|
Associated factors of Dietary Diversity among Students of Gopalganj Science and Technology University: Using an Ordinal Logistic Regression
Rakibul Hasan1, Gazi Mohammad Mahbub1,[*]
1Assistant Professor, Department of Economics, Gopalganj Science and Technology University.
|
|
Keywords |
|
Abstract |
|
Food consumption, Dietary diversity, University students, Associated factors, Ordinal Logistic Regression. |
|
Background: Dietary diversity, a reflection of nutritional sufficiency, is essential to the health and academic performance of university student. They are frequently deal with changing lifestyles, fewer food alternatives, irregular meals, and budgetary restraints that impact their dietary diversity score. This study measures the dietary diversity score (DDS) and examines at how university students' dietary diversity is influenced by socioeconomic factors. Methods: Using the simple random sample technique, 411 students participated in cross-sectional study. The Food and Agriculture Organization's suggested Dietary Diversity score includes 12 food groups that are categorized as low, medium, and high, which is a dependent variable. Independent variable included demographic, socio-economic and anthropometric factors. Ordinal logistic regression and Chi-Square Test of Independence were conducted for appropriate estimation. Results: Among the students in 12 food categories, the highest consumption rates around more than 80% were found in food grains, vegetables, and white tubers and roots whereas milk, fruits, and sweets were consumed at the lowest rates. Considering eating habit, gender is significantly associated with skipping dinner and religion is associated with skipping breakfast. An ordinal logistic regression revealed that personal expense (B = .000, p = .009), academic result (B = −1.059, p= .011), and family income (B = −0.667, p = .023) were significant predictors. Conclusion: The results highlight the significance of demographic and socio-economic status in determining dietary diversity. In order to improve their eating habits as well as dietary diversity score, university authority and policymakers should take these variables into account. |
Introduction
University life is a time of change, characterized by new social experiences, academic pressures, and a greater degree of freedom. Students frequently develop new eating habits during this time that can have a big impact on their long-term health and nutritional status (Almoraie et.al. 2024). Most of the students facing inadequate dietary habits due to insufficient guardian nursing, long time academic activities, lack of nutrition understanding, low socio-economic status impact on their overall dietary patterns. Additionally, they were exposed to changes in dietary habits that led to their consuming more unhealthy foods, which had an adverse effect on their general health. Food choices are significantly influenced by stress, and the extent of this influence differs by gender (Ndlovu, 2017). Although it is commonly recognized that maintaining healthy food habits is essential to preventing diseases in adulthood, there has been a growing concern about the potential negative effects on public health arising from bad dietary patterns among youth and students over the past twenty years (Yun, Ahmad & Quee 2018, Ogden, et. al. 2006). The majority of youth consider the move to college or university as a time of greater independence, especially when it comes to food preferences. These food choices are related to the lack of appropriate dietary guidance available to them while they are studying, the difficulties they face in their academic fields, and the stress of choosing, buying, and preparing meals (Mogeni & Ouma 2022, Al Qauhiz 2010). Their unhealthy dietary habits may lead to a variety of chronic illnesses, such as obesity, hypertension, asthma, heart conditions, and malnutrition, which have an adverse impact on students' academic performance and general health (Abdel-Megeid, et. al. 2011, Qauhiz 2010).
Individuals' dietary diversity is evaluated to see how adequate and high-quality their diets are, which is important for health outcomes. It acts as a nutrient intake indicator and correlates with a number of health indicators, such as biological aging, child growth, and the treatment of chronic diseases. Dietary diversity reflects access to a diversity of foods, which is essential for nutrient suitability (Hailemariam et al., 2018). Higher dietary diversity scores (DDS) are associated with improved nutrient intake, as seen in studies where individuals with higher DDS had better nutrient profiles (Kojima et al., 2020). For populations with chronic conditions, such as diabetes and hypertension, dietary diversity is vital for managing health and preventing complications (Kal et al., 2016). Assessing dietary diversity in individuals with varying nutritional needs and health statuses involves a multifaceted approach that considers both the quantity and quality of food consumed. Dietary diversity denotes to the variety of distinct meals or dietary groups consumed within a specified time frame (often in 24 hrs), irrespective of the frequency of intake. (Vam, W. 2008).
Rational of the Study
University or college students have a high rate of poor eating habits and are associated with increased occurrence of lifestyle and chronic diseases. (Mogeni & Ouma 2022). According to 48th annual report-2021 of University Grant Commission of Bangladesh, there were total 44,41,717 students in public (including affiliated and constituent Colleges/Madrasas) and private universities among them 54 percent are male student and 46 percent are female and notable amount 93 percent of student are studied in public university where only 7 percent studied in private university. University or college students are considered young adults generally defined as 18 to 22 or 18 to 25 years old during the transition from adolescence to adulthood (Arnett 2023). This age group is frequently neglected as a distinct demographic with distinct needs, whereas female, pregnant women, children, and adolescents receive more attention. In Bangladesh, the youth population (age group 15-24) was 18.16% in 2011 and it has increased to 19.11% in 2022 which number is about 3 crore and 15 lakh (Population and Housing Census 2022). If this age group (15-24) is healthy, they can contribute more to the country's development than unhealthy ones. This study studied the respondents' BMI and dietary variety scores together with their socioeconomic status and eating habits to provide information about their health.
The research was conducted solely at GSTU, which has approximately 10,000 students (per year 1505 enrolled) including around 2,000 hall residents, while nearly 4,000 reside in mess arrangements near the campus. University halls typically provide structured accommodation and student-support facilities, whereas mess settings vary in services and amenities. Therefore, the findings may not be generalizable to students from other universities or to non-residential student populations beyond GSTU due to differences in living arrangements, institutional environments, and support systems.
