresented and, second, by weighting the data. Weighting was performed using the anesrake R package (ANES Raking Implementation, version 0.70, 2011, J. Pasek, University of Michigan). Deidentified data, with no link to identifiers, was provided by BOI to the authors of this article for use in the analysis. Therefore, this project did not meet the definition of human subjects research by the University of Minnesota Institutional Review Board and did not require University of Minnesota Institutional Review Board review.
Formative Research and Survey Development Results from the preliminary focus group interviews described what constitutes an eating or drinking occasion and how the type of eating occasion was related to the emotions, social context, and mindfulness throughout a day. The variety of personal participant characteristics allowed for a differentiation in eating occasion characteristics based on the presence of children, employment, age, and tendency to plan. Various triggers to consumption occasions were identified as well as goals that varied with personal characteristics and cues to end the occasion. These triggers and goals were similar to those identified in previous focus group and individual interviews with adults [29,30]. Women identified similar internal and external drivers perceived as needs surrounding eating occasions reflecting a wide range in emotional gratification [29]. From interviews with adults, factors influencing food choice decisions were identified from 5 main categories of cultural ideals, personal factors, resources, social factors, and present contexts [30]. The triggers and goals identified in the current study were based on similar factors. Focus group findings were also used to develop definitions for eating occasions used in the survey, particularly for nontraditional occasions such as snacks, treats, and mini-meals. The consistency of terminology used by focus groups participants for these occasions contributed to the formulation of the terms and ensured that they were distinct from one another. All measures included in the survey were developed by BOI and subject to review by IFIC Foundation and IFIC Foundation’s stakeholders with the objective of providing a broad description of the goals and motivations that drive food consumption in the United States. The measures that provided data for this article
317
Downloaded by [University of Nebraska, Lincoln] at 10:20 08 October 2014
Eating Occasions by Demographic Characteristics included questions dealing with planning, triggers for deciding to engage in the consumption occasion, goals, and cues for ending the occasion. The questionnaire was divided into 3 parts. The first part involved a series of respondent-level questions that asked about food and beverage categories consumed on the previous day and the number of eating and drinking occasions that the respondent had the previous day by time of day. They were given specific instructions to report for a 24-hour period from midnight to midnight. Respondents were instructed that each time they began to eat or drink something on the previous day it should be thought of as a separate occasion. They were asked to indicate the number of separate occasions (1, 2, 3 or more) for each of 7 time periods where they had consumed anything. In the second part of the questionnaire, respondents were asked to report on no more than 3 of the occasions they listed in part 1. If more than 3 eating and/or drinking occasions were listed in part 1, 3 separate eating and/or drinking occasions were randomly selected for the respondent to report on. Then for each occasion, respondents were asked a series of occasion-level questions about what was consumed, what kind of occasion it was (for this they were asked which of the following best describes the occasion—breakfast: standard morning occasion, lunch: standard mid-day occasion, dinner: standard evening occasion, snack: an occasion between standard meals, treat: an infrequent occasion that includes an especially good tasting food/beverage, mini-meal: small meal that is typically consumed at multiple occasions through the day, or other) and the situational variables surrounding the occasion such as when the occasion occurred, who was there, and what the respondent was doing. Respondents were asked about their planning of the occasion, triggers (or cues) experienced, and goals they had for the occasion. Lastly, the respondents were asked another series of respondent-level questions that included questions about demographic characteristics and height and weight to enable calculation of BMI.
Pretesting Several IFIC Foundation staff members and trustees with nutrition and evaluation expertise and other IFIC Foundation stakeholders reviewed the survey in its entirety multiple times before administration. A pretest was conducted in which approximately 100 respondents completed the survey to determine whether reasonable patterns of response were forming. Response patterns were examined for consistency with the expectations formed by the qualitative formative process and phrasing for some questions was modified according to findings. In addition, Research Now examined the pretest data to identify points in the survey where respondents were having problems. Then those areas were reviewed for clarity or any dimensions that potentially caused problems for the respondent and revised to address these issues prior to final administration.
