Skip to main content

Fruit and vegetable consumption across population segments: evidence from a national household survey



The 2002 World Health Report documented that low fruit and vegetable intake are among the top ten risk factors contributing to attributable mortality and up to three million lives could be saved each year by adequate consumption of F&V across the globe, leading an examination of behavioral preferences of the individual and family social, environmental, and behavioral factors that constitute perceived barriers to fruit and vegetable consumption.


The study examines factors affecting the choice of eating fruits and vegetables by household members and calculates eating frequency probabilities of different population-origin associated with personal attributes and behavior.


Turkish Health Survey (THS) 2019 data from the Turkish Statistical Institute (TSI) national representative household panel is applied. Estimating a random-effect bivariate probit model of fruit and vegetable choice, we calculated marginal probabilities of choosing fruits and vegetables, the joint probability of choosing both, and conditional probabilities between choosing to eat either, detecting consumption synergy.


The role of uncontrolled variables in choosing to eat fruits and vegetable (F&V) differs between the decision of an average family and the decision of individual family members. The attitude is positive for an average family and contrasts with the negative attitude among some family members. Most individual and family attributes inversely affect fruit and vegetable choice across different groups, while a positive relationship exists between the likelihood of fruit and vegetable choice and attributes such as age, marital status, education, weight, having health insurance, income, and time and forms of physical activity.

Conclusion and recommendation

Instead of a general policy for the implementation of a healthy and balanced nutrition program to improve fruit and vegetable eating frequency, it appears more effective to adopt programs with distinct characteristics that segregate society into different cohorts. We suggest appropriate policies and offer suitable approaches to reach targeted groups.


The 2002 World Health Report documented that low fruit and vegetable intake are among the top ten risk factors contributing to attributable mortality and up to three million lives could be saved each year by adequate consumption of F&V across the globe [1]. The daily, habitual intake of F&V can prevent major non-communicable diseases (NCDs) (some cancers and cardiovascular diseases) while eating a variety of produce daily provides the body with the most micronutrients, dietary fiber, and essential non-nutrients [2]. F&V, rich in phytochemicals such as polyphenols, terpenoids, and organic sulfides, play a role in adipose tissue growth and differentiation, apoptosis of adipocytes, and lipid and energy metabolism [3]. About 85% of the global disease burden attributable to low fruit and vegetable intake (LFVI) includes 31% of ischemic heart disease, 15% of cancers, and 11% of strokes [4]. Also, inadequate consumption of F&V causes 7.5% (from lack of fruits) and 7.6% (from lack of vegetables) of deaths from heart disease, stroke, and type 2 diabetes [5] While lowering food and beverage consumption provides a significant fall in both body mass index (BMI) [6] and regulates blood pressure (BP) and blood sugar [7], increased food and beverage consumption can alleviate the risk of depression [8] with more robust mental health [8] and provide positive psychological effects on humans [9]. Several countries, therefore, implemented public education campaigns to increase F&V consumption. Campaigns in Europe and North America have increased F&V consumption. The American National Academy of Sciences urged a change in F&V consumption habits of low-income families using the “Five a Day” program, which has been well-received over the past two decades [10].

Although most chronic diseases can be prevented by lifestyle choices such as healthy eating and physical activity [5, 10], in past studies, fruit and vegetable consumption decisions have been limited to socio-demographic, economic, and behavioral factors. Specific traits included gender, age, and education, while other characteristics accounted for psychological, behavioral, and biological factors [11, 12]. Measures of income and location complete the list. The majority of empirical studies focus on individual behavior, not accounting for a family, and therefore are difficult to implement in settings where individual decisions about fruit and vegetable consumption are made at the family level. Studies are lacking on the effects of factors on the fruit and vegetable consumption decisions of individuals within a family.

An examination of behavioral preferences of the individual and family social, environmental, and behavioral factors that constitute perceived barriers to fruit and vegetable consumption is needed. Identifying characteristics that sustain balanced, daily F&V consumption in emerging economies such as Turkey could increase inadequate intake and reduce the risk factors contributing to attributable mortality. The current study, therefore, aims to determine the empirical heterogeneity of the relationship between fruit and vegetable intake and individual and family traits using a random-effect bivariate probit model. The study uses nationally representative intra-family multi-level data from Turkey. The bivariate random-effects model tests the presence and strength of factors after controlling for within-family observable traits in each consumption event. This study, first in applying a simultaneous approach of both food types to F&V consumption, contributes to the existing literature by aiming to control the intra-familial heterogeneity. Considering the heterogeneity in dietary habits and food culture of individuals within the family and understanding what influences dietary choices in F&V consumption is crucial for the development of effective interventions leading to sustainable behavioral change. Increased F&V consumption is supported by the well-established link between F&V consumption, health maintenance, and disease prevention and is of great concern to policymakers. The benefits from higher F&V consumption are not limited to individuals but lower the costs of health services by public and private providers. As Turkey’s population grows, the potential cost savings could be substantial.