Research Objectives
The specific research questions addressed by this study were: (i) what is the DD situation among university students by measuring DD? and (ii) what are the factors that affect significantly on DD score? To get the answer of the research question, the paper's particular goals are:
Review of Literature
Dietary Diversity Score (DDS), compared to any other tools, can indeed provide a good estimate of the nutrient adequacy of a population's diet (Kennedy, Ballard & Dop, 2011). DDS is calculated by counting the number of different food groups consumed within a specific timeframe, and it's a widely used indicator of diet quality. Studies have shown a strong correlation between DDS and the nutrient adequacy of a diet, indicating that greater dietary diversity generally leads to better nutrient intake.
Dietary diversity among university students is influenced by various factors, including socioeconomic status, food security, and personal dietary habits. Beside this health-related research requires anthropometric indicators to be one of the mandatory factors that includes weight, height, head circumference, BMI, and body circumferences like waist and hip (Mogeni, & Ouma, 2022). Many students suffer from insufficient nutritional diversity, according to research, as a result of a confluence of lifestyle decisions, economic difficulties, and scholastic demands. Micronutrient intake was positively and significantly connected with dietary diversity scores, which were defined by family eating patterns at the population level across a one-day recall period (Kennedy, 2009). Adolescents' diets in underdeveloped nations are less varied, particularly in terms of fruits and vegetables, which results in low energy consumption and inadequate intake of micronutrients (Ochola & Masibo, 2014). The study indicated that there are significant socioeconomic differences between public and private universities in Bangladesh, which mostly affect their monthly spending and eating patterns. Where the correlation between economic status and dietary diversity score was shown to be significant, with private university students scoring better on food intake (Rizwan et all 2023).
Research continuously shows that a sizable fraction of college students has less than ideal dietary diversity. For example, according to a study conducted in Bangladesh, almost half of college students have inadequate dietary diversification (Moon et al., 2024). Similar to this, studies conducted in a variety of contexts show that a large number of students fall short of food-based dietary guidelines and frequently consume inadequate amounts of fruits and vegetables and a lot of fast food, sugary snacks, and convenience meals (Avram et al., 2024). University time is a crucial because that time of dietary decisions can have a big impact on long-term health consequences. Beside with Sociodemographic factors, cooking methods, independent living, academic pressure significantly influenced on their dietary pattern (Kaewpradup et al. 2024, Ochola, S., & Masibo, P. K. 2014). According to Mogeni & Ouma 2022, the majority of students said they followed healthy lifestyle choices, including as regular exercise, avoiding from alcohol and tobacco, and eating a balanced diet. Their BMI was normal (69.4%), and they cooked their own food (57.0%). In coastal Kenya, there were significant variations in the consumption of fruits and vegetables, eating habits, consumption of a balanced diet, and dietary diversity. Poor nutritional status and improper dietary/eating habits are now directly associated with the growing incidence and rising prevalence of lifestyle and chronic illnesses, such as diabetes, obesity, and cardiovascular diseases (Nasreddine et al., 2020; Tanton et al., 2015). To sustain health and support appropriate physical and mental growth, a person's diet must be sufficient in both amount and variety throughout their life. (Haque, S., and others, 2023). Despite the fact that dietary diversity score (DDS) has been shown in numerous studies to be a valid and widely used indicator of nutrient adequacy and diet quality across populations, including university students. There is still little targeted research on the precise factors influencing dietary diversity among university students in Gopalganj Science and Technology University, especially when taking into account the combined effects of socioeconomic status, anthropometric measures, and lifestyle factors like academic pressure and independent living.
Methodology
Students who are enroll in a university and lived in university hall and private mess and age between 18-25 were randomly selected from the university area from November 2024 to December 2024. Total 411 students 31% (129 respondents) are female and 69% (282 respondents) males were invited to participate in the study. All the participants where a written informed consent was obtained.
Data collection
Each student who consented to participate was given a self-report questionnaire where information was split into 3 categories. First part of questionnaire included about personal and academic information like respondent age, BMI score, marital status, living status, academic result, department and family socio-economic information like family education and occupation status, income level, religious status. Part two consist of different question about meal pattern and food habit like sources of meals, skipping meals, frequency of their meals and final segment shows a dietary sheet known as a DDS score sheet (12 items) that records food intake over the previous 24 hours that contain 12 items food such as Food grain/ Cereals, White tubers and roots, Vegetables, Fruits, Meat and organ meat, Fish, Eggs, Legumes, nuts & seeds, Milk and milk products, Oil and fats, Sweet, Spices, condiments, beverages. The study was explained to the students, who also received guidance on how to honestly and completely complete the questionnaire.
Sample Size determination
Pourhoseingholi (2023) and Naing et. al. (2006) suggested that there is popular basic formula required to determine the minimum sample size for a prevalence study. They suggested the simple formula adopted from Daniel (1999) which is:
Where, n=sample size, Z = Z statistic for a level of confidence, P = expected prevalence or proportion, and d = precision
Considering, For the level of confidence of 95%, which is conventional, Z value is 1.96, prevalence or population proportion is 50%, then P is equal to 0.5. if investigators want the width of CI as 10% (0.1), d should be set at 0.05. Then the formula is calculated as follows:
=384.16
The optimal sample size of 422 was achieved by accounting for a 10% non-response rate, which was used to compensate for potential non-responses.
Measurements
Dietary Diversity Score: In this study we show the dietary diversity score which is combination of 12 food groups is used in guidelines created by FANTA and approved by FAO for measuring individual level diversity that records food intake over the previous 24 hours. The goal of this structure is to establish a relationship between household dietary energy intake and dietary diversity (Kennedy, 2009). The amount of different food groups that each participant consumed during the recall period was added up to determine the Dietary Diversity Score (DDS). Participants were divided into three groups according to their total DDS: High DDS: scores between 9 and 12, Medium DDS: scores between 6 and 8, and Low DDS: scores of 5 or below (≤ 5) (Seyoum et al. 2024).
