Nutrition 31 (2015) 324–330
Contents lists available at ScienceDirect
Nutrition journal homepage: www.nutritionjrnl.com
Applied nutritional investigation
Stress and dietary behaviour among first-year university students in Australia: Sex differences Keren Papier B.P.H. (Hons) a, b, Faruk Ahmed Ph.D. c, *, Patricia Lee Ph.D. c, Juliet Wiseman Ph.D. d a
National Centre for Epidemiology and Population Health, Australian National University, Canberra, Australia Population Health Department, Queensland Institute of Medical Research, Brisbane, Queensland, Australia Public Health, School of Medicine and Griffith Health Institute, Griffith University, Gold Coast, Queensland, Australia d Nutrition and Dietetics, School of Allied Health Sciences and Griffith Health Institute, Griffith University, Gold Coast, Queensland, Australia b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 26 April 2014 Accepted 15 August 2014
Objective: The aim of this study was to examine the relationship between stress and food selection patterns by sex among first-year undergraduate students studying in an Australian university. Methods: Participating in this cross-sectional study were 728 (331 men and 397 female students) first-year students, ages >18 y, attending the Gold Coast Campus of Griffith University. Data were collected using a self-administered questionnaire consisting of three sections: sociodemographic information, stress measures, and a 7-d food frequency questionnaire. Results: More than half (52.9%) of the participants were found to suffer from some level of stress, with relatively more female students (57.4%) suffering than men (47.4%). Men who experienced mild to moderate levels of stress were two to three times more likely to eat cereal foods (P < 0.01), fish/seafood (P < 0.001), and protein powder (P < 0.05). They also tended to eat more meat alternatives (P < 0.05), highly processed foods (P < 0.05), and to drink more alcohol (P < 0.05) than unstressed male students. However, they were less likely to consume vegetables and fruit (P < 0.05) compared with their unstressed counterparts. The trend analysis results indicated significant dose–response patterns in the relationship between stress level and consumption of cereal food, meat alternatives, vegetables and fruit (negative trend), highly processed food, protein powder, beverages and alcoholic beverages (all P < 0.05). Female students who experienced mild to moderate stress were 2.22 times more likely to eat processed food (P < 0.01) than unstressed female students. Female students who experienced severe stress were less likely to consume meat alternatives (P < 0.05) than their unstressed counterparts. Significant dose–response trends were found in the relationship between stress levels and the consumption of meat alternatives, vegetables and fruit (both negative trends), and processed food (all P < 0.01). Conclusion: These results demonstrate a clear difference in food selection patterns between stressed male and female students, with stress being a more significant predictor of unhealthy food selection among male students. Further research is needed using a qualitative approach to understand how stress and eating behavior are related in university students. Ó 2015 Elsevier Inc. All rights reserved.
Keywords: Stress Dietary pattern Students Australia
Introduction This study received funding from the Griffith University, Australia. KP was responsible for data collection, data entry and cleaning, and wrote the first draft manuscript. FA conceived the idea for this study, and contributed to the study design, writing, and critical revision of the manuscript. PL contributed to the study design, data analysis and interpretation of results, and critical revision of the manuscript. JW contributed to the study design, writing, and critical revision of the manuscript. The authors have no conflict of interest to declare. * Corresponding author. Tel.: þ61 7-55527874; fax: þ61 7-55528799. E-mail address: F.ahmed@griffith.edu.au (F. Ahmed). http://dx.doi.org/10.1016/j.nut.2014.08.004 0899-9007/Ó 2015 Elsevier Inc. All rights reserved.