318
Survey Administration Overall the survey averaged 23 minutes to complete. The fieldwork was conducted between March 10 and April 14, 2013. From an estimate of initial invitations (n = 74,660), 8444 respondents (11%) clicked to begin screening, with 172 of these screened out by the screening criteria. Based on quotas, 3544 were over quota (they qualified on the original set of criteria but were in quota groups that had already been filled when they began screening). A number of potential respondents (n = 1871) abandoned the survey after qualifying, resulting in 2857 respondents completing the survey. Of the 2857 respondents completing the survey, data for some were not available for gender (n = 1), presence of children (n = 54), and height and/or weight (n = 89). Respondents ≥65 years of age with children in the household (n = 11) were not included. Data from these 155 respondents were not included in the analysis, resulting in a final sample of 2702 respondents.
Occasions The aim of this study was to provide an estimate of the true proportion of eating occasions. Therefore, the distribution of the occasions was reapportioned such that the sample of occasions was weighted to proportionally give each day of the week equal contribution (14% of total occasions). No other reapportioning of the occasion data was done. In support of this treatment, an analysis was done where a comparison was made of weighted versus unweighted occasion data for the questions regarding triggers and goals. Results between the two never differed by more than 1% point, so the effect of the weighting by day of the week is marginal. The initial 2857 respondents reported on 8141 eating/drinking occasions. After removing data from 155 respondents with missing data or who had children but were ≥65 years of age, 7713 occasions remained. Data from occasions where respondents classified occasion type as “other” were also removed (n = 457 occasions) leaving 7256 occasions for further analysis (Table 1). To avoid pseudoreplication from having multiple data points from the same individual, for individuals who reported on multiple eating occasions of a given type, a single occasion of each type was randomly chosen resulting in the elimination of an additional 567 occasions for a final sample of 6689 eating/drinking occasions.
Statistical Analyses Statistical analysis was completed using R, version 3.0.1 [31]. Descriptive statistics were used to describe characteristics of the sample, eating occasion frequency, characteristics, and goals. Association between eating occasions and BMI was tested for using quantile regression of the median due to notable skew in BMI. Eating occasions were totaled for each individual
VOL. 33, NO. 4
Eating Occasions by Demographic Characteristics Table 1. Eating Occasions (Unsampled)1 by Demographic Characteristics Variable All (n = 7256) Gender Children
Downloaded by [University of Nebraska, Lincoln] at 10:20 08 October 2014
Age
BMI
Currently trying to lose weight
Category Male Female Yes No 18–24 25–34 35–44 45–54 55–64 65–74 75–80
Journal of the American College of Nutrition Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uacn20
Associations Between Eating Occasion Characteristics and Age, Gender, Presence of Children and BMI Among U.S. Adults a
c
b
Marla Reicks PhD , Dennis Degeneffe MBA , Aaron Rendahl PhD , Marianne Smith Edge MS d
e
f
f
RD , Katie Burns BA , Brian O’Meara MS & Greg Blevins MBA a
Department of Food Science and Nutrition, University of Minnesota, St. Paul, Minnesota
b
School of Statistics, University of Minnesota, St. Paul, Minnesota
c
Consumer Centric Solutions, LLC, St. Paul, Minnesota
d
International Food Information Council Foundation, Washington, D.C.
e
APCO Worldwide, Washington, D.C.
f
Blevins-O'Meara Insights, Bloomington, Minnesota Published online: 20 Aug 2014.