Anthropometric Data: Participants self-reported their heights and weights then researcher measured the body mass index (BMI) by using standardized methods. Body weight measured in Kg and height of the respondent measured in meters. Body mass index (BMI) was calculated as weight (kg) divided by height (meters) squared. According to the National Institutes of Health, adults were classified based on their BMI to underweight BMI value is less than and equal to 18, normal BMI range is 18.5-24.9 where overweight BMI value is 25 - 29.9 and finally obese BMI is greater and equal to 30. BMI (Body Mass Index) is categorized as follows: Underweight (less than 18.5), Healthy Weight (18.5 to 24.9), Overweight (25 to 29.9), and Obese (30 or greater)
Other Socio-economic variables: Five educational levels of students first, second, third, fourth years and master's program were included in this study. Beside these four educational levels were considered by the parents: no education, primary and secondary and higher education. The parents worked as day laborers, self-employed (own business), and job holder. The parents' monthly income was divided into three categories: less than 12,000, between 12,001 and 21,000, and more than 21,000 BDT, following income categorization used in the Household Income and Expenditure Survey (Bangladesh Bureau of Statistics, 2016). All dependent and independent variables and other categorical variables were selected based on the literature review from sultan et. al. (2019), kumar et al (2020), Kabir et. al (2018) and Moon et al. (2024).
Data processing
We checked the filled surveys for consistency in order to clean them. After that, the data were put into an Excel sheet, and sorting and frequency distribution were used to perform logical tests. Lastly, the statistical product analysis was used to examine the data by using SPSS.
Analytical techniques for logistic regression
According to Bilder, C. R., & Loughin, T. M. (2014), Ordinal regression is a statistical technique that is used to predict behavior of ordinal level dependent variables with a set of independent variables.
Table 1: List of variables for Ordinal logistic regression in the study
|
Variable Name |
Measurement Scale |
Coding/description |
|
Dependent Variable |
|
|
|
Dietary Diversity Score (DDS) |
Ordinal |
1= Low DDS 2= Medium DDS 3= High DDS |
|
Independent Variables |
|
|
|
Gender
|
Categorical
|
0=Female 1=Male |
|
Religion
|
Categorical
|
0=Non-Muslim 1=Muslim |
|
Education of HH
|
Ordinal |
0=no education 1=primary 2=secondary 3=higher |
|
Family Income (Monthly)
|
Ordinal |
0=low income 1=Middle income 2= high income |
|
Employment of HH
|
Categorical |
1= self employed 2=Job holder/ Employee 3= Day labor |
|
BMI Status
|
Ordinal |
0=Underweight 1= Healthy weight 2= Overweight 3= Obese |
|
Academic Result |
Continuous |
Continuous |
|
Personal Expense (Monthly) 1 unit = 1000 Tk |
Continuous |
Continuous |
|
Family Member |
Continuous |
Continuous |
Model Specification
Regression analysis and descriptive statistics were employed in the study to meet its goals. Ordinal Logistic Regression model will use because dependent variable has ordinal form like low dietary diversity, medium dietary diversity and high dietary diversity (O'Connell, 2011). The ordinal logistic regression model can be defined as
logit (P(Y ≤ j)) = βjo + βj1X1 + ⋯ + βjpxp.
for j=1,…., J-1 and P predictors. Due to the parallel lines assumption, the intercepts are different for each category but the slopes are constant across categories, which simplifies the equation above to
logit (P(Y ≤ j)) = βjo + β1X1 + ⋯ + βpxp.
Where as The log odds is also known as the logit, so that
= logit (P(Y ≤ j))
To examine the association between categorical variables, a Chi-Square Test of Independence was conducted and a chi-square test was run with a 0.05 α threshold. Chi square (χ2) tests were performed to compare dietary diversity, meal pattern and BMI outcomes of different group like gender, religion, income, education etc. Statistical significance was defined at a p-value < 0.05 which is 5% significant level and <0.10 which is 10%. All test were assessed with SPSS (version 25, IBM).
Results and Discussion
Demographic characteristics of respondents
The table presents the demographic and socio-economic distribution of 411 respondents. In terms of gender, males found the majority with 69% (282 individuals), while females represent 31% (129 individuals). The religious affiliation is predominantly Muslim at 83% (342), compared to 17% (69) non-Muslim participants. The distribution by year of study reveals that the largest group is in the 4th year (38%), followed by 3rd year (25%), Masters (16%), 2nd year (14%), and 1st year (7%).
Table 2: Demographic characteristics and socio-economic status of the respondents
|
Sample Size (N=411) |
Frequency Distribution |
Percentage Distribution |
|
Gender |
|
|
|
Male |
282 |
69% |
|
Female |
129 |
31% |
|
Religion |
|
|
|
Muslim |
342 |
83% |
|
Non-Muslim |
69 |
17% |
|
Year of the Study |
|
|
|
1st Year |
30 |
7% |
|
2nd Year |
56 |
14% |
|
3rd Year |
104 |
25% |
|
4th Year |
156 |
38% |
|
Masters |
65 |
16% |
|
Living Status of family |
|
|
|
Rural |
351 |
85% |
|
Urban |
60 |
15% |
|
Education of Household Head |
|
|
|
No education |
53 |
13% |
|
Primary education |
118 |
29% |
|
Secondary education |
177 |
43% |
|
Higher education |
63 |
15% |
|
Income of Family (monthly) |
|
|
|
Low-income |
139 |
34% |
|
Medium-income |
169 |
41% |
|
High-income |
103 |
25% |
|
Employment of HH (n=409) |
|
|
|
Self Employed |
291 |
71% |
|
Employee |
36 |
9% |
|
Day Labor |
78 |
19% |
Sources: Authors calculation based on survey data
Family socio-economic status of the respondents
The education level of household heads varied: 13% had no formal education, 29% had completed primary education, 43% had attained secondary education, and 15% had higher education. Regarding household income, a substantial portion of the sample (41%) fell into the medium-income (12000 to 21000 BDT) category, followed by 34 % in the low-income group and only 25% which is 103 respondents out of 411 family had more than 21000 BDT. This data indicates that the majority of households are clustered in the low to middle-income range. Employment data (n = 409) revealed that 71% of household heads were self-employed, indicating a significant reliance on informal or small-scale economic activities. Only 9% were salaried employees, and 19% worked as day laborers. Family size data showed that 82% of households had less than and equal to 6 members which is called nuclear family and 18% had more than 6 members their mean value of family member is 5.24.