For many students, the transition from high school environment to university is accompanied by the emotional and/or psychological distress [1]. The transfer into a new physical and social environment is often accompanied by new relationships, financial demands, and expectations may bring with it increased levels of psychological distress [2]. Recent Australian literature reported that the occurrence and complexity of psychological
K. Papier et al. / Nutrition 31 (2015) 324–330
distress among university students is increasing [2]. Of even more concern, the prevalence of psychological distress among Australian university students surpasses that of the same age non-student population and that of the general Australian population [2,3]. There is substantial evidence that stress can affect an individual’s health, not only through direct physiological processes but also by changing behaviors that affect health [4,5]. One such health behavior is diet [5]. Stress has been shown to affect the amount of food consumed. Some studies have shown that individuals tended to increase consumption of high-caloric and high-fat snack foods when feeling stressed [4,5], whereas other studies reported that individuals ate less of all foods when under stress [6]. Stress also has been associated with the selection of foods consumed [5]. Studies have shown an increase in the preference for carbohydrate-rich foods during times of distress [7,8]. This increase has been partially attributed to the relationship between carbohydrate intake and serotonin brain activity [9]. Serotonin is a neurotransmitter in the central nervous system that has the capacity to alter moods and behaviors including anxiety, depression, and anger [10]. Furthermore, the intake of snack-type foods, ready-to-eat foods, and sweet foods such as chocolate, cakes, and ice cream, was found to increase among students experiencing stress [5,9–11], whereas the intake of healthy foods such as vegetables tended to decrease [4,5]. Studies have shown that food selection patterns among stressed individuals can vary by country of residence [8], dieting status [11], sex [5,12], or a combination of these factors. The findings surrounding the relationship between food selection, stress, and sex remain inconclusive. For instance, one study that compared food selection between stressed and unstressed men found that stress had no effect on food selection [12], whereas a similar study reported that stress was associated with an increase in the consumption of red meat, pizza, and soda among male participants [13]. Similarly, some studies have reported that although female students tended to be more health conscious than men during non-stress periods (i.e., opting for low-calorie foods) [14,15], when stressed they were likely to increase consumption of high-sugar, high-fat foods and unhealthy snacking, while reducing intakes of healthier foods like vegetables [15]. The university transition period is crucial for the establishment of future health behaviors [4,16,17]. Therefore, a negative association between stress and healthy dietary behavior would be particularly concerning for young university students. Overall, the studies that have assessed the relationship between psychological stress and dietary behaviors have either done so using the general population [13], by studying one sex [7, 12] or without taking into account total dietary behavior [5,16]. To our knowledge, no research in Australia has assessed the relationship between stress and food selection patterns of university students by sex while accounting for various sociodemographic factors [15,18–21]. Thus, the aim of this study was to describe the level of stress among first-year students in an Australian university, and to assess the relationship between stress and food selection patterns by sex. This study may provide information to help health professionals develop appropriate and holistic interventions for helping young adults cope with stress and simultaneously maintain healthy eating habits.
325
Griffith University has five campuses. The selection of this campus was based on its diverse academic cohort, its large number of international students, and its size as the biggest of the five campuses. In all, 728 first-year students (331 men and 397 female students) from four different schools participated in the study. The study was approved by the Griffith University Human Research Ethics Committee. Sampling To be included in this study, students had to be enrolled in a first-year course at Griffith University. All students were approached in their lecture halls during weeks 10 to 13 of the second semester of 2012 and during weeks 10 to 13 of the first semester in 2013. These two periods were selected because they are the two most stressful of the academic year. All of the students were informed about the purpose of the research at the beginning of their lecture and a self-administrated questionnaire was distributed at the end of the lecture. Over the 8-wk data collection period, 800 first-year students were approached. Of these students, 728 agreed to participate, making the overall response rate of participants 91%. Of the 375 men and 425 female students approached, 331 (response rate 88.3%) and 397 (response rate 93.4%) participated in the study. Data collection This needs a reference or explanation. The questionnaire was comprised of three sections: sociodemographic, stress measures, and dietary pattern for selected food items. Sociodemographic