The current study also provides extensive empirical insights for food industry managers to create effective and dynamic marketing strategies. Food distributors and retailers can offer consumer nutritional education programs using knowledge of distinct individual profiles within a family generated from the current study. Formulating marketing strategies and nutritional programs tailored to consumer consumption patterns have the potential to be an effective intervention tool promoting greater fruit and vegetable intake. The current study provides robust information about diverse fruit and vegetable consumption, for example, the habits of those who consume fruits and vegetables separately (measured by marginal probability), those who consume both together (i.e., quantifying joint probability), and those who consume fruit (vegetables) when also consuming vegetables (fruit) (measured by conditional probability). In this context, information obtained from different consumption probability components will provide policymakers with qualified insights for more efficient use (e.g., balanced choice of nutrients) and distribution of national resources. Also, inferences derived from the current study facilitate the implementation of comparable actions in other countries with similar consumer socio-demographic, economic, and eating pattern characteristics as in Turkey.

Fruit and vegetable consumption and nutrient deficiency in Turkey

In Turkey, fruit and vegetable consumption follows the pattern of several major emerging and developed economies. The annual F&V consumption in Turkey was 32.87/122.33 kg, while in Brazil, Canada, China, France, Greece, England, and the United States was 19.97/26.97 kg, 29.65/71.43 kg, 36.77/330.68 kg, 30.98/65.06 kg, 53.15/77.13 kg, 22.23/55.93 kg, and 20.15/65.98 kg, respectively, in 2019 [13]. Per-capita fruit and vegetable consumption in Turkey is high when compared to other listed countries. The daily energy intake of fruits and vegetables per capita in Brazil, Canada, China, France, Greece, Turkey, England, and the USA, was 20/18 kcal, 44/58 kcal, 42/225 kcal, 46/52 kcal, 69/49 kcal, 41/70 kcal, 36/46 kcal, and 31/44 kcal, respectively [13]. Turkey’s per capita daily fruit and vegetable energy intake was also higher than the other listed countries. However, the Turkish population suffers from nutrient deficiencies similar to other countries. The deficiency involves several key vitamins and micronutrients essential for healthy, productive living. Vitamin A deficiency, which also affects iron levels, was found among 17% of pregnant women in Turkey [14]. Tomatoes, green peppers, and cucumbers that provide vitamin A are commonly grown and eaten vegetables in Turkey. Turkey was the world’s third-largest tomato producer in 2019 [15].

Consumption of F&V offsets the lack of knowledge about the importance of folic acid among Turkish consumers, especially during pregnancy [16], and is particularly important for those who do not take folic acid-containing supplements. Increased folic acid is needed to reduce the incidence of neural tube defects in Turkey [17]. Leafy greens, broccoli, and asparagus, among vegetables, and citrus fruit and bananas, among fruits, are valuable sources of folic acid. Citrus and bananas are domestically produced and accessible. Among micronutrients, iron deficiency has been verified among reproductive-age women in a study on anemia in Turkey. Iron deficiency is preventable at a relatively low cost and through targeted education campaigns [18]. Such a campaign could publicize that eating domestically produced fruits (e.g., citrus, banana) and vegetables (e.g., tomatoes, green peppers, and cucumbers,) supplies the body with iron (Table 4).

Hypokalemia is a new phenomenon associated with the inadequate intake of potassium resulting from a decreased level of potassium in the soil where crops grow [19]. A comparison of the potassium content of fruits and vegetables in the United States between 1999 and 2015 showed a decrease in the case of several fruits and vegetables commonly consumed in Turkey. For example, among fruits, lower potassium content was verified in apples, apricots, bananas, figs, kiwis, melons, and watermelons. Among vegetables, decreased potassium content was reported in broccoli, cabbage, cauliflower, onions, peppers, and potatoes, among others. Turkey was the fifth producer of onions in the world in 2019 [20]. Turkish consumers have been more knowledgeable about dietary fiber consumption than consumers in several other countries [21].

A study of Turkish consumers reported that the average vegetable consumption for four age groups was about once per day, except that in the case of the elderly the average was slightly above (1.6 consumption occurrences per day) [21]. Fruit consumption was considerably higher and increased from 1.4 eating occurrences per day for young adults to 2 eating occasions for the elderly. Eggplant, green peppers, and yellow onions are good sources of fiber among the popular vegetables in Turkey (Table 5). Among fruits that are produced in Turkey, citrus, banana, and apples provide even more fiber per serving than most vegetables (Table 4). Bananas and apples also provide small amounts to vitamin B12, not reported sources used to create Table 4. Turkey is among the leading apple-producing countries in the world [20].

The limited consumption of fruits and vegetables as a source of vitamins and key nutrients contrasts with the recent surge in internet searches about vitamins exacerbated by the COVID-19 epidemic [22]. Often the bioavailability of key nutrients in F&V surpasses that of supplements [23], although the issue is complex and requires additional research [24]. In the meantime, Turkish consumers can enjoy the abundant domestic supply of fresh produce and benefit from the established link between its consumption and disease prevention and health maintenance.