Dietary Diversity Score (DDS) and Body Mass Index (BMI) Status of the respondents
The participants' diets were relatively diverse, as indicated by the average Dietary Diversity (DD) score of 8.03 (SD = 1.93) which indicates mean value of score fall in medium DDS category. The data also presents the mean value of Body Mass Index (BMI) which is 21.61. The lowest mean value of BMI is 14.11 and highest BMI mean value is 33.45.
|
|
|
Figure 1: DDS Status and BMI Status of the respondents
The respondents' percentage distribution of Body Mass Index (BMI) and Dietary Diversity Status (DDS) is shown in the pie chart. According to the findings, most participants (52%) had a medium DDS, which denotes a moderate level of dietary variety. A high DDS was attained by 38% of the respondents, indicating that they ate a variety of food types. Only 10% of respondents reported having poor DDS, which is evidence of a lack of dietary diversity. However, the majority (81%) fell into the healthy weight range, whereas only 1% were classed as obese, 9% were underweight, and 9% were overweight.
Food habit behavior of the university students
Table shows, majority 67% which is 275 out of 411students sometimes skip breakfast, and 14% do so daily, indicating poor morning eating habits only 80 students never skip their breakfast. In contrast, dinner is more consistently consumed, with 70% of total respondent never skipping it. Junk food intake is common, as 79% report occasional consumption, though only 5% consume it daily and 16% avoid it altogether. Notably, around 66% sometimes feel weakness after meals, and 4.9% experience it daily, suggesting potential dietary inadequacies. However, 29.4% never report such weakness, implying that nearly one-third maintain adequate nutritional intake.
Table 3: Dietary Behaviors among university students
|
|
Skipping Breakfast |
Skipping Dinner |
Eating junk food |
Post-meal weakness |
||||
|
|
Total |
Percentage |
Total |
Percentage |
Total |
Percentage |
Total |
Percentage |
|
Everyday |
56 |
14% |
26 |
6% |
21 |
5% |
20 |
4.9% |
|
Never |
80 |
19% |
287 |
70% |
64 |
16% |
121 |
29.4% |
|
Sometime |
275 |
67% |
98 |
24% |
326 |
79% |
270 |
65.7% |
Sources: Authors calculation based on survey data
In this survey, data shows that 70% of respondents eat 3–4 meals daily, indicating adequate intake, while 30% consume only 1–2 meals, suggesting possible food insecurity or insufficient dietary practices. Most of the university students come into the campus by the university bus in the morning and they engage in academic and extra-curricular activities during that day. Around the day they eat variety of food consume and some are not consumed. The bar graph shows that how student tackle the short-term hunger in this campus.
Figure 2: Bar Graph of Tackling short term hunger
Among 411 students 48.2% (n =198) said they relied on unhealthy food to satisfy their short-term hunger, while 22.6% (n =93) said they were still hungry. Of the participants, only 29.2% (n = 120) said they had access to healthy meals like fruits.
Table presents the association between students’ food habits and selected demographic variables (gender, religion and faculty) using chi-square test where p-values indicate statistically significant association. The results for each variable are summarized in following table.
Table 4: Chi-square test result on food taking behavior among university students
|
Food habit among students |
Gender |
Religion |
|
|
p-value |
|
|
Skipping Breakfast |
.991 |
.069** |
|
Skipping Dinner |
.011* |
.729 |
|
Eating Junk Food |
.281 |
.760 |
|
Tackling Short-term hunger |
.019* |
.025* |
|
Feel weakness after taking meal |
.012* |
0.55** |
|
*p < 0.05: Statistical significance (5%) **p < 0.1: Statistical significance (10%) |
||
Sources: Authors calculation based on survey data
The analysis revealed a significant association between gender and food habit among student were skipping dinner (p = .011), tackling short-term hunger (p = .019), feeling weakness after meals (p = .012), suggesting that male and female students differ in their dinner habits, meet up short term hunger and weakness after meals. On the other hand, Religion was marginally associated with skipping breakfast (p = .069) findings suggest religious affiliation may influence breakfast habits, with non-Muslim students more likely to skip breakfast daily despite being a smaller proportion of the sample. Religion is also near significant association with weakness after meals (p = .055), suggesting that religious practices or cultural norms may influence breakfast habits. A significant relationship was found between religion and tackling short-term hunger (p = .025). Only food habit of eating junk food is not statistically significant or no association among gender, religion and faculty.
Dietary Diversity among University Students
The figure presents the percentage distribution of food group consumption among 12 different food groups and indicates the proportion of respondents who consumed each group ("Yes") versus those who did not ("No"). The highest consumption was observed in the Food grain/cereals category, with 99.51% of individuals reporting intake, indicating it is a staple in most diets. White tubers & roots and Vegetables also showed high consumption rates of 83.45% and 80.29%, respectively.