The sociodemographic section gathered information on 1) area of study and study status; 2) age, sex, marital status, living situation; 3) hours worked per week, and 4) anthropometric and health-related data (e.g., body weight and height, exercise, weight loss, and smoking). Stress assessment Stress among the participants was assessed using version 21 of the Depression Anxiety Stress Scale (DASS). The DASS has been validated successfully for different populations and is a popular tool for assessing the severity of the main symptoms for depression, anxiety, and stress among researchers within and outside of Australia [22]. This study only used the stress scale section of the DASS. This section consists of 14 items that assess difficulty relaxing; nervous arousal; and being easily upset/agitated, irritable/overreactive, and impatient [23]. The DASS scale uses a 4-point Likert scale of frequency or severity to rate the degree of stress experienced by participants during the previous week [22]. The 4-point scale ranges from 0, meaning that the participant felt that the item did not apply at all, to 3, whereby the participant felt that the item applied very much, or most of the time [23]. Individuals’ stress scores were calculated through the sum of all of scores from the 14 items and through the comparison of those scores to the cutoff scores for normal, moderate, and severe stress levels found in the DASS manual. Dietary intake The dietary patterns of the study participants were assessed using the Commonwealth Scientific and Industrial Research Organization food frequency questionnaire (FFQ). The CSIRO FFQ was used with permission of the CSIRO Preventative Health Flagship. The FFQ selected for this study was chosen based on its repeated validation success for the Australian adult population [24,25]. Information on the frequency of food intake was assessed based on the students’ previous 1-wk period using an 8-ranged scale. Food frequency intake responses ranged from never to 3þ times per day. The frequency of consumption for each food or beverage in the past week was converted into a daily equivalence as follows: not in the past week (0.00/d), once a week (0.14/d), two to three times a week (0.35/d), four to six times a week (0.70/d), once daily (1/d), twice daily (2/d), and 3 times daily (3/d). The FFQ was focused on the frequency of selected food items only and information on the portion size was not included. The food categories listed in the FFQ included cereal foods (breakfast cereal, white and wholemeal bread, rice, and pasta); fish and seafood (fresh and canned seafood); meat and chicken (beef, pork, and lamb); offal (liver and kidney); dairy and eggs (yogurt, whole fat and skimmed fat milk, and cheese varieties); meat alternatives (nuts, legumes, and tofu); vegetables (starchy and leafy) and fruit; processed food (fast food, meat pies, sausages, and kebab); highly processed food (jams, bakery sweets, crisps, ice cream, and chocolate or muesli bars); protein powder; nonalcoholic beverages (energy drinks, sodas, juices, and flavored milk); warm beverages (tea and coffee); and alcoholic beverages (mixers, beers, and spirits).
Participants and methods Statistical analysis Study participants A cross-sectional study design was used to collect data from undergraduate students studying at the Gold Coast campus of Griffith University, Australia.
Univariate analysis comprised of simple frequency distribution of selected variables. The proportion of male and female students who exceeded the specific cutoff scores for various stress levels was calculated. As the distributions of all
326
K. Papier et al. / Nutrition 31 (2015) 324–330
food categories were positively skewed (majority of the participants had small amount of consumption), a Mann-Whitney U-test was performed to compare the difference in the intake of each food category between male and female participants. Finally, associations between stress and dietary behaviors were assessed using logistic regression analysis for men and female students separately. Each of the food categories were treated as outcome (dependent) variables with stress treated as the independent variable. All food categories were divided into two groups using median intake cutoffs under each sex group. The independent variable (stress) was categorized into three levels: no stress (treated as the reference category), mild to moderate stress, and severe stress (severe and extremely severe were combined). Studies have found that stress is associated with various sociodemographic factors [26,27]. Therefore, the data were adjusted for potential confounders including marital status, study status, living situation, working hours, frequency of exercise, body mass index (BMI), whether participants were trying to lose weight, and smoking status. Age was eliminated in the logistic models due to a relatively large proportion (w13%) of missing values for both male and female participants. Results of the regression analyses are presented as odds ratios adjusted for confounding variables. A P-value for trend was produced in conjunction with the logistic regression analysis to examine significant dose–response trends between the dependent variables (food categories) and the independent variable (stress). A P 0.05 was considered statistically significant. Data was analyzed by using Statistical Package for the Social Sciences (SPSS) version 21.0-computer software (manufactured by the IBM Corporation in Armonk, NY, USA).
