Eating Behavior of Brazilian College Students: Influences of Lifestyle,Negative Affectivity,and Personal Characteristics

Introduction

The phrase “eating behavior” is used to describe food management and the thoughts and feelings that may influence food intake (i.e., how food is selected and attained) (Alvarenga, 2019). The importance of eating behavior to a population’s health status has been well recognized (Alvarenga & Koritar, 2019; Karlsson et al., 2000; Toral & Slater, 2007; WHO, 1998). Based on Karlsson et al. (2000), eating behavior can be assessed in three dimensions: (a) conscious restriction of food to control or lose weight (cognitive restraint or CR); (b) the ingestion of large amounts of food in a short period of time (uncontrolled eating or UE); and (c) food intake due to emotions such as sadness, depression, and loneliness (emotional eating or EE).

Various studies have found that a variety of biological, sociocultural, psychological and emotional factors may influence eating behavior (Alvarenga, 2019; Canetti et al., 2002; Carraça et al., 2014; Nahas et al., 2000; Toral & Slater, 2007). Among sociocultural aspects, lifestyle includes actions that reflect a person’s beliefs and values as they are applied to daily living (Carraça et al., 2014; Nahas et al., 2000; WHO, 1998). Lifestyle assessment, according to Nahas, Barros, and Francalacci (2000), involves physical components (such as nutrition, physical activity, and preventive actions) and psychological components (such as social relationships and stress control). Thus, a healthier lifestyle may lead to healthier eating behaviors (Carraça et al., 2014). Jallinoja et al. (2010) theorized that emotions may mediate the relationship between lifestyle and eating behavior. For example, a healthy lifestyle might be considered restrictive and prohibitive (Alvarenga & Koritar, 2019; Jallinoja et al., 2010) when eating pleasure is neglected (Alvarenga & Koritar, 2019; Jallinoja et al., 2010; Nahas et al., 2000), resulting in a feeling of emotional deprivation that might lead to episodes of compensatory uncontrolled or emotional eating (Alvarenga, Polacow, et al., 2019).

An individual’s diet not only supplies physiological and nutritional needs, but it also fulfills emotional needs (Canetti et al., 2002; Jallinoja et al., 2010; Karlsson et al., 2000). Negative affectivity is the propensity to experience negative emotional states (Watson & Clark, 1984), such as depression, anxiety, and stress (Lovibond & Lovibond, 1995). Some studies indicate that coping with emotional situations, especially negative ones, may be associated with variations in food consumption and the adoption of inappropriate eating behaviors (Canetti et al., 2002; Geliebter & Aversa, 2003; Macht, 2008; Saat et al., 2014).

An individual’s demographic and anthropometric characteristics, such as sex, weight status, and age, may also influence eating behavior (Abdella et al., 2019; Canetti et al., 2002; Jáuregui-Lobera et al., 2014; León-Vázquez et al., 2017; Medeiros et al., 2017; Sepulveda et al., 2008). Studies by Jáuregui-Lobera et al. (2014), and Medeiros et al. (2017) found that females had higher scores for emotional eating and cognitive restraint than males. León-Vázquez et al. (2017) reported that individuals in conditions of overweight or obesity had higher uncontrolled eating scores and a higher prevalence of high-risk eating behaviors than individuals without these conditions. Abdella et al. (2019) found that younger individuals presented a higher risk of emotional and uncontrolled eating than older people. Other personal characteristics that were found to influence eating behavior include the individual’s economic stratum, in that eating behavior reflecting body image or weight concerns may differ according to an individual’s high and low purchasing power (Silva et al., 2017).

The concepts of eating behavior, negative affectivity, and lifestyle have been investigated among university students (Aceijas et al., 2016; Camacho et al., 2016; Jáuregui-Lobera et al., 2014; Medeiros et al., 2017). When entering university, students experience significant life changes as they further explore personal identity, develop autonomy, and assume new responsibilities (Nelson et al., 2008; Sepulveda et al., 2008), including, for example, selecting and preparing their own meals. In addition, exposure to stressors may increase for university students, as academic, social, and personal demands increase (Nelson et al., 2008), possibly leading, in turn, to exhaustion and thoughts about dropping courses. According to Canetti, Bachar and Berry (2002), the management of emotions can involve changes toward unhealthy eating behavior such as turning to food to seek relief from negative affect. In addition, immature time-management skills can contribute to adopting unfavorable lifestyle habits (Nelson et al., 2008), such as eating fast-food and becoming more sedentary. Therefore, the university student population may be considered susceptible (Aceijas et al., 2016) to adopting dysfunctional eating behaviors and may benefit from identifying risk factors for unhealthy habits so as to minimize their negative health impacts.