In contrast, Milk and milk products had the lowest intake, with only 37.47% consuming them and 62.53% not consuming them, indicating risk of calcium/vitamin D deficiency. Other food groups with relatively lower intake include Fruits (48.66%), Sweet (47.69%), and Legumes nut & seeds (63.75%). Moderate consumption levels were found for Meat & organ meat (63.75%), Fish (65.21%), and Eggs (68.13%) that means over 30% of respondents do not consume eggs, fish, and meat that also indicate limited access or dietary restrictions. Beside this 78.59 percent cooked with cooking oil and fats. Overall, the table illustrates a varied pattern of food group consumption, suggesting that while staples like grains are widely consumed, other nutrient-rich groups like dairy, fruits, and sweets are less frequently included. Using Chi-square tests, this table also examines the relationship between students' food group consumption and their gender, religion, household education, income, and BMI.
Figure 3: Types of food groups consumed by the university students
The results reveal several significant associations, indicating that demographic and socioeconomic factors influence dietary choices.
Table presents the statistical associations between the consumption of various food items and key socioeconomic and demographic variables among 411 respondents. Significant relationships (p < .05) and marginal associations (p < .10) are highlighted. According to the food consumption
Table 5: Food consumption associated with socioeconomic and demographic factor
|
Food list |
Total N=411 |
Gender
|
Religion
|
Education of HH |
Family Income |
BMI
|
|
|
|
p-value |
||||
|
Food grain/ Cereals |
409 |
.338 |
.524 |
.389 |
.237 |
.924 |
|
White tubers & roots |
343 |
.214 |
.203 |
.892 |
.376 |
.319 |
|
Vegetable |
330 |
.339 |
.389 |
.598 |
.667 |
.099** |
|
Fruits |
200 |
.000* |
.345 |
.001* |
.898 |
.471 |
|
Meat and Organ Meat |
262 |
.000* |
.003* |
.143 |
.568 |
.024* |
|
Fish |
268 |
.979 |
.998 |
.750 |
.121 |
.879 |
|
Egg |
280 |
.072** |
.089** |
.020* |
.929 |
.650 |
|
Legumes nut & Seeds |
262 |
.475 |
.171 |
.090** |
.171 |
.020* |
|
Milk |
154 |
.464 |
.391 |
.001* |
.447 |
.070** |
|
Oil |
323 |
.381 |
.125 |
.029* |
.864 |
.936 |
|
Sweet |
196 |
.918 |
.008* |
.246 |
.001* |
.987 |
|
Spices, condiments & beverages |
274 |
.007* |
.401 |
.664 |
.000* |
.173 |
|
*p < 0.05: Statistical significance (5%), **p < 0.1: Statistical significance (10%) |
||||||
Sources: Authors calculation based on survey data
Gender was found to be significantly associated with the consumption of several food items: Fruits (p = .000), Meat and organ meat (p = .000), Spices, condiments & beverages (p = .007) and Egg consumption showed a marginal association with gender (p = .072). These results suggest notable gender-based dietary preferences, particularly in the intake of protein-rich and flavorful food items. Religious affiliation significantly influenced with Meat and organ meat consumption (p = .003) and sweet consumption (p = .008) but a marginal relationship was found with egg consumption (p = .089). These findings suggest that cultural and religious norms may influence food choices among students, particularly in animal-based and sweet food items.
The education level of the household head significantly associated with Fruit consumption (p=.001), Egg consumption (p = .020), Milk consumption (p = .001) and Oil consumption (p = .029) but a marginal association was also found with legumes, nuts, and seeds (p = .090). These results highlight the role of educational attainment in promoting dietary diversity and nutrient-rich food intake. Income level was significantly associated with only Sweet consumption (p = .001) and Spices, condiments & beverages (p = .000). Whereas BMI status was significantly associated with Meat and organ meat (p = .024), Legumes, nuts, and seeds (p = .020) and marginal associations were observed with Vegetables (p = .099) and Milk (p = .070). Overall, gender, religion, education, income, employment and BMI were all significantly linked to specific food groups. While staple foods like cereals, tubers, and fish showed no significant variation.
Logistic Regression Model Result
In order to determine whether the final model suited the data better than a model that just included the intercept, a model fitting test was performed. The results showed that the addition of predictors improved model fit χ²(15) = 39.968, p <.001. The model's fit to the observed data was evaluated using the Goodness-of-Fit statistics. Both the Deviance Chi-Square test χ²(751) = 687.872, p =.951) and the Pearson Chi-Square test χ²(751) = 785.392, p =.186 were not significant. There is no discernible difference between the observed and projected values, indicating that the model fits the data well, as both p-values are higher than.05. As a result, the data is adequately fitted by the ordinal logistic regression model.
|
Table 6: Model Fitting Information |
||||
|
Model |
-2 Log Likelihood |
Chi-Square |
df |
Sig. |
|
Intercept Only |
729.227 |
|
|
|
|
Final |
689.259 |
39.968 |
15 |
.000 |
|
Link function: Logit. |
||||
Sources: Authors calculation based on survey data
The model explains a moderate amount of the variance in the outcome variable, according to the -pseudo-R-squared values. In particular, the McFadden R2 was.055, the Nagelkerke R2 was.115, and the Cox and Snell R2 was.097. These figures imply that between 9.7% and 11.5% of the variation in the dependent variable can be explained by the model. The proportional odds assumption was evaluated using the Test of Parallel Lines. With χ²(15) = 12.634 and p =.631, the result was not significant, suggesting that the assumption is correct. As a result, using the logit link function in conjunction with ordinal logistic regression is appropriate.