Results Of the participants, 45.5% (331) were men and 54.5% (397) were female students (Table 1). Mean (SD) age of the participants was 21.5 (2.8) y for men and 21.2 (3) y for female students. About 73% of all participants were single, only 1.1% was separated/ divorced and the rest were married or living with partner. However, a significantly higher proportion of men were single compared with the female students (81.2% versus 67.6%, P < 0.001). More than half (54.5%) of the participants were living on their own, with 8.9% living on campus and 45.6% living off campus, and the rest were living with family. A large majority (83.7%) of the participants were domestic students with almost equal numbers among men and female students. Forty-eight percent of participants were working at a paid part-time job for >10 h/wk, whereas others were full-time students or working 3 d/wk, another 31.7% reported doing physical exercise once or twice a week and the rest (18.8%) reported either never or rarely doing the exercise. Men tended to exercise more frequently than any female students (P < 0.001). The prevalence rates of overweight (BMI 25–29.9 kg/m2) and obesity (BMI 30 kg/m2) were 41.3% and 21.4% among men and female students, respectively; whereas only 2.8% of men and 7.7% of female students were considered underweight (P < 0.001). Less than one-fourth of men were trying to lose weight, whereas almost half (48.9%) of the female students wanted to lose weight (P < 0.001). Among all student participants, about 7% were smokers (Table 1). About 53% of all participants were found to suffer from some level of stress, with relatively more female students (57.4%) than men (47.4%) (P ¼ 0.031) suffering. Twelve percent of all female students were suffering from severe stress levels, 30% had moderate stress, and another 15.6% had mild stress. Among the men, only 6.6% were suffering from severe stress levels, 25.4% had moderate stress, and 15.4% suffered from mild stress (Table 1). Table 2 summarizes the differences in consumption of different food categories between male and female participants. Men had significantly higher intakes of cereal foods, meat and chicken, offal (liver/kidney), fish/seafood, dairy and eggs, processed food, protein powder, and alcoholic beverages than
Table 1 Sociodemographic characteristics of study participants by sex N ¼ 728
Age group (y) (mean/SD) 18–20 21þ Marital status Single Separated/divorced Married/partnership Study status Domestic International Living situation On-campus accommodation Off-campus accommodation At home with family Working hours/wk (paid employment) 0–10 11 Exercise (times/wk) Never or rarely 1–2 3 BMI (kg/m2) 0–18.49 18.5–24.99 25–29.99 30 Trying to lose weight Yes No Smoking status Smoker Nonsmoker Stress level Normal Mild Moderate Severe Very severe *
Total
Male
n (%)
n
Female %
n
P-value* %
21.48 (2.84)
21.23 (2.96)
133 157
45.9 54.1
173 174
49.9 50.1
533 (73.2) 268 8 (1.1) 2 181 (25.1) 60
81.2 0.6 18.2
265 6 121
67.6 1.5 30.9
604 (83.7) 277 118 (16.3) 51
84.5 15.5
327 67
83.0 17.0
24
7.3
40
10.1
330 (45.6) 145
44.2
185
46.9
329 (45.5) 159
48.5
170
43.0
306 (48) 331 (52)
0.315
Contents lists available at ScienceDirect
Nutrition journal homepage: www.nutritionjrnl.com
Applied nutritional investigation
Stress and dietary behaviour among first-year university students in Australia: Sex differences Keren Papier B.P.H. (Hons) a, b, Faruk Ahmed Ph.D. c, *, Patricia Lee Ph.D. c, Juliet Wiseman Ph.D. d a
National Centre for Epidemiology and Population Health, Australian National University, Canberra, Australia Population Health Department, Queensland Institute of Medical Research, Brisbane, Queensland, Australia Public Health, School of Medicine and Griffith Health Institute, Griffith University, Gold Coast, Queensland, Australia d Nutrition and Dietetics, School of Allied Health Sciences and Griffith Health Institute, Griffith University, Gold Coast, Queensland, Australia b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 26 April 2014 Accepted 15 August 2014
Objective: The aim of this study was to examine the relationship between stress and food selection patterns by sex among first-year undergraduate students studying in an Australian university. Methods: Participating in this cross-sectional study were 728 (331 men and 397 female students) first-year students, ages >18 y, attending the Gold Coast Campus of Griffith University. Data were collected using a self-administered questionnaire consisting of three sections: sociodemographic information, stress measures, and a 7-d food frequency questionnaire. Results: More than half (52.9%) of the participants were found to suffer from some level of stress, with relatively more female students (57.4%) suffering than men (47.4%). Men who experienced mild to moderate levels of stress were two to three times more likely to eat cereal foods (P < 0.01), fish/seafood (P < 0.001), and protein powder (P < 0.05). They also tended to eat more meat alternatives (P < 0.05), highly processed foods (P < 0.05), and to drink more alcohol (P < 0.05) than unstressed male students. However, they were less likely to consume vegetables and fruit (P < 0.05) compared with their unstressed counterparts. The trend analysis results indicated significant dose–response patterns in the relationship between stress level and consumption of cereal food, meat alternatives, vegetables and fruit (negative trend), highly processed food, protein powder, beverages and alcoholic beverages (all P < 0.05). Female students who experienced mild to moderate stress were 2.22 times more likely to eat processed food (P < 0.01) than unstressed female students. Female students who experienced severe stress were less likely to consume meat alternatives (P < 0.05) than their unstressed counterparts. Significant dose–response trends were found in the relationship between stress levels and the consumption of meat alternatives, vegetables and fruit (both negative trends), and processed food (all P < 0.01). Conclusion: These results demonstrate a clear difference in food selection patterns between stressed male and female students, with stress being a more significant predictor of unhealthy food selection among male students. Further research is needed using a qualitative approach to understand how stress and eating behavior are related in university students. Ó 2015 Elsevier Inc. All rights reserved.