Materials and methods

The cross-sectional data used in the study was compiled from Turkey Health Survey (THS) 2019. The survey, conducted by the Turkish Statistical Institute (TSI), has been carried out biennially since 2008 in cooperation with the Statistical Office of the European Union (SOEU). The surveys are conducted in the last quarter of the year (October, November, and December), and the total number of observations is determined as per the survey modules of the SOEU. The total sample size was 9470 household addresses and 8166 families (i.e., 88% participation rate) interviewed. The 1304 non-participating households include either those who accepted the administrative fine or those who were not present during the survey (e.g., due to vacation or visiting relatives). The participants were 18 years old or older. Twelve statistical subdivisions of the country were included in the Nomenclature of Territorial Units for Statistics (NUTS).

Survey information about the respondent characteristics includes gender, age, marital status, education, employment status, body mass index, health insurance, participation in physical exercise, leisure, tobacco and alcohol use, time allocated to walking and other types of physical exercise, occupation, history of depression,Footnote 1 and the number of chronic diseases, among others. A separate set of questions asked about family traits such as the number of children in different age groups (under 7; 7 to 14), number of adults, income categories, and residential location. Several other questions were about education, age, and income.

Our empirical specification in the current study is laid down from the discrete random utility theory [25], in which a household maximizes its random utility function subject to a fixed budget:

$$\mathop {\max }\limits_{q,c} \{ U\left( {Yq,c;h} \right)|p^{\prime}q + c = m\}$$

where \(q=[{q}_{1},{q}_{2}]{^{\prime}}\) is a vector containing quantities of fruits \(({q}_{1})\) and vegetables \(({q}_{2})\) with their corresponding unit prices \(p = [p_{1} ,p_{2} ]^{\prime }\), respectively. While c is a composite commodity for all but these two food products, m reflects the budget, and h is a vector of the household, but mostly characteristics of an individual. While we set \(Y=diag\left({y}_{1},{y}_{2}\right)\) as a diagonal matrix with each (random) binary indicator \({y}_{i}\) indicating potential consumption of \({q}_{i}\) [26]. Assuming that the utility function U(Yq,c;h) is strictly quasi-concave and increases with increasing positive values of Yq and c leads to the notional demand of the two foods. Such notional demands are obtained by solving the utility maximization problem in Eq. (1), in which a vector of optimal quantities demanded of both foods is a function of their prices without non-negativity constraints, household budget, as well as characteristics of individuals [26]. The use of individuals and household characteristics in demand analysis with cross-sectional data dates back to [27].

Our dependent variables are consumption probabilities, as we only observe how often these two foods are consumed daily, not their quantities in a household. If the individual consumes one or more servings of fruit or vegetables daily, we defined them as the latent variables (y1 for fruits and y2 for vegetables). The random-effects bivariate probit model consists of households (i = 1,…,N), two heterogeneity parameters, α1, and α2, defined for the family members (j = 1,…,M) by the exogenous variables x1 and x2, and possibly associated error terms ε1 and ε2, unit variances, correlation coefficient (τ), and two latent variables, y1, and y2, that are normally distributed:

$$\begin{aligned} y_{1,ij}^{*} & = x^{\prime}_{1,ij} \beta_{1} + \alpha_{1,i} + \varepsilon_{1,ij} \\ y_{2,ij}^{*} & = x^{\prime}_{2,ij} \beta_{2} + \alpha_{2,i} + \varepsilon_{2,ij} ,\,\,\,\,\,i = 1,...,N\,\,and\,\,j = 1,...,M \\ \end{aligned}$$


$$\begin{aligned} \varepsilon_{it} & = \left( \begin{gathered} \varepsilon_{1,ij} \hfill \\ \varepsilon_{2,ij} \hfill \\ \end{gathered} \right) \approx i.i.d.N\left[ {\left( \begin{gathered} 0 \hfill \\ 0 \hfill \\ \end{gathered} \right);\left( \begin{gathered} 1\,\,\,\,\tau \hfill \\ \tau \,\,\,\,1 \hfill \\ \end{gathered} \right)} \right] \\ \alpha_{i} & = \left( \begin{gathered} \alpha_{1,i} \hfill \\ \alpha_{2,i} \hfill \\ \end{gathered} \right) \approx i.i.d.N\left[ {\left( \begin{gathered} 0 \hfill \\ 0 \hfill \\ \end{gathered} \right);\left[ \begin{gathered} \sigma_{1}^{2} \,\,\,\,\,\,\,\,\,\,\,\rho \sigma_{1} \sigma_{2} \hfill \\ \rho \sigma_{1} \sigma_{2} \,\,\,\,\,\,\,\,\sigma_{2}^{2} \,\, \hfill \\ \end{gathered} \right]} \right] \\ \end{aligned}$$