Table 7: SPSS result of goodness-of-fit and Pseudo R-square
|
Goodness-of-Fit |
Pseudo R-Square |
||||
|
|
Chi-Square |
df |
Sig. |
Cox and Snell |
.097 |
|
Pearson |
785.392 |
751 |
.186 |
Nagelkerke |
.115 |
|
Deviance |
687.872 |
751 |
.951 |
McFadden |
.055 |
|
Link function: Logit. |
|||||
Sources: Authors calculation based on survey data
An ordinal logistic regression was conducted to examine the socio-demographic and economic predictors that impact on dietary diversity status (DDS), categorized as low, medium, and high.
Table 8: Revealed results of the parameter estimates for ordinal logistic regression.
|
|
Estimate |
Sig. |
95% Confidence Interval |
||
|
Lower Bound |
Upper Bound |
||||
|
Academic Result |
|
-1.059 |
.011* |
-1.874 |
-.244 |
|
Personal Expense |
|
.174 |
.009* |
.043 |
.306 |
|
Family Member |
|
.071 |
.242 |
-.048 |
.189 |
|
Gender (ref=male) |
Female |
-.198 |
.379 |
-.639 |
.243 |
|
Religion (ref= Muslim) |
Non-Muslim |
-.239 |
.384 |
-.777 |
.299 |
|
Education of HH (ref=higher education) |
No education |
.267 |
.549 |
-.606 |
1.140 |
|
Primary |
-.220 |
.544 |
-.929 |
.489 |
|
|
Secondary |
-.610 |
.073** |
-1.278 |
.058 |
|
|
Family Income (ref=high income) |
Low-income |
-.677 |
.023* |
-1.260 |
-.095 |
|
Middle-income |
-.304 |
.244 |
-.816 |
.207 |
|
|
Employment of HH (ref=Day Labor) |
Self Employed |
.489 |
.089** |
-.074 |
1.051 |
|
Employee |
.049 |
.922 |
-.939 |
1.038 |
|
|
BMI Status (ref=Obese) |
Underweight |
1.189 |
.268 |
-.917 |
3.295 |
|
Healthy Weight |
1.156 |
.257 |
-.844 |
3.155 |
|
|
Overweight |
1.439 |
.174 |
-.637 |
3.515 |
|
|
Link function: Logit. a. This parameter is set to zero because it is redundant. Ref= reference category *p < 0.05: Statistical significance (5%), **p < 0.1: Statistical significance (10%) |
|||||
Sources: Authors calculation based on survey data
The results from the ordinal logistic regression identified several significant predictors of dietary diversity status (DDS) among individuals. Academic performance emerged as a significant predictor, with a negative estimate (β = -1.059, p = .011), indicating that students with lower academic achievement were more likely to fall into the lower DDS categories. This finding reinforces prior evidence that higher diet quality linked to better academic performance (Al‑Saadi, et al., 2020). Personal expense was also a significant predictor (β= .174 p = .009), suggesting that students with greater financial capacity had improved dietary diversity. This supports the concept that economic resources enable access to a wider variety of food groups, enhancing diet quality (Kennedy et al., 2011). Similarly, respondents from low-income households were significantly less likely to achieve higher DDS scores (β = -0.677, p = .023), in line with research highlighting poverty and food insecurity reduce dietary variety in rural Bangladeshi adolescents (Khan et al., and Islam et al., 2020).
Education of Household Head (Secondary) are marginally
associated with lower dietary diversity compared to higher
education (β=-0.610, p=0.073) that means having a
household head with secondary education is marginally
associated (10% level) with lower dietary diversity compared
to higher education. This trend is consistent with research in
Nepal, where adolescents whose mothers had secondary education
experienced significantly higher odds of inadequate dietary
diversity compared to peers whose mothers had higher education
(Dahal et al., 2022). Employment of Household Head (Estimate =
0.489, p = .089) that means
students whose household head is self-employed are marginally
more likely to have higher dietary diversity compared to those
whose household head is a day laborer.
However, it is important to critically note that several theoretically relevant variables gender, religion, family member, and BMI status did not significantly predict DDS (p > .05). Academic performance, personal expenses, and family income are the most significant predictors of dietary diversity. Household education and employment also show marginal influence, suggesting the broader socioeconomic environment plays a crucial role in students' dietary quality.
Limitations
This study, which focuses on students from Gopalganj Science and Technology University, has several limitations, including its cross-sectional design, dependence on self-reported dietary data, and restricted geographical scope. The results may not be generalizable to other institutions or regions, and additional research is required to examine long-term trends and external factors.
Conclusion
The study demonstrates that gender, religion, education, income, employment status, and BMI significantly influence the consumption of various food groups, particularly protein-rich and nutrient-dense items. These results align with existing literature emphasizing the role of socio-demographic and cultural factors in shaping dietary behavior (Ruel, 2003; Kennedy et al., 2011).
The analysis of logistic regression identified academic performance, personal expense, and household income as significant predictors of dietary diversity status (DDS). Specifically, those with lower academic achievement and limited financial resources were more likely to exhibit lower dietary diversity. This highlights a clear linkage between socioeconomic status, educational outcomes, and nutritional quality, aligning with existing literature that connects better academic and economic conditions to healthier dietary behaviors highlighting the role of educational and economic factors in shaping students’ dietary behavior. Students with better academic outcomes and greater financial capacity demonstrated higher DDS, consistent with prior research linking socio-economic status to healthier diets (Mellor et al., 2014; Kennedy et al., 2011; Ruel, 2003). While other variables like gender, religion, and BMI were not significant, their theoretical relevance suggests a need for broader models. The results underscore the importance of addressing structural and contextual barriers to improve dietary diversity among students.
However, several expected predictors such as gender, religion, household head’s education, employment type, and BMI status were not statistically significant in this model. These findings may be attributed to sample characteristics, contextual variables not captured (e.g., food availability, cultural norms), or limitations in measurement.