Keywords: Stress Dietary pattern Students Australia
Introduction This study received funding from the Griffith University, Australia. KP was responsible for data collection, data entry and cleaning, and wrote the first draft manuscript. FA conceived the idea for this study, and contributed to the study design, writing, and critical revision of the manuscript. PL contributed to the study design, data analysis and interpretation of results, and critical revision of the manuscript. JW contributed to the study design, writing, and critical revision of the manuscript. The authors have no conflict of interest to declare. * Corresponding author. Tel.: þ61 7-55527874; fax: þ61 7-55528799. E-mail address: F.ahmed@griffith.edu.au (F. Ahmed). http://dx.doi.org/10.1016/j.nut.2014.08.004 0899-9007/Ó 2015 Elsevier Inc. All rights reserved.
For many students, the transition from high school environment to university is accompanied by the emotional and/or psychological distress [1]. The transfer into a new physical and social environment is often accompanied by new relationships, financial demands, and expectations may bring with it increased levels of psychological distress [2]. Recent Australian literature reported that the occurrence and complexity of psychological
K. Papier et al. / Nutrition 31 (2015) 324–330
distress among university students is increasing [2]. Of even more concern, the prevalence of psychological distress among Australian university students surpasses that of the same age non-student population and that of the general Australian population [2,3]. There is substantial evidence that stress can affect an individual’s health, not only through direct physiological processes but also by changing behaviors that affect health [4,5]. One such health behavior is diet [5]. Stress has been shown to affect the amount of food consumed. Some studies have shown that individuals tended to increase consumption of high-caloric and high-fat snack foods when feeling stressed [4,5], whereas other studies reported that individuals ate less of all foods when under stress [6]. Stress also has been associated with the selection of foods consumed [5]. Studies have shown an increase in the preference for carbohydrate-rich foods during times of distress [7,8]. This increase has been partially attributed to the relationship between carbohydrate intake and serotonin brain activity [9]. Serotonin is a neurotransmitter in the central nervous system that has the capacity to alter moods and behaviors including anxiety, depression, and anger [10]. Furthermore, the intake of snack-type foods, ready-to-eat foods, and sweet foods such as chocolate, cakes, and ice cream, was found to increase among students experiencing stress [5,9–11], whereas the intake of healthy foods such as vegetables tended to decrease [4,5]. Studies have shown that food selection patterns among stressed individuals can vary by country of residence [8], dieting status [11], sex [5,12], or a combination of these factors. The findings surrounding the relationship between food selection, stress, and sex remain inconclusive. For instance, one study that compared food selection between stressed and unstressed men found that stress had no effect on food selection [12], whereas a similar study reported that stress was associated with an increase in the consumption of red meat, pizza, and soda among male participants [13]. Similarly, some studies have reported that although female students tended to be more health conscious than men during non-stress periods (i.e., opting for low-calorie foods) [14,15], when stressed they were likely to increase consumption of high-sugar, high-fat foods and unhealthy snacking, while reducing intakes of healthier foods like vegetables [15]. The university transition period is crucial for the establishment of future health behaviors [4,16,17]. Therefore, a negative association between stress and healthy dietary behavior would be particularly concerning for young university students. Overall, the studies that have assessed the relationship between psychological stress and dietary behaviors have either done so using the general population [13], by studying one sex [7, 12] or without taking into account total dietary behavior [5,16]. To our knowledge, no research in Australia has assessed the relationship between stress and food selection patterns of university students by sex while accounting for various sociodemographic factors [15,18–21]. Thus, the aim of this study was to describe the level of stress among first-year students in an Australian university, and to assess the relationship between stress and food selection patterns by sex. This study may provide information to help health professionals develop appropriate and holistic interventions for helping young adults cope with stress and simultaneously maintain healthy eating habits.