In this study, we aimed to (i) verify the impact of lifestyle components, negative affectivity and personal characteristics (such as sex, age, weight status and economic stratum) on dimensions of eating behavior among college students; and (ii) verify a presumption that negative affectivity mediates the relationship between lifestyle and eating behavior among university students.

Method

Study Design, Sampling Method, and Participants

This was an observational cross-sectional study for which we used a non-probabilistic convenience sample. We invited research participation from undergraduate students who were enrolled in 2017 and 2018 at a public Brazilian university. Participant inclusion criteria included being aged between 18–35 years. Individuals with diabetes and pregnant females were excluded from participation. We calculated a minimum sample size, based on Kim’s (2005) proposal that considers the degrees of freedom of the models, the desired statistical power and the statistical significance level (α). The degrees of freedom of the models ranged from 789 to 1,103. We considered 80% power and a 5% significance level. We considered the higher estimate of 54 as a minimum sample size so as to encompass all models. We included an additional 20% to account for possible losses, giving us a need for a total sample of 68 students.

Regarding sample recruitment, we began this process inside the School of Pharmaceutical Sciences in São Paulo State University. Subsequently, we invited other São Paulo State University students. We contacted the responsible professor of each class and scheduled a day and time for data collection. After the instructor’s authorization, we invited the students to participate in our study during class hours. We presented the study’s objectives and instructions for completing the assessment tools, emphasizing that their participation was voluntary and anonymous. All participants gave written informed consent. The study protocol was approved by the Ethics Committee for Human Research of the School of Pharmaceutical Sciences of UNESP at the Araraquara campus (CAAE: 63553516.4.0000.5426).

Our final study sample consisted of 1,109 participants (M age = 20.9, SD = 2.7 years; 65.7% females). Participants’ average body mass index (BMI) was 23.3 (SD = 4.1) kg/m². Their further demographic information is presented in Table 1.

OPEN IN VIEWER

Table 1.

Sample Characteristics.

Characteristic	n (%)
Course area
Human sciences	717 (64.7)
Exact sciences	85 (7.6)
Biological and Health sciences	307 (27.7)
Course year
First	401 (36.3)
Second	277 (25.1)
Third	198 (18.0)
Fourth	179 (16.2)
Fifth	48 (4.4)
Class period
Morning	176 (16.8)
Afternoon	145 (13.8)
Night	358 (34.1)
Full-time	371 (35.3)
Did you have thoughts about dropping out of the course?
Frequently	119 (10.7)
Sometimes	491 (44.4)
Never	497 (44.9)
Economic Stratum (average monthly income†)
A($4,952.39 USD)	283 (26.4)
B($1,639.61 USD)	618 (57.7)
C($468.42 USD)	165 (15.4)
D and E($139.50 USD)	5 (0.5)
Anthropometric weight status (BMI range)
Underweight (<18.5 kg/m²)	95 (8.7)
Normal Range/Eutrophy ([18.5–25.0[ kg/m²)	692 (63.4)
Pre-obesity ([25.0–30.0[ kg/m²)	236 (21.6)
Obesity (≥30.0 kg/m²)	69 (6.3)

Note. †Brazilian Economic Classification Criteria 2019. Brazilian Reals (BRL) were converted into American dollars (USD) (exchange rate in March 2020 – 1 USD = 5.16 BRL – available in https://www.bcb.gov.br).

Most participants were enrolled in Humanities and Social Sciences courses, were in the first two years of their undergraduate course (61.4%), studied during daytime (65.9%), had thought of dropping out (often or sometimes, 55.1%), belonged to economic stratum B, and had normal weight (eutrophics).

Results

Table 2 presents an analysis of the pathways between global lifestyle components, negative affectivity, and demographic variables in relation to the dimensions of eating behavior, assessed separately. Furthermore, in Figures 1 to 3 we present the structural model considering the pathways of specific lifestyle components and their impact on negative affectivity and eating behavior.

OPEN IN VIEWER

Table 2.

Structural Model to Assess the Impact of Lifestyle Components (Physical And Psychological), Negative Affectivity, and Sex, Age, Weight Status and Economic Stratum on the Dimensions of Eating Behavior (Cognitive Restraint, Uncontrolled Eating, and Emotional Eating) of College Students (n = 1,109).