To cite this article: Marla Reicks PhD, Dennis Degeneffe MBA, Aaron Rendahl PhD, Marianne Smith Edge MS RD, Katie Burns BA, Brian O’Meara MS & Greg Blevins MBA (2014) Associations Between Eating Occasion Characteristics and Age, Gender, Presence of Children and BMI Among U.S. Adults, Journal of the American College of Nutrition, 33:4, 315-327, DOI: 10.1080/07315724.2014.887485 To link to this article: http://dx.doi.org/10.1080/07315724.2014.887485
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Original Research
Associations Between Eating Occasion Characteristics and Age, Gender, Presence of Children and BMI Among U.S. Adults
Downloaded by [University of Nebraska, Lincoln] at 10:20 08 October 2014
Marla Reicks, PhD, Dennis Degeneffe, MBA, Aaron Rendahl, PhD, Marianne Smith Edge, MS, RD, Katie Burns, BA, Brian O’Meara, MS, Greg Blevins, MBA Department of Food Science and Nutrition (M.R.), and School of Statistics (A.R.), University of Minnesota, St. Paul, Minnesota; Consumer Centric Solutions, LLC, St. Paul, Minnesota (D.D.); International Food Information Council Foundation, Washington, D.C. (M.S.E.); APCO Worldwide, Washington, D.C. (K.B.); Blevins-O’Meara Insights, Bloomington, Minnesota (B.O., G.B.) Key words: diets, obesity, general nutrition, pediatric and geriatric considerations, preventative nutrition and chronic disease Objective: To describe how frequency and characteristics of traditional meal and non-meal occasions vary by age, gender, presence of children, and body mass index (BMI). Design: A cross-sectional survey was administered to a national demographically balanced sample of adults via an online market research panel. Setting: Online survey. Subjects: Survey respondents were in the 18- to 80-year-old age range and had consumed any food or beverage at home or away from home the previous day. The sample included 2702 adults reporting on 6689 eating/drinking occasions. Most (80.3%) had no children at home; 43.5% were male and about two thirds were overweight/obese. Measures of outcome: Eating occasion characteristics and goals by age, gender, presence of children, and BMI. Results: Older respondents were more likely to report planning traditional meal occasions and report on a breakfast occasion than younger respondents. Two prominent reasons that triggered consumption occasions were habit and hunger/thirst with one dominant benefit of satisfying hunger or thirst. Habit and nutrition played a larger role as a goal for eating occasions for older compared to younger respondents. When children were present in the household, respondents had a goal of connecting with “family, friends, or colleagues” at dinner compared to those without children. Few gender differences were noted; however, women more often reported goals of satisfying hunger/thirst and taste at lunch than men. BMI levels were related to a range of triggers, goals, and behaviors but not as prominently as the relationships observed with age. Those with BMI ≥ 30 were less health conscious regarding dinner and breakfast consumption compared to those with a lower BMI. Conclusions: Among demographic variables, age differences were noted in relation to eating occasion characteristics more often than other demographic characteristics or BMI. Understanding these differences can be beneficial in tailoring promotion of healthful intake at specific eating occasions for particular subgroups.
INTRODUCTION
sions per day from 1977 (5.9 eating occasions) to 2006 (6.6 eating occasions) and a decrease in the time between eating occasions by about an hour [3]. Nationally representative data (1988–1994) also showed that the most common eating patterns for U.S. adults were based on a traditional breakfast, lunch, and dinner and two snacks (32%) and breakfast, lunch, and dinner and one snack (15%) [4]. Current studies regarding the association between eating frequency and snacking frequency and obesity have produced conflicting results. In various demographic subgroups, studies have found positive [5,6], inverse [7,8], or
Eating patterns of individuals in the United States based on frequency of traditional meal and non-meal occasions have been implicated in the increasing prevalence of obesity [1]. Data from nationally representative samples of U.S. adults over the past 30 years have shown that total energy intake from snacks increased from 18% in 1977–1978 to 24% in 2003–2006 based on an increased frequency [2]. The increased frequency of snacking resulted in an increase in the total number of eating occa-
Address correspondence to: Marla Reicks, University of Minnesota, 1334 Eckles Avenue, St. Paul, MN 55108. E-mail: [email protected]
Journal of the American College of Nutrition, Vol. 33, No. 4, 315–327 (2014) C American College of Nutrition Published by Taylor & Francis Group, LLC 315
Downloaded by [University of Nebraska, Lincoln] at 10:20 08 October 2014
Eating Occasions by Demographic Characteristics no associations [9,10]. A recent study that examined relationships between eating frequency and snacking frequency and dietary intake showed that frequency was related to energy intake and not body mass index (BMI) in midlife women, with obese women having greater energy intake compared to normal-weight women who consumed the same number of meals and snacks [11]. Therefore, the types of foods chosen and the amounts consumed at any occasion may be more important than the number of occasions. Within eating occasions, several factors have been associated with the choice of foods and beverages. Previous studies have indicated that planning food intake in advance is related to better self-control, consumption of healthful foods, and advanced stage of change for healthful eating [12–14]. Implementation intentions have been defined as the generation of plans for when, where, and how one can meet dietary goals [15]. Results from a review indicated a positive relationship between implementation