325
Griffith University has five campuses. The selection of this campus was based on its diverse academic cohort, its large number of international students, and its size as the biggest of the five campuses. In all, 728 first-year students (331 men and 397 female students) from four different schools participated in the study. The study was approved by the Griffith University Human Research Ethics Committee. Sampling To be included in this study, students had to be enrolled in a first-year course at Griffith University. All students were approached in their lecture halls during weeks 10 to 13 of the second semester of 2012 and during weeks 10 to 13 of the first semester in 2013. These two periods were selected because they are the two most stressful of the academic year. All of the students were informed about the purpose of the research at the beginning of their lecture and a self-administrated questionnaire was distributed at the end of the lecture. Over the 8-wk data collection period, 800 first-year students were approached. Of these students, 728 agreed to participate, making the overall response rate of participants 91%. Of the 375 men and 425 female students approached, 331 (response rate 88.3%) and 397 (response rate 93.4%) participated in the study. Data collection This needs a reference or explanation. The questionnaire was comprised of three sections: sociodemographic, stress measures, and dietary pattern for selected food items. Sociodemographic The sociodemographic section gathered information on 1) area of study and study status; 2) age, sex, marital status, living situation; 3) hours worked per week, and 4) anthropometric and health-related data (e.g., body weight and height, exercise, weight loss, and smoking). Stress assessment Stress among the participants was assessed using version 21 of the Depression Anxiety Stress Scale (DASS). The DASS has been validated successfully for different populations and is a popular tool for assessing the severity of the main symptoms for depression, anxiety, and stress among researchers within and outside of Australia [22]. This study only used the stress scale section of the DASS. This section consists of 14 items that assess difficulty relaxing; nervous arousal; and being easily upset/agitated, irritable/overreactive, and impatient [23]. The DASS scale uses a 4-point Likert scale of frequency or severity to rate the degree of stress experienced by participants during the previous week [22]. The 4-point scale ranges from 0, meaning that the participant felt that the item did not apply at all, to 3, whereby the participant felt that the item applied very much, or most of the time [23]. Individuals’ stress scores were calculated through the sum of all of scores from the 14 items and through the comparison of those scores to the cutoff scores for normal, moderate, and severe stress levels found in the DASS manual. Dietary intake The dietary patterns of the study participants were assessed using the Commonwealth Scientific and Industrial Research Organization food frequency questionnaire (FFQ). The CSIRO FFQ was used with permission of the CSIRO Preventative Health Flagship. The FFQ selected for this study was chosen based on its repeated validation success for the Australian adult population [24,25]. Information on the frequency of food intake was assessed based on the students’ previous 1-wk period using an 8-ranged scale. Food frequency intake responses ranged from never to 3þ times per day. The frequency of consumption for each food or beverage in the past week was converted into a daily equivalence as follows: not in the past week (0.00/d), once a week (0.14/d), two to three times a week (0.35/d), four to six times a week (0.70/d), once daily (1/d), twice daily (2/d), and 3 times daily (3/d). The FFQ was focused on the frequency of selected food items only and information on the portion size was not included. The food categories listed in the FFQ included cereal foods (breakfast cereal, white and wholemeal bread, rice, and pasta); fish and seafood (fresh and canned seafood); meat and chicken (beef, pork, and lamb); offal (liver and kidney); dairy and eggs (yogurt, whole fat and skimmed fat milk, and cheese varieties); meat alternatives (nuts, legumes, and tofu); vegetables (starchy and leafy) and fruit; processed food (fast food, meat pies, sausages, and kebab); highly processed food (jams, bakery sweets, crisps, ice cream, and chocolate or muesli bars); protein powder; nonalcoholic beverages (energy drinks, sodas, juices, and flavored milk); warm beverages (tea and coffee); and alcoholic beverages (mixers, beers, and spirits).
Participants and methods Statistical analysis Study participants A cross-sectional study design was used to collect data from undergraduate students studying at the Gold Coast campus of Griffith University, Australia.