Recommendation
Based on the observed relationship between low academic performance and dietary diversity, educational institutions should integrate nutrition education programs targeting students with poor academic outcomes. These initiatives could improve both diet and learning performance, as suggested by Mellor et al. (2014). Given the significant influence of household income and expenses on dietary diversity, policymakers should consider implementing subsidized food programs or financial support for low-income students to ensure equitable access to a nutritious and diverse diet (Ruel, 2003). As gender and religious norms significantly affect food choices, culturally sensitive nutrition awareness campaigns should be developed in collaboration with community leaders to encourage balanced dietary practices.
To enhance dietary diversity, nutrition interventions should address cultural norms, educational disparities, and economic inequalities. Policymakers must implement context-specific strategies, including nutrition education and food subsidies, to ensure equitable access to diverse, health-promoting foods across populations.
References
Abdel-Megeid, F.Y., Abdelkarem, H.M. and El-Fetouh, A.M., 2011. Unhealthy nutritional habits in university students are a risk factor for cardiovascular diseases. Saudi Medical Journal, 32(6), pp.621–627. https://smj.org.sa/content/smj/32/6/621.full.pdf
Abu Ansar Md Rizwan, Banik, S.K., Anny, N.A., Ferdush, J., Nokshi, S.S., Hossain, A.M. and Hossen, A.B.M.M., 2023. Comparative analysis of dietary diversity and food consumption patterns among public and private university students in Bangladesh. International Journal of Biological and Pharmaceutical Sciences Archive, 6(2), pp.191–198. https://doi.org/10.53771/ijbpsa.2023.6.2.0123
Al Qauhiz, N.M., 2010. Obesity among Saudi female university students: dietary habits and health behaviors. The Journal of the Egyptian Public Health Association, 85(1–2), pp.45–59. https://doi.org/10.21608/jepha.2010.1457
Almoraie, N.M., Alothmani, N.M., Alomari, W.D. and Al-Amoudi, A.H., 2024. Addressing nutritional issues and eating behaviours among university students: a narrative review. Nutrition Research Reviews, pp.1–45. DOI: https://doi.org/10.1017/S0954422424000088
Arnett, J.J., 2023. Emerging adulthood: The winding road from the late teens through the twenties. Oxford University Press. https://academic.oup.com/book/55230
Avram, C., Nyulas, V., Onisor, D., Georgescu, I. M., Szakacs, J., & Ruta, F. (2024). Food Behavior and Lifestyle Among Students: The Influence of the University Environment. Nutrients, 17(1), 12. https://doi.org/10.3390/nu17010012
Bangladesh Bureau of Statistics (BBS), 2016. Household Income and Expenditure Survey 2016. Dhaka: Ministry of Planning, Government of the People’s Republic of Bangladesh.
Bangladesh Bureau of Statistics (BBS), 2022. Population and Housing Census 2022: Preliminary Report. Dhaka: Ministry of Planning, Government of the People’s Republic of Bangladesh. https://sid.portal.gov.bd/sites/default/files/files/sid.portal.gov.bd/publications/01ad1ffe_cfef_4811_af97_594b6c64d7c3/PHC_Preliminary_Report_(English)_August_2022.pdf
Bolo, A., Verger, E., Fouillet, H. and Mariotti, F., 2024. Exploring multidimensional and within-food group diversity for diet quality and long-term health in high-income countries. Advances in Nutrition, 100278. https://doi.org/10.1016/j.advnut.2024.100278
Dahal, M., Basnet, A., Khanal, S., Baral, K. and Dhakal, S., 2022. Gender difference in food choice and eating practice and their association with health among students of Kathmandu, Nepal. Journal of Obesity, 2022(1), p.2340809. https://doi.org/10.1155/2022/2340809
Daniel, W.W., 1999. Biostatistics: A foundation for analysis in the health sciences. 7th ed. New York: John Wiley & Sons.
Déré, K.A.L., Djohan, Y.F., Koffi, K.G., Manhan, K., Niamké, A.G. and Tiahou, G.G., 2016. Individual dietary diversity score for diabetic and hypertensive patients in Côte d’Ivoire. International Journal of Nutrition, 2(1), pp.38–47. DOI : 10.14302/issn.2379-7835.ijn-16-943
Hailemariam, T., Girmay, T. and Girmay, G., 2018. Determinants of individual dietary diversity score of children less than five years old in the southern zone of Tigray, Ethiopia. African Journal of Food, Agriculture, Nutrition and Development, 18(1). 10.18697/ajfand.81.16400
Kabir, A., Miah, S. and Islam, A., 2018. Factors influencing eating behavior and dietary intake among resident students in a public university in Bangladesh: a qualitative study. PLOS ONE, 13(6), e0198801. https://doi.org/10.1371/journal.pone.0198801
Kaewpradup, T., Tangmongkhonsuk, M., Chusak, C., Siervo, M. and Adisakwattana, S., 2024. Examining sociodemographic factors, reasons, and barriers in the diversity of fruit and vegetable intake among undergraduate students. Nutrients, 16(6), p.779. https://doi.org/10.3390/nu16060779
Kennedy, G., Ballard, T. and Dop, M., 2011. Guidelines for measuring household and individual dietary diversity. Rome: Food and Agriculture Organization.