The coefficient τ is the degree of relationship between two latent variables (y1 and y1), coefficients σi denote the associated standard deviations of the two heterogeneous coefficients (αi), ρ represents the correlation coefficient between the two heterogeneous coefficients, while βi denotes the coefficients of regressor variables affecting the latent variable. In our example, the fixed-effects model (in its most basic form) may control for unmeasured variables that are constant between individuals but vary between families, by explicitly including a separate cutoff term (αi) for each family in each regression equation above. However, the reason for using the random-effects model instead of the fixed-effects model is that, in addition to the confounding variables originating from the family, we cannot fully predict how all confounding factors specific to individuals (such as the individual’s talent, ability, digestion, appetite, other health factors, etc.) will react from family to family and the probability of fruit and vegetable consumption. The same is true for mixed-effects. In addition, the concern that many of our control variables might have collinear relationships with individual unmeasured or confounding variables led us to the random-effects model. Our random-effects model is also equivalent to the random parameter model, with only the constant term being random. Although the triggers (regressors) of each latent variable could be different, they are assumed to be equal in this study (x1 = x2). The observed model is:

$$\begin{aligned} y_{1,ij} & = 1\,\,(y_{1,ij}^{*} > 0) \\ y_{2,ij} & = 1\,\,(y_{2,ij}^{*} > 0) \\ \end{aligned}$$

where y1 is 1 if an individual consumes fruits one or more servings (portions) per day, 0 otherwise. Similarly, y2 is 1 if an individual consumes vegetables one or more servings per day, and 0 otherwise. In other words, it reflects that the daily intake of at least one or more servings of fruits and vegetables is related to the characteristics of the individual and the family.

The classical transformation of the observed variables and the corresponding conditional composite likelihood functions are defined, respectively, asFootnote 2:

$$\begin{aligned} q_{1,ij} & = 2y_{1,ij} - 1 \\ q_{2,ij} & = 2y_{2,ij} - 1 \\ \ell_{i} (y_{i} |x_{ij} ,\beta ,\sigma_{i}^{2} ,\rho ) & = \int\limits_{ - \infty }^{ + \infty } {\int\limits_{ - \infty }^{ + \infty } {f_{i} (y_{i} |x_{i} ,\alpha_{i} ,\beta ,\tau )} } *g\left( {\alpha_{i} |\sigma_{1}^{2} ,\sigma_{1}^{2} ,\rho } \right)d\alpha_{1,i} d\alpha_{2,i} \\ \end{aligned}$$

While some studies took advantage of simulations to optimize the above likelihood function [28], this study employs the Gauss-Hermite quadratic technique [29]. The random-effects bivariate probit model is essentially identical to the bivariate random parameter probit model when the constant coefficient is assumed to be random.Footnote 3

Probability values are derived differently when using a random-effects bivariate probit model. For example, besides examining the probability of consuming fruits and vegetables together (e.g., the joint decision), it is possible to obtain the conditional consumption probability of fruits (vegetables) given the consumption probability of vegetables (fruits). The expected values of marginal (e.g., individual), joint, and conditional probabilities of F&V intake are:

$$\begin{aligned} \Pr \left[ {y_{k,ij} = 1|x_{k,ij} } \right] & = \Phi \left( {x^{\prime}_{k,ij} \hat{w}_{k} } \right),\,\,k = 1,2 \\ \Pr \left[ {y_{1,ij} = 1,y_{2,ij} = 1|x_{k,ij} } \right] & = \Phi_{2} \left( {x^{\prime}_{1,ij} \hat{w}_{1} ,x^{\prime}_{2,ij} \hat{w}_{2} ;\tau } \right) \\ \Pr \left[ {y_{k,ij} = 1|y_{m,ij} = 1} \right] & = \frac{{\Pr \left[ {y_{1,ij} = 1,y_{2,ij} = 1|x_{k,ij} } \right]}}{{\Pr \left[ {y_{m,ij} = 1|x_{k,ij} } \right]}},\,\,k \ne m \\ \end{aligned}$$

where \(\hat{w}_{k} = \frac{{\hat{\beta }_{k} }}{{\sqrt {1 + \hat{\sigma }_{k}^{2} } }}\) and \(\Phi \,\) and Φ2 are univariate and bivariate cumulative density functions, respectively. Differentiation of each corresponding probability value concerning any regressor is the marginal (unitary) effect in the case of a continuous variable and a unitary effect if the explanatory variable is binary. The standard deviations of marginal effects were obtained using the delta method.

Descriptive and preliminary results

Table 1 shows the descriptive statistics of dependent and confounding variables. The consumption rates of one or more servings per day were 47% for fruits and 57% for vegetables. Daily fruit and vegetable consumption is common, with nearly one in two individuals eating one or more servings per day. Considering that in Turkish culture, meals, especially dinners, are served in the presence of the whole family with a vegetable salad served first, a high vegetable intake is expected. Fruit intake, on the other hand, varies according to the individual’s preferences, taste, and psychological state at the time it is offered. Approximately 40% of individuals consume both foods together (Table 3). About 70% of those who consume one or more servings of vegetables a day get fruit, while about 85% of those who consume one or more servings of fruit a day consume vegetables. Among those who consume vegetables, the habit of consuming fruit is weaker than the habit of consuming vegetables among those who consume fruits. In addition, the intake of both foods is at a very low rate. The fact that those who consume fruits are more conscious in terms of health also increases the odds of consuming vegetables. However, to boost fruit intake among those who consume vegetables and increase consumption of both foods together as a part of daily life in individuals, there is a need for intervention initiatives in the country.

Table 1 Means and VIF scores of dependent and explanatory variables

The individuals participating in the survey (Table 1) included 45.5% male, 54.5% female, 68.6% married, about 18% university graduates, 16% obese as well as 6% extremely obese, 46% smokers, and 5% consuming alcohol. Further, 3% participated in weight lifting at least four times a week, and 16% engaged in sports for at least one hour a day. Descriptive statistics of other variables are given in Table 1. The variance inflation factor (VIF) was calculated to verify the absence of multicollinearity among the independent variables, where a VIF value of less than 5 indicates that there is no significant collinear relationship between our regressors in this study.

Table 2 shows the estimators of the maximum likelihood function. The correlation coefficient (τ) between the two consumption decisions was positive and statistically significant (Table 2). While vegetables are mostly used within meals and served in the form of salads, it is a classic Turkish tradition to serve fruit immediately after a meal or a few hours after dinner. Therefore, when uncontrollable factors affect the vegetable intake decision, they also likely affect the probability of fruit intake. The correlation coefficient (ρ) between heterogeneities in the consumption probabilities of F&V was negative and statistically significant. The heterogeneity, for example, in the consumption probability of fruits has an inverse effect on the heterogeneity of vegetable consumption. Interestingly, while the two food intake patterns are positively affected by uncontrollable variables among families, they are negatively affected among family members. This is a very important result because the average attitudes of family members may differ from those of the average family. The Likelihood Ratio (LR) test (LR = −4051.87, degrees of freedom (df) = 2, p < 0.000) rejected the hypothesis that both correlation coefficients were simultaneously zero, indicating that the bivariate random-effects probit model was superior to the two discrete (e.g., separate) binary random-effects probit models. Also, using the Lagrangian Multiplier (LM) test (LM = 3923.08, df = 4, p < 0.000), the goodness of fit of the proposed bivariate-random effects probit model outperforms the bivariate probit model,Footnote 4 which ignores all intra-family heterogeneities. Using the LR test, we also determined that all regressors used in the random-effects bivariate probit model are the main source of variation in the probability of simultaneous consumption decision (LR = 80.00, df = 80, p < 0.000). The following discussion will focus on the proposed model. Also, the overlap between the actual values and the estimated values in all calculated probabilities is strong evidence of the suitability of the method used in the data analysis (Table 3).

Table 2 Maximum likelihood estimates from the panel random-effects bivariate probit model
Table 3 Marginal effects of explanatory variables on eating one or more servings of fruits and vegetables in Turkey using the bivariate random-effects probit model

Discussion and policy implications

The marginal effects derived under different probabilities show that individual socio-demographic, economic, and habitual factors significantly influence their F&V intake likelihood (Table 3). The marginal effects vary greatly as the calculated probability differs. First, we will focus on statistically significant dummy variables. Apart from fruit intake among individuals who consumed one or more servings of vegetables per day, in all other probabilities men consumed significantly fewer fruits and vegetables than women. Meanwhile, the probability of vegetable intake in the population (i.e. marginal probability, Prob(yk = 1|xik), where k = 1,2) is significantly lower among males than females (11.83%), and the probability of vegetable intake given the probability of one or more servings of fruit per day (7.42%) (i.e., the conditional probability). The joint probability of both foods (7.91%) is also lower among males in Turkey. The results show that the gender distinction in fruit and vegetable intake is obvious, as well as heterogeneity between the segments of the population. Men consume much fewer vegetables than women, which is followed by the probability of consuming two foods together and consuming vegetables among those who consume fruit. The result is consistent with the study of young American adults where men consumed approximately 8.5% less F&V than women [30]. Similarly, Malaysian men were 4.8% and 9.3%, respectively, less likely to consume fruits and vegetables per day than women [31]. A recent study confirmed that female-headed households consume more F&V [32]. In Turkey, being a male household head reduces the probability of monthly consumption of fresh or frozen vegetables by 2.4% and lowers the expenditures by 5.9 Turkish Lira (TL) in comparison to female household heads [12]. In general, the factors impacting health are expected to affect women more than men. For example, women who watch health television programs are likely to be more aware of the health benefits of F&V than men. Therefore, such findings may be attributed to men’s poorer nutritional knowledge and men being less likely to know the benefits associated with consuming the recommended intake. By expanding the scope and duration of health-related topics on television programs throughout the day, local and regional public health-related interventions may influence F&V consumption in the family. In a country like Turkey with a male-dominated workforce, policymakers at the national and local levels can increase male F&V consumption probability by urging employers to include a daily F&V option in lunches or promote F&V eating by offering them free or discounted prices in the workplace. Additionally, employers can ensure that vending machines include F&V options.