Univariate analysis comprised of simple frequency distribution of selected variables. The proportion of male and female students who exceeded the specific cutoff scores for various stress levels was calculated. As the distributions of all
326
K. Papier et al. / Nutrition 31 (2015) 324–330
food categories were positively skewed (majority of the participants had small amount of consumption), a Mann-Whitney U-test was performed to compare the difference in the intake of each food category between male and female participants. Finally, associations between stress and dietary behaviors were assessed using logistic regression analysis for men and female students separately. Each of the food categories were treated as outcome (dependent) variables with stress treated as the independent variable. All food categories were divided into two groups using median intake cutoffs under each sex group. The independent variable (stress) was categorized into three levels: no stress (treated as the reference category), mild to moderate stress, and severe stress (severe and extremely severe were combined). Studies have found that stress is associated with various sociodemographic factors [26,27]. Therefore, the data were adjusted for potential confounders including marital status, study status, living situation, working hours, frequency of exercise, body mass index (BMI), whether participants were trying to lose weight, and smoking status. Age was eliminated in the logistic models due to a relatively large proportion (w13%) of missing values for both male and female participants. Results of the regression analyses are presented as odds ratios adjusted for confounding variables. A P-value for trend was produced in conjunction with the logistic regression analysis to examine significant dose–response trends between the dependent variables (food categories) and the independent variable (stress). A P 0.05 was considered statistically significant. Data was analyzed by using Statistical Package for the Social Sciences (SPSS) version 21.0-computer software (manufactured by the IBM Corporation in Armonk, NY, USA).
Results Of the participants, 45.5% (331) were men and 54.5% (397) were female students (Table 1). Mean (SD) age of the participants was 21.5 (2.8) y for men and 21.2 (3) y for female students. About 73% of all participants were single, only 1.1% was separated/ divorced and the rest were married or living with partner. However, a significantly higher proportion of men were single compared with the female students (81.2% versus 67.6%, P < 0.001). More than half (54.5%) of the participants were living on their own, with 8.9% living on campus and 45.6% living off campus, and the rest were living with family. A large majority (83.7%) of the participants were domestic students with almost equal numbers among men and female students. Forty-eight percent of participants were working at a paid part-time job for >10 h/wk, whereas others were full-time students or working 3 d/wk, another 31.7% reported doing physical exercise once or twice a week and the rest (18.8%) reported either never or rarely doing the exercise. Men tended to exercise more frequently than any female students (P < 0.001). The prevalence rates of overweight (BMI 25–29.9 kg/m2) and obesity (BMI 30 kg/m2) were 41.3% and 21.4% among men and female students, respectively; whereas only 2.8% of men and 7.7% of female students were considered underweight (P < 0.001). Less than one-fourth of men were trying to lose weight, whereas almost half (48.9%) of the female students wanted to lose weight (P < 0.001). Among all student participants, about 7% were smokers (Table 1). About 53% of all participants were found to suffer from some level of stress, with relatively more female students (57.4%) than men (47.4%) (P ¼ 0.031) suffering. Twelve percent of all female students were suffering from severe stress levels, 30% had moderate stress, and another 15.6% had mild stress. Among the men, only 6.6% were suffering from severe stress levels, 25.4% had moderate stress, and 15.4% suffered from mild stress (Table 1). Table 2 summarizes the differences in consumption of different food categories between male and female participants. Men had significantly higher intakes of cereal foods, meat and chicken, offal (liver/kidney), fish/seafood, dairy and eggs, processed food, protein powder, and alcoholic beverages than
Table 1 Sociodemographic characteristics of study participants by sex N ¼ 728
Age group (y) (mean/SD) 18–20 21þ Marital status Single Separated/divorced Married/partnership Study status Domestic International Living situation On-campus accommodation Off-campus accommodation At home with family Working hours/wk (paid employment) 0–10 11 Exercise (times/wk) Never or rarely 1–2 3 BMI (kg/m2) 0–18.49 18.5–24.99 25–29.99 30 Trying to lose weight Yes No Smoking status Smoker Nonsmoker Stress level Normal Mild Moderate Severe Very severe *
Total
Male
n (%)
n
Female %
n
P-value* %
21.48 (2.84)
21.23 (2.96)
133 157
45.9 54.1
173 174
49.9 50.1
533 (73.2) 268 8 (1.1) 2 181 (25.1) 60
81.2 0.6 18.2
265 6 121
67.6 1.5 30.9
604 (83.7) 277 118 (16.3) 51
84.5 15.5
327 67
83.0 17.0
24
7.3
40
10.1
330 (45.6) 145
44.2
185
46.9
329 (45.5) 159
48.5
170
43.0
306 (48) 331 (52)
0.315