Independent → Dependent	B	β	Standard error	p
Cognitive Restraint (CR) #
FYS → NEG-A (β1)	0.232	0.153	0.090	0.010
PSY→ NEG-A (β2)	−0.810	−0.543	0.103	<0.001
DO → NEG-A (β3)	0.412	0.267	0.050	<0.001
FYS → CR	1.293	0.807	0.138	<0.001
PSY → CR	−0.503	−0.318	0.128	<0.001
NEG-A → CR (β4)	0.015	0.014	0.052	0.779
Sex → CR	0.262	0.154	0.056	<0.001
Age → CR	0.011	0.038	0.009	0.223
Weight status → CR	0.345	0.191	0.060	<0.001
EcStratum → CR	0.181	0.081	0.073	0.014
FYS → NEG-A → CR (β1*β4)‡	0.003	0.002	0.011	0.767
PSY →NEG-A → CR (β2*β4)‡	−0.012	−0.008	0.042	0.777
DO → NEG-A → CR (β3*β4)‡	0.006	0.004	0.022	0.780
Uncontrolled Eating (UE) ##
FYS → NEG-A (β1)	0.280	0.186	0.106	0.008
PSY→ NEG-A (β2)	−0.887	−0.574	0.121	<0.001
DO → NEG-A (β3)	0.416	0.271	0.050	<0.001
FYS → UE	−0.101	−0.123	0.053	0.054
PSY → UE	0.018	0.021	0.063	0.773
NEG-A → UE (β4)	0.128	0.233	0.028	<0.001
Sex → UE	0.073	0.083	0.030	0.013
Age → UE	−0.014	−0.090	0.005	0.006
Weight status → UE	0.171	0.183	0.034	<0.001
EcStratum → UE	0.012	0.010	0.038	0.749
FYS → NEG-A → UE (β1*β4)‡	0.036	0.043	0.016	0.027
PSY →NEG-A → UE (β2*β4)‡	−0.114	−0.134	0.031	<0.001
DO → NEG-A → UE (β3*β4)‡	0.053	0.063	0.013	<0.001
Emotional Eating (EE) ###
FYS → NEG-A (β1)	0.269	0.181	0.102	0.008
PSY→ NEG-A (β2)	−0.868	−0.570	0.118	<0.001
DO → NEG-A (β3)	0.415	0.269	0.050	<0.001
FYS → EE	0.067	0.047	0.083	0.424
PSY → EE†	−0.124	−0.086	0.101	0.220
NEG-A → EE (β4)	0.283	0.298	0.047	<0.001
Sex → EE	0.512	0.336	0.062	<0.001
Age → EE	−0.037	−0.142	0.009	<0.001
Weight status → EE	0.423	0.261	0.061	<0.001
EcStratum → EE	0.035	0.017	0.063	0.582
FYS → NEG-A → EE (β1*β4)‡	0.076	0.054	0.030	0.013
PSY →NEG-A → EE (β2*β4)‡	−0.246	−0.170	0.052	<0.001
DO → NEG-A → EE (β3*β4)‡	0.117	0.080	0.023	<0.001

B: unstandardized pathways; β: standardized pathways; FYS: physical component of lifestyle; PSY: psychological component of lifestyle; NEG-A: negative affectivity. Independent variables: DO: thoughts of dropping out of the undergraduate course (0 = no; 1 = yes); Sex (0 = male and 1 = female); weight status (0 = not overweight or obese and 1 = overweight or obese) EcStratum (Economic Stratum: 0= lower middle income; 1 = higher middle income); age: continuous variable. r² = explained variance; χ²/df: chi-square per degrees of freedom ratio; CFI: Comparative Fit Index; TLI: Tucker-Lewis index; RMSEA: Root Mean Square Error of Approximation.

^†Model without mediator variable (NEG-A): β = −0.274; p < 0.001.

^#Cognitive Restraint: r² = 0.55; χ²/df = 3.08; CFI = 0.95; TLI = 0.96; RMSEA = 0.04.

^##Uncontrolled eating: r² = 0.113; χ²/df = 2.88; CFI = 0.94; TLI = 0.95; RMSEA = 0.04.

^###Emotional eating: r² = 0.295; χ²/df = 3.25; CFI = 0.95; TLI = 0.96; RMSEA = 0.05.

^‡Bootstrap simulation: Sobel’s test.

OPEN IN VIEWER

Figure 1.

Hypothetical Structural Causal Model Designed to Verify the Impact of Lifestyle First-Order Components, Negative Affectivity and Sample Characteristics on the Cognitive Restraint. Note: Full arrows indicate significant pathways (p<.05); Dotted arrows indicate non-significant pathways (p>.05).

OPEN IN VIEWER

Figure 2.