intentions for adults and consumption of healthful foods [15]. Previous studies have shown that primary factors in the initiation of daily food and beverage consumption occasions include mood, hunger, distraction [16], visual and olfactory cues [17], and habit [18]. These relationships have typically been studied with respect to eating behaviors characterized by excessive restraint [16] or disordered eating [19] or in laboratory settings; however, fewer studies have focused on triggers to eating occasions in everyday situations or with nonclinical samples. Previous studies have also focused on a single or very few factors that might initiate eating occasions as opposed to a more comprehensive approach. Several approaches have been used to explain need states or goals that an individual may have at specific eating occasions. One approach involved segmentation of eating occasions by situation-based needs or need states among midlife women [20]. Primary need states included indulgence, comfort, nutrition and health, nurturing others, and convenience, with variable intakes of energy, fat, sweets, fruits, and vegetables by eating occasion need segment [20]. Situational context was the basis for the differences in intake of sweetened beverages, cereals, and entr´ees among young adults who recorded eating occasion characteristics with personal digital assistants [13]. Qualitative 24-hour recalls were used in another study to characterize demands within eating occasions [21]. Results showed that adults followed purposeful routines based on daily schedules to meet demands and values. Means-end theory and practice [22] is another approach used to explain needs related to food choice. This theory would indicate that food choice is based on the perceived benefits that the food can provide to the individual. These benefits may be organized at several levels from lower order functional needs to higher level psychosocial needs. Studies of eating occasion characteristics have examined not only stimuli that initiate consumption or goals that impact food choice within the occasion but also motives for ending the occasion [23]. These motives have typically included internal cues,
316
such as satiation, or external cues, such as when others were finished eating, with a relationship observed between responsiveness to type of cue and weight status. Eating occasion characteristics and goals are likely to be dependent on the characteristics of the individual and his or her environment. For example, young adults may be less likely to have routine eating occasions based on instability in living arrangements, household composition, and employment compared to older adults. A large number of eating occasions of young adults were characterized as occurring without structure and in less traditional settings [13], which could contribute to less healthful food choices. In another study, eating frequency among adults varied by age group, education, income, and ethnicity but not by gender [4]. However, the type of foods selected for snacks varied by gender in a sample of Swiss adults, with women choosing more healthful options than men [24]. Using longitudinal data over 7 years from 3 urban centers, Laroche et al. [25] found that becoming a parent did not have major effects on food or nutrient intake; however, intake of saturated fat decreased compared to nonparents. Parents also experienced a nonsignificant increase in energy intake of 79 calories/day compared to nonparents (p = 0.058) and mothers gained more weight than women who did not become a mother over the same period [26]. Eating occasion characteristics and needs may also differ based on a tendency for emotional eating, level of responsivity to environmental cues, and extent of disinhibition, which may vary with weight status [27]. A better understanding of demographic characteristics and BMI on eating occasion situational factors and goals will allow for effective tailoring of interventions to improve eating patterns and the type and amounts of foods consumed at specific eating occasions. To date, many studies have examined the influence of triggers, goals, and planning or the influence of demographic characteristics and BMI on eating patterns and intake, but these studies have typically not also focused on differences by eating occasion type [4,15–18,25–27]. Therefore, the purpose of this study was to describe how frequency and characteristics of traditional meal and non-meal occasions, including triggers and goals (based on rational, physiological, and emotional dimensions), vary by demographic and physical characteristics including age, gender, presence of children, and BMI.
MATERIALS AND METHODS Overview A 2-phase study was conducted by Blevins-O’Meara Insights (BOI), an independent marketing research firm, commissioned by the International Food Information Council (IFIC) Foundation. Preliminary focus group interviews in 3 regions of the United States identified potential patterns and differences with respect to situational context and characteristics and
VOL. 33, NO. 4
Eating Occasions by Demographic Characteristics
Downloaded by [University of Nebraska, Lincoln] at 10:20 08 October 2014
emotions and motivations for specific eating and drinking occasions. These focus group findings were used as the basis for the development of an online questionnaire administered to a nationally representative sample of adults 18–80 years of age to describe perceptions and approaches related to daily eating and drinking occasions including frequency, underlying motivations, and situational context. The study was intended to provide a broad description of the goals and motivations that drive food consumption in the United States. This article will focus only on selected measures to describe the relationship between age, gender, presence of children in the household, and BMI and eating occasion characteristics and needs/goals.