In all explored consumption likelihoods (i.e., different population segments), married individuals are more likely to consume fruits and vegetables than individuals with different marital statuses, and these individuals consume relatively more vegetables. A study in Uganda found that married participants were more likely to consume five or more servings of fruit and/or vegetables in a typical week [33]. Such a level of association has been supported by other studies [34]. Marriage and friendship form the basis of a regular meal pattern that includes fruits and vegetables and provides an opportunity for social interaction for greater food consumption. Such behaviors are explained by the theoretical model of social integration and social control associated with married couples’ efforts to control each other’s health behaviors [35]. In an average household of married couples, a spouse, especially a woman, keeps such foods at home to prepare a rich meal for themselves and their family. In empirical studies conducted in the United Kingdom [36], Canada [37], and Turkey [12], married households have been found to consume more F&V than households not married.

Education creates a set of social, behavioral, and psychological foundations that promote healthy lifestyles for individuals. A non-linear positive relationship was found between an individual’s education and F&V intake. The better educated choose to eat healthy produce, including F&V [12, 38, 39]. In all probabilities examined, secondary and university graduates are more likely to eat F&V. The effects of education on the probability of fruit intake among vegetable consumers and the probability of vegetable intake among fruit eaters were relatively lower than the other probabilistic effects from all education strata. Meanwhile, the probability of fruit intake among those who consume vegetables was found to be higher than the probability of intake of vegetables among those who consumed fruit. The fact that individuals who make vegetable consumption a part of their daily life pay the same attention to their daily fruit intake (along with their increasing education level) is proof that they may well be aware of the benefits of nutrition derived from fruits. We also found that the probability of consuming vegetables among fruit eaters was the lowest among all other probabilities as the individual’s education level increased. From the least educated to the most educated, it would be the right decision to switch to public programs that include the disease-fighting aspects of vegetables and their nutritional contribution to health.

The lesser educated, often unemployed or working lower-paying jobs and likely residing in suburbs face a greater risk of food insecurity or food budget constraints. In such cases, subsidies for F&V may be beneficial. F&V consumption should be encouraged for children two years old and older to assure they eat the recommended amount for reducing the risk of chronic diseases resulting from conditional such as vitamin deficiency. To this extent, the ministries of health and education can cooperate by adding topics addressing healthy nutrition to the school curricula and help mitigate household health expenditures while expanding the national healthcare system. Such coordinated efforts raise awareness and subsequently increase F&V consumption.

Compared with students and military service individuals, the probability of most F&V consumption increases among wage earners, though insignificant effects. Interestingly, among the employed, the probability of fruit consumption decreases for those who make vegetable consumption a part of their daily life, while the probability of consumption of vegetables increases significantly among fruit eaters. On the other hand, among job seekers, the probability of fruit consumption and the consumption of both foods together decrease. Interestingly, among the different layers of the explored population, fruit and vegetable consumption habits increase among the retired, while it decreases among homemakers, but in the case of the latter group, the results are statistically insignificant. Results show also that the employed may have decided that plant-based diets have become popular as a way to mitigate the diet’s environmental footprint and improve human health and animal welfare. The current study results coincide with Heo et al. [40] that employees had a low frequency of F&V consumption. In a study conducted in Ghana, the self-employed household head or working in the public sector had a reduced fresh F&V expenditure [41]. Terin et al. [12] reported that the consumption probabilities of fresh and frozen F&V were 2.5% and 1.78%, respectively, lower than that in the households of unemployed heads in Turkey. Although the current study does not identify the underlying reasons for such a consumption pattern, providing F&V as a snack in the workplace may increase fruit and vegetable intake for employees.

Obese and morbidly obese individuals tend to have higher F&V uptake in the distinct segments of the examined population. However, fruit consumption intake is statistically significant among obese individuals who make a habit of daily vegetable consumption. Meanwhile, the morbidly obese are more likely to have higher fruit intakes and are almost twice as likely to consume vegetables and both foods as compared to obese people. Tohill et al. [42] reported the F&V intake of individuals with a normal BMI was lower than the overweight or obese. Consuming fruit more than four times a week may have increased the risk of obesity [43]. In an Australian study, obese and overweight women were more likely to be in the highest quartile for combined fruit and vegetable intake and meet the “2 and 5” target or consume five or more servings of fruit and vegetables per day [44]. The energy supply due to F&V consumption of those with high BMI can be controlled by ensuring their participation in appropriate, regular physical activities along with dietary programs.

This study found that individuals with general health insurance are likely to increase their fruit intake, while those with private health insurance are more likely to consume vegetables both in the population (i.e., the marginal probability) and among fruit-consuming vegetable eaters (i.e., the conditional probability) but their effects found insignificant. F&V bill programs (also called bill of material programs, BoM,) can stimulate increased consumption of F&V among those living in poverty [45]. This simulation model typically allows a healthcare provider to “prescribe” eating fresh F&V to patients with diet-related chronic diseases while they receive nutritional guidance in a clinical setting. Such F&V bill programs have stimulated healthy eating habits [7, 45]. A study using private health insurance as a confounding factor in F&V intake in Australia showed a positive result, given that those with private health insurance tend to be highly educated and high-income individuals [46].

On the other hand, individuals who have the habit of cycling for at least ten minutes a day are more likely to consume one or more servings of fruit a day, consume more among those who eat both foods together, and are more likely to consume fruits among those who consume vegetables. Cyclists who experience energy loss due to exercise are expected to make up for the such loss by eating more fruit. Parallel to such a finding, compared with those with no or less than ten minutes of walking a day, the more time spent walking, the more likely individuals are to consume more fruit, vegetables, or both, and consume vegetables among daily fruit eaters. However, such an effect is more pronounced among those who make vegetable consumption a part of their daily life, almost twice as much as those who consume fruit.

A similar result exists among workaholics, where those who rest less than four hours a day are more likely to consume fruit and vegetables, while those who make their daily fruit intake a lifestyle increase their vegetable intake. This behavior increases the risk of unbalanced nutrition and severe health problems. Since those individuals tend to be highly educated, enjoy a relatively high income, and have regular medical checkups, they may be receptive to messages encouraging eating fruits and vegetables as snacks. Exercise and high consumption of F&V have a significant positive impact on the health of individuals [1].

Policies encouraging individuals to participate in various forms of physical exercise result in a healthier society and increase F&V consumption. Additionally, the creation of educational curricula that support the upbringing of healthier individuals (e.g., including courses that motivate individuals into participating in regular sports activities in school) would encourage individuals both to regularly exercise and consume healthier foods. On the other hand, as physical labor from moderate to vigorous intensity increases, individuals can meet their increased energy and other nutritional needs with fruit and vegetable intake. Heavy manual labor workers are relatively more likely to consume fruits and vegetables than lower manual labor workers, and they are also likely to consume more vegetables than fruits. Those who work jobs that are more physically demanding may replace some of their lost energy and nutritional needs with fruit and vegetables.

Individuals in the two extreme income quintiles, especially very poor families, have a very low probability of fruit and vegetable consumption in all segments of the population, while very wealthy families have a higher likelihood of fruit and vegetable intake compared to families in the middle-income bracket. Considering that very poor families spend almost all of their income on food, it is evident that they do not have access to sufficient nutritious food and are vulnerable to inadequate food consumption. The state could provide food and other services for low-income pregnant women and children up to a specific nutritional-risk age by developing a supplemental food program for women, infants, and children. On the other hand, when compared to families in the two extreme income brackets, we found that those with higher incomes made more fruits and vegetables a part of their daily lives, their vegetable intake outweighing fruit consumption. There is ample evidence in the literature confirming such a correlating relationship [11, 12].

The worldwide increase in family income and public health education in the last decade has contributed to the F&V consumption increase [38]. Affordability has been a reason for low-income households’ limited F&V consumption [47]. Yen et al. [31] reported that individuals in the lowest income group tend to consume fewer vegetables compared to those in the highest income category (e.g. individuals in the highest income group in England consumed more F&V [48]. Several other studies stressed that F&V consumption increased with income [12, 41]. In our study, as income increases, individuals prefer vegetable portions to fruit portions. Considering the nutritional values provided by different ranges of seasonal vegetables, it is expected that individuals will tend to prefer relatively more expensive vegetables to lower-priced vegetables as income increases.

Finally, in the examination of the link between regions (the last of the binary variables), where a person resided in the country and F&V consumption shows a positive relationship. As compared to individuals residing in the East Anatolia region, residents of Marmara consumed more F&V, whereas individuals residing in the Aegean, Mediterranean, Black Sea, and Southeastern Anatolia regions consumed fewer F&V. The result for Central Anatolians was insignificant. Compared to the reference region, individuals residing in the Black Sea region have very low fruit and vegetable intakes, while in regions with a confirmed statistically significant effect, the probability of individuals consuming fruits and vegetables is almost identical. In regions with low F&V consumption, encouraging more F&V consumption and the individual willingness to change eating habits is of great importance. Particularly, individuals should be motivated and encouraged to participate actively in lifestyle intervention programs (i.e., the Healthy Portions program) and to learn about the benefits of F&V through government-run public service advertisements in the media. Support programs encouraging rural populations can emphasize increased F&V consumption of raised produce.