Kennedy, G.L., 2009. Evaluation of dietary diversity scores for assessment of micronutrient intake and food security in developing countries. Wageningen: Wageningen University and Research. https://www.proquest.com/openview/a872ac30862c6ed0dc12f6f2ae69460d/1?pq-origsite=gscholar&cbl=2026366&diss=y
Khan, M.N., Rahman, M.M., Rahman, M.M., Khan, M.M.H., Rahman, M.S., Jesmin, S.S. and Islam, M.R., 2020. Household food insecurity and dietary diversity among early adolescents in a poor rural area of Bangladesh: a cross-sectional study. Public Health Nutrition, 23(9), pp.1536–1547. https://doi.org/10.1017/S1368980019004021
Kojima, Y., Murayama, N. and Suga, H., 2020. Dietary diversity score correlates with nutrient intake and monetary diet cost among Japanese adults. Asia Pacific Journal of Clinical Nutrition, 29(2), pp.382–394. https://search.informit.org/doi/10.3316/informit.286873244117699
Kumar, A., Ayub, A., Roy, R., Rai, A., Ameta, B., Latheef, A., Rani, A., Aysha, M., Ali, A.M., Azad, A.V. and Ashtami, K.H., 2020. Assessment of diet diversity and eating pattern of undergraduate students: a pan India study. https://pdfs.semanticscholar.org/190a/81f228d4322cafc759382d310813f123beb7.pdf
Mellor, J.M., Dolan, C.B. and Rapoport, R.B., 2014. Child health and academic achievement: evidence from the NLSY97. Economics of Education Review, 38, pp.100–106. https://doi.org/10.1016/j.econedurev.2013.10.006
Mogeni, B.K. and Ouma, L.O., 2022. Dietary patterns, behaviours, and their associated factors among university students in coastal Kenya. Cogent Food & Agriculture, 8(1). https://doi.org/10.1080/23311932.2022.2132873
Moon, M.I., Islam, T., Tohan, M.M., Shah, B.R., Juhi, N.J., Anni, T.T., Islam, A.M.M., Akter, K.M., Howlader, M.H. and Rahman, M.A., 2024. Dietary Diversity among Students of Khulna University, Bangladesh. Dr. Sulaiman Al Habib Medical Journal, 6(3), pp.92-104. https://journals.lww.com/dshmj/fulltext/2024/06030/dietary_diversity_among_students_of_khulna.2.aspx
Naing, L., Winn, T.B.N.R. and Rusli, B.N., 2006. Practical issues in calculating the sample size for prevalence studies. Archives of Orofacial Sciences, 1, pp.9–14. https://aos.usm.my/docs/Vol_1/09_14_ayub.pdf
Nanjundeswaraswamy, T.S. and Divakar, S., 2021. Determination of sample size and sampling methods in applied research. Proceedings on Engineering Sciences, 3(1), pp.25–32. https://oaji.net/articles/2021/8593-1615826931.pdf
Ndlovu, N.P., 2017. Relationship between diet quality, nutrition status and academic performance of first and non-first-generation university students in Durban (Doctoral dissertation). https://openscholar.dut.ac.za/bitstream/10321/2619/1/NDLOVU_NP_2017.pdf
Nisar, N., Qadri, M.H., Fatima, K. and Perveen, S., 2009. Dietary habits and lifestyle among the students of a private medical university Karachi. Journal of the Pakistan Medical Association, 59(2), pp.98–101. https://pubmed.ncbi.nlm.nih.gov/19260574/
Ochola, S. and Masibo, P.K., 2014. Dietary intake of schoolchildren and adolescents in developing countries. Annals of Nutrition and Metabolism, 64(Suppl. 2), pp.24–40. https://doi.org/10.1159/000365125
Ogden, C.L. et al., 2006. Prevalence of overweight and obesity in the United States, 1999–2004. JAMA, 295(13), pp.1549–1555. https://jamanetwork.com/journals/jama/fullarticle/202627
Pourhoseingholi, M.A., Vahedi, M. and Rahimzadeh, M., 2013. Sample size calculation in medical studies. Gastroenterology and Hepatology from Bed to Bench, 6(1), pp.14–17. https://pmc.ncbi.nlm.nih.gov/articles/PMC4017493/
Ruel, M.T., 2003. Operationalizing dietary diversity: a review of measurement issues and research priorities. The Journal of Nutrition, 133(11), pp.3911S–3926S. https://doi.org/10.1093/jn/133.11.3911S
Seyoum, A., Tsegaye, A., Samuel, A., & Amede, T. (2024). Factors influencing under five children dietary diversity score: a cross-sectional study in enset-based farming systems of Sidama Regional State, Ethiopia. J Human Nutr Food Sci, 12(1), 10-47739.
Sultana, M., Hasan, T. and Shaheen, N., 2019. Energy and nutrient intake and dietary diversity among female residential students of Bangladesh. Current Research in Nutrition and Food Science, 7(1), pp.244. https://doi.org/10.12944/CRNFSJ.7.1.24
University Grants Commission of Bangladesh, 2021. 48th Annual Report 2021. Available at: https://ugc.portal.gov.bd/sites/default/files/files/ugc.portal.gov.bd/annual_reports/104b8c32_dbe2_4c07_8228_76b999beff2a/2024-01-23-05-49-e04bbf524cb5120252822612c3fab6df.pdf
Vam, W. (2008). Food consumption analysis: Calculation and use of the food consumption score in food security analysis: Technical guidance sheet. Vulnerability Analysis and Mapping Unit, World Food Programme: Rome, Italy. https://www.humanitarianlibrary.org/sites/default/files/2014/02/wfp188184.pdf?utm_source=chatgpt.com
Yun, T.C., Ahmad, S.R. and Quee, D.K.S., 2018. Dietary habits and lifestyle practices among university students in Universiti Brunei Darussalam. The Malaysian Journal of Medical Sciences (MJMS), 25(3), pp.56. doi: 10.21315/mjms2018.25.3.6
Corresponding Author. E-mail: gazimohammedmahbub@gmail.com
DOI: doi.org/10.64172/ssr.2025.i3.09