Subjects In the first formative phase, a series of 12 2-hour focus group interviews were conducted with a demographically reflective sample of the current U.S. population, including men and women aged 18–69 in 3 cities (Minneapolis, Minnesota [n = 15], Atlanta, Georgia [n = 15], and Phoenix, Arizona [n = 12]). The qualitative research was designed to generate a broad set of ideas, behaviors, and language to include in the quantitative survey. A diverse set of consumers was screened to include people from various demographic groups and provide representation across major U.S. Census regions. Several respondents from each of the following demographic groups were included: younger (18– 44 years) and older (45–69 years); male and female; Hispanic, African American, and white; married and single; having either no children, young children (0–3 years), or older (4 + years) children at home; and being employed or not employed. Because geographic dispersion was also considered valuable in addressing the broad objective of the study, 42 respondents were interviewed across these demographic variables in Atlanta, Minneapolis, and Phoenix. For the second phase, an online survey was conducted using a national online panel of consumers who had been recruited and maintained by Research Now, a private research company under the parent company e-Rewards, headquartered in Plano, Texas. Research Now uses a variety of recruiting practices to build and maintain their panel, including the recruiting of credit card holders, airline flyers, and magazine subscribers. This has resulted in a panel that includes about 6.5 million respondents. From the Research Now panel, an initial sample of individuals was selected so that it included a nationally representative geodemographically balanced cross section of consumers on the basis of age range, household income, gender, region, and race/ethnicity [28]. The selected panel members were sent email invitations to participate in a survey and were provided a Research Now web link. Upon entry to the website, respondents were asked to verify that they were in the 18- to 80-year-old age range and that they had consumed any food or beverage either at home or away from home the previous day. Those who qualified to participate in the survey by meeting these require-
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION
ments then continued with the remainder of the questionnaire. Respondents were informed of the remainder as follows: “You will be asked to recall how many different eating or drinking occasions you had yesterday and what you ate and/or drank on each occasion.” For their participation, respondents were paid $5.50 by Research Now in electronic currency redeemable for various consumer goods and services. Response rates varied by group, with lower responses from groups including those with low income, those who were ethnically diverse, young males, and those with lower education levels. Discrepancies were mitigated by first going through supplemental waves of invitations to those groups that were underrepresented and, second, by weighting the data. Weighting was performed using the anesrake R package (ANES Raking Implementation, version 0.70, 2011, J. Pasek, University of Michigan). Deidentified data, with no link to identifiers, was provided by BOI to the authors of this article for use in the analysis. Therefore, this project did not meet the definition of human subjects research by the University of Minnesota Institutional Review Board and did not require University of Minnesota Institutional Review Board review.
Formative Research and Survey Development Results from the preliminary focus group interviews described what constitutes an eating or drinking occasion and how the type of eating occasion was related to the emotions, social context, and mindfulness throughout a day. The variety of personal participant characteristics allowed for a differentiation in eating occasion characteristics based on the presence of children, employment, age, and tendency to plan. Various triggers to consumption occasions were identified as well as goals that varied with personal characteristics and cues to end the occasion. These triggers and goals were similar to those identified in previous focus group and individual interviews with adults [29,30]. Women identified similar internal and external drivers perceived as needs surrounding eating occasions reflecting a wide range in emotional gratification [29]. From interviews with adults, factors influencing food choice decisions were identified from 5 main categories of cultural ideals, personal factors, resources, social factors, and present contexts [30]. The triggers and goals identified in the current study were based on similar factors. Focus group findings were also used to develop definitions for eating occasions used in the survey, particularly for nontraditional occasions such as snacks, treats, and mini-meals. The consistency of terminology used by focus groups participants for these occasions contributed to the formulation of the terms and ensured that they were distinct from one another. All measures included in the survey were developed by BOI and subject to review by IFIC Foundation and IFIC Foundation’s stakeholders with the objective of providing a broad description of the goals and motivations that drive food consumption in the United States. The measures that provided data for this article
317
Downloaded by [University of Nebraska, Lincoln] at 10:20 08 October 2014
Eating Occasions by Demographic Characteristics included questions dealing with planning, triggers for deciding to engage in the consumption occasion, goals, and cues for ending the occasion. The questionnaire was divided into 3 parts. The first part involved a series of respondent-level questions that asked about food and beverage categories consumed on the previous day and the number of eating and drinking occasions that the respondent had the previous day by time of day. They were given specific instructions to report for a 24-hour period from midnight to midnight. Respondents were instructed that each time they began to eat or drink something on the previous day it should be thought of as a separate occasion. They were asked to indicate the number of separate occasions (1, 2, 3 or more) for each of 7 time periods where they had consumed anything. In the second part of the questionnaire, respondents were asked to report on no more than 3 of the occasions they listed in part 1. If more than 3 eating and/or drinking occasions were listed in part 1, 3 separate eating and/or drinking occasions were randomly selected for the respondent to report on. Then for each occasion, respondents were asked a series of occasion-level questions about what was consumed, what kind of occasion it was (for this they were asked which of the following best describes the occasion—breakfast: standard morning occasion, lunch: standard mid-day occasion, dinner: standard evening occasion, snack: an occasion between standard meals, treat: an infrequent occasion that includes an especially good tasting food/beverage, mini-meal: small meal that is typically consumed at multiple occasions through the day, or other) and the situational variables surrounding the occasion such as when the occasion occurred, who was there, and what the respondent was doing. Respondents were asked about their planning of the occasion, triggers (or cues) experienced, and goals they had for the occasion. Lastly, the respondents were asked another series of respondent-level questions that included questions about demographic characteristics and height and weight to enable calculation of BMI.
Pretesting Several IFIC Foundation staff members and trustees with nutrition and evaluation expertise and other IFIC Foundation stakeholders reviewed the survey in its entirety multiple times before administration. A pretest was conducted in which approximately 100 respondents completed the survey to determine whether reasonable patterns of response were forming. Response patterns were examined for consistency with the expectations formed by the qualitative formative process and phrasing for some questions was modified according to findings. In addition, Research Now examined the pretest data to identify points in the survey where respondents were having problems. Then those areas were reviewed for clarity or any dimensions that potentially caused problems for the respondent and revised to address these issues prior to final administration.
318
Survey Administration Overall the survey averaged 23 minutes to complete. The fieldwork was conducted between March 10 and April 14, 2013. From an estimate of initial invitations (n = 74,660), 8444 respondents (11%) clicked to begin screening, with 172 of these screened out by the screening criteria. Based on quotas, 3544 were over quota (they qualified on the original set of criteria but were in quota groups that had already been filled when they began screening). A number of potential respondents (n = 1871) abandoned the survey after qualifying, resulting in 2857 respondents completing the survey. Of the 2857 respondents completing the survey, data for some were not available for gender (n = 1), presence of children (n = 54), and height and/or weight (n = 89). Respondents ≥65 years of age with children in the household (n = 11) were not included. Data from these 155 respondents were not included in the analysis, resulting in a final sample of 2702 respondents.
Occasions The aim of this study was to provide an estimate of the true proportion of eating occasions. Therefore, the distribution of the occasions was reapportioned such that the sample of occasions was weighted to proportionally give each day of the week equal contribution (14% of total occasions). No other reapportioning of the occasion data was done. In support of this treatment, an analysis was done where a comparison was made of weighted versus unweighted occasion data for the questions regarding triggers and goals. Results between the two never differed by more than 1% point, so the effect of the weighting by day of the week is marginal. The initial 2857 respondents reported on 8141 eating/drinking occasions. After removing data from 155 respondents with missing data or who had children but were ≥65 years of age, 7713 occasions remained. Data from occasions where respondents classified occasion type as “other” were also removed (n = 457 occasions) leaving 7256 occasions for further analysis (Table 1). To avoid pseudoreplication from having multiple data points from the same individual, for individuals who reported on multiple eating occasions of a given type, a single occasion of each type was randomly chosen resulting in the elimination of an additional 567 occasions for a final sample of 6689 eating/drinking occasions.
Statistical Analyses Statistical analysis was completed using R, version 3.0.1 [31]. Descriptive statistics were used to describe characteristics of the sample, eating occasion frequency, characteristics, and goals. Association between eating occasions and BMI was tested for using quantile regression of the median due to notable skew in BMI. Eating occasions were totaled for each individual
VOL. 33, NO. 4
Eating Occasions by Demographic Characteristics Table 1. Eating Occasions (Unsampled)1 by Demographic Characteristics Variable All (n = 7256) Gender Children
Downloaded by [University of Nebraska, Lincoln] at 10:20 08 October 2014
Age
BMI
Currently trying to lose weight
Category Male Female Yes No 18–24 25–34 35–44 45–54 55–64 65–74 75–80