Hypothetical Structural Causal Model Designed to Verify the Impact of Lifestyle First-Order Components, Negative Affectivity and Sample Characteristics on the Uncontrolled Eating. Note: Full arrows indicate significant pathways (p<.05); Dotted arrows indicate non-significant pathways (p>.05).

OPEN IN VIEWER

Figure 3.

Hypothetical Structural Causal Model Designed to Verify the Impact of Lifestyle First-Order Components, Negative Affectivity and Sample Characteristics on the Emotional Eating. Note: Full arrows indicate significant pathways (p<.05); Dotted arrows indicate non-significant pathways (p>.05).

We found a significant impact of lifestyle second-order components on negative affectivity in all dimensions of eating behavior. Negative affectivity, however, had a significant impact, and it mediated the relation between the students’ thoughts about dropping a course only in the uncontrolled eating and emotional eating models.

Physical lifestyle components only had a direct impact on cognitive restraint, and this impact was related to nutrition. While the psychological component of lifestyle also had a direct impact on cognitive restraint; it was not possible to determine what specific factor related to it. Thus, we suggest that the psychological component of lifestyle be considered as a second-order factor (i.e. with its two components of social relationship and stress management) for evaluating its impact on eating behavior.

Regarding mediation analyses, we verified that, for global models, there were no mediation pathways between lifestyle and eating behavior, except for the relationship between psychological lifestyle on emotional eating. When we analyzed the specific models, we observed that negative affectivity influenced the relation between Stress Management and both Uncontrolled Eating and Emotional Eating, making negative affectivity a mediator variable.

The sample characteristics of sex and weight status contributed significantly to all dimensions of eating behavior. Females and people with higher BMIs showed higher cognitive restraint, uncontrolled eating and emotional eating than males and people with lower BMIs. Age was significant only for uncontrolled eating and emotional eating, and economic stratum was significant only for cognitive restraint. Thus, younger students had higher uncontrolled eating and emotional eating; and those with higher income presented higher cognitive restraint.

Discussion

In the present study, we found that lifestyle factors directly or indirectly impacted different dimensions of eating behavior, and negative affectivity had a significant impact on uncontrolled eating and emotional eating. Moreover, negative affectivity mediated the relationship between thoughts of dropping out of a course and uncontrolled eating and emotional eating. Sex, age, weight status and economic stratum were also relevant for the assessment of eating behavior.

Our finding of a positive impact of the physical component of lifestyle on negative affectivity is inconsistent with prior research suggesting that negative affectivity is associated with less healthy lifestyles (Alves et al., 2019; Badger et al., 2019). These results may be explained by the these participants’ past motivations (Pelletier et al., 2004; Ryan et al., 1997) and beliefs (Alvarenga & Koritar, 2019; Coveney & Bunton, 2003; Jallinoja et al., 2010), since Pelletier et al. (2004) noted that extrinsic motivations can compromise the maintenance of behavior in the long term as the motives for the behavior then come from external sources that are not inherent to the individual. In addition, the feeling that a healthy lifestyle is restrictive and prohibitive (Alvarenga & Koritar, 2019) may reduce an individual's commitment to maintain healthy behaviors (Jallinoja et al., 2010) or promote negative affectivity (Polivy, 1996). One of the challenges in promoting healthier lifestyles is the awareness of a possible need to balance the pleasure obtained by eating with the practice of physical exercise (Jallinoja et al., 2010), challenging the paradigm that pleasure and health cannot concurrently exist (Alvarenga & Koritar, 2019; Coveney & Bunton, 2003; Jallinoja et al., 2010). Future studies should investigate the role of motivations and beliefs that lead to and control adopted lifestyles, so that more inclusive and humanized education and intervention strategies can be planned to promote sustainable and effective improvements in the overall health of the population.

The impact of emotions on eating behavior has been addressed in previous studies (Canetti et al., 2002; Macht, 2008; Saat et al., 2014). People can use food in an attempt to regulate emotional processes (Canetti et al., 2002), and when facing stressful situations, some people may lose control over eating. Negative affectivity can trigger episodes of uncontrolled eating and emotional eating (Canetti et al., 2002; Geliebter & Aversa, 2003; Macht, 2008; Saat et al., 2014). Our results reinforce the need to consider not only the biological and nutritional factors related to eating behavior but also the emotional/psychological aspects, in order to develop an integrated and comprehensive assessment of the impact of emotions on eating behavior.