Focusing on continuous variables, the increasing age of individuals increases the probability of F&V consumption in all segments of the population. However, the significant increase is in fruit intake, while the least uptake is in vegetable intake among fruit consumers. A plentiful intake of fruits and vegetables is important at any age, as such foods are the top source of fiber and other nutrients essential for good health. The intake of such nutrients should be made a habit early in life, as many of these nutrients help prevent or mitigate the risk of disease as it progresses over the years. Studies in Italy [49], the United Kingdom [50], and Malaysia [31] reported that an increase in the age of the household head increased fruit consumption. A similar relationship between the age of individuals and F&V consumption probability was obtained in studies conducted in Turkey [12] and South Korea [31]. Health literacy campaigns (e.g., creating public service ads, messages, advertisements, and social media content with private and public collaboration) regarding disease prevention and health maintenance by eating fruits and vegetables that are easy to understand can be implemented through mass media. Providing additional support for fruit and vegetable consumption among the young and middle-aged should be considered, such as the creation of mobile phone apps (e.g., businesses can help increase F&V intake by digitizing their messages regarding the availability and wholesomeness of F&V).

As time spent on playing sports increases, F&V intakes increase. F&V intake is likely to be higher because such activities require more energy. Meanwhile, tobacco and alcohol intake negatively affect both food intakes, but the effect of tobacco intake is both much more pronounced and far more potent than alcohol. Fruit and vegetable intakes are also affected by the appetite suppression properties of those two habits. It is particularly noteworthy that the probability of eating fruit by smokers is approximately 2.5 times lower than that of vegetables, suggesting that smoking may replace fruit snacking. Furthermore, the probability of reduced fruit consumption by smokers is higher among all considered factors. Yen et al. [31] found that the probability of fruit consumption by smokers was 6.72% lower than that of non-smokers. Daily smoking rates were less in groups with high F&V consumption compared to others [52]. A study of Canadian adults reported that those who never smoked, former smokers, and the elderly consumed more F&V than other groups [53].

The presence of children between 7 and 14 years old in the family significantly, although by a relatively small percent, lowers the probability of both fruit and vegetable consumption. Contrarilty, the increase in the number of adults in the household is likely to increase the intake of one or more servings of fruit per day but their impacts are insignificant. Adults with children may eat fewer fruits and vegetables as they are more likely to have to accommodate children who want to eat salty snacks, pizza, cheese, beef, ice cream, cake or cookies, high-fat foods such as sausage or processed meats, and nut-butters. F&V consumption can be increased in those households through creative business efforts. For example, parents’ preferences can be shaped by social media and electronic messages (e.g., Short Message Service (SMS), e-mail) containing information about F&V benefits. On the other hand, an increase in the number of adults in the family significantly increases the probability of vegetable intake and triggers the odds of consuming vegetables among those who consume fruit daily.


The current study investigated the relationship between F&V intake decisions of family members eating one or more servings of fruits and vegetables a day and individual and family characteristics. This approach fills the gap by accounting for intra-family heterogeneity in cross-sectional studies of F&V consumption. Additionally, F&V consumption likelihoods often depend on simultaneous decisions. The fact that both the intra-family heterogeneity parameter and the relationship between the two eating decisions have significant and inverse effects indicates that the attitudes and decisions of each family can differ greatly. It is more reasonable to evaluate the results obtained the reflection of different characteristics of family members on the decision of distinct consumption likelihoods, rather than an extraordinary expectation. Studies that ignore the multilevel data structure and simultaneous fruit and vegetable consumption decisions are not expected to reach such counterfactual inferences.

Being male, regardless of job type, having an income, residing in Aegean, Mediterranean, Black Sea, and Southeastern Anatolia regions, smoking, alcohol consumption, and presence of children under age 7 and ages 7 to 14 in a household were factors associated with less likely F&V consumption. However, a positive influence on F&V intake decision was associated with marital status, education, BMI, general and private health insurance coverage, bicycling, walking, resting, performing physically demanding jobs, high income, residence in Marmara region, age, time spent on sports, and number of adults in the household. Elderly individuals consume more fruit, and fruit consumption is higher among those who also consume vegetables. The results show that increased fruit consumption is associated with more frequent vegetable consumption. Similarly, as an individual has more education, fruit intake among those who consume vegetables is higher than among those who consume only fruit. Considering the relationship between high education level and high income, it is clear that the priority of individuals in this group is high fruit intake despite relatively high prices. To increase F&V consumption and develop sustained F&V consumption, policies raising awareness of F&V benefits within the family are desirable in addition to efforts at the society level. Enhancing F&V consumption could lead to healthier current and future generations and possible health cost reduction.

The results suggest segmenting individuals by their socioeconomic, demographic, lifestyle, and income factors to develop and implement separate, segment-specific policies rather than attempting a one-size-fits-all policy. For example, there is a need to formulate and implement policy to increase F&V consumption among low- and middle-income households in Turkey. The elderly will benefit from information about the importance of vitamins in F&V in improving physical and mental health. State-run television channels can promote “healthy lifestyles” characterized by F&V consumption and affect housewives which then encourage family members to eat more F&V. Information about F&V can be conveyed to the illiterate population visually or verbally through social media via cell phones. Family-oriented programs should exploit the multi-faceted benefits of F&V by emphasizing the functional ingredients of fruits and vegetables and targeting specific segments of the population such as smokers. A separate path promoting F&V consumption can be fruit and vegetable snacks accessible for workers at places