We found a better physical lifestyle to be associated with higher cognitive food restraint, similar to the findings of McLean and Barr (2003). The goal to consciously control food intake to maintain or lose weight is, generally, in line with a lifestyle considered healthier. However, it should be noted that food restriction may be associated with extreme weight loss behaviors such as fasting, purging and the use the laxative substances (Alvarenga, Figueiredo, et al., 2019) that can cause disordered eating, physiological problems and compensatory behaviors leading to weight gain. Thus, the behavior of food restraint should be evaluated carefully, as it may trigger benefit or harm to an individual’s health depending on the magnitude of restriction. On the other hand, we found that the psychological component of lifestyle had a negative impact on cognitive restraint, but no influence on uncontrolled eating. In the lifestyle first-order components evaluation, these results may indicate that stress management can favor emotional balance and higher well-being (Nahas et al., 2000; WHO, 1998). Under this condition, an individual may feel that their social environment is safe and pleasant (Evers et al., 2013; Nahas et al., 2000), possibly contributing to lower cognitive food restraints and favoring emotional eating, due to positive affect (Evers et al., 2013). Although we did not evaluate positive affect in the present study, we found psychological lifestyle to be negatively associated with negative affectivity, a possible sign of the presence of positive affect.

Regarding our mediation analysis, for lifestyle second-order components we observed no significant mediations for eating behavior dimensions, except for the relation between psychological lifestyle on emotional eating. This result can be interpreted from the definition of emotional eating as the control of the eating process by an emotional feeling or situation (Canetti et al., 2002; Geliebter & Aversa, 2003; Karlsson et al., 2000). From this perspective, lifestyle behaviors may not have had a direct impact on emotional eating, but may have indirectly contributed to emotions that, in turn, may have interfered with the eating behavior, including episodes of uncontrolled eating (Canetti et al., 2002; Evers et al., 2013; Geliebter & Aversa, 2003).

In general, females have been found to be more likely to present emotional eating and cognitive restraint than males (Jáuregui-Lobera et al., 2014; Medeiros et al., 2017), while eating disorders have been more prevalent among individuals with higher BMIs (León-Vázquez et al., 2017). Our results are in accordance with those past findings, as we found females and individuals with higher BMIs to be more prone to cognitive restraint, uncontrolled eating, and emotional eating. Such results can be explained by the social pressure (Clark, 2019; Stunkard & Wadden, 1992) perceived by females and people with obesity conditions to limit their food intake to maintain or lose weight and reach a socially acceptable standard (Toral & Slater, 2007). If the imposed restraint generates a feeling of emotional deprivation (Alvarenga, Polacow, et al., 2019; Polivy, 1996), under stressful situations, then these individuals may lose control over their attempts at restraint (Macht, 2008) and engage in episodes of uncontrolled eating and emotional eating, leading to weight gain (Canetti et al., 2002); they may then resume food restraint when they perceive themselves outside the socially imposed attractiveness standard (Alvarenga, Polacow, et al., 2019). Thus, females and people with obesity conditions are more vulnerable to adopting extreme behaviors that may harm their health.

The negative association of age with uncontrolled eating and emotional eating can be explained by O’Brien (2015), who studied the psychological functioning of individuals of different ages and stated that the cognitive skill set for coping with different situations and emotions increases with age. Thus, the risk of binge-eating for coping with aversive emotional states would be lower in older than in younger people (Canetti et al., 2002; O’Brien, 2015).

The positive association we found between the participants’ economic stratum and their cognitive restraint can be associated with the various different body concerns that people with different income levels can present. Silva et al. (2017) found in a Brazilian sample that individuals with higher income were more concerned about their weight than their counterparts with lower income. Considering these concerns, these higher income individuals may engage in restrictive eating strategies to maintain or lose body weight, including severe and/or extreme strategies that can risk their health (Alvarenga, Figueiredo, et al., 2019).

Thoughts of dropping out of a course might be associated with negative emotions (Canetti et al., 2002; Macht, 2008) that trigger episodes of emotional eating and uncontrolled eating (Canetti et al., 2002; Saat et al., 2014). We observed this pattern in mediation relationships. Thus, health-promoting strategies should take into account the biopsychosocial reality of the target population to achieve consistent and effective results.

Conclusion

In this study of college students, the physical component of lifestyle was positively associated with negative affectivity, highlighting the importance of nutrition and physical activity for health status and prevention of mental health problems. We found the psychological component, in turn, to be negatively associated with negative affectivity, reinforcing the relevance of maintaining positive social relationships and emotional control. Lifestyle, negative affectivity, and sample characteristics such as sex, age, weight status and economic stratum impacted dimensions of eating behavior among college students, indicating that these factors should be considered when planning preventive, educational and interventional strategies for eating behavior in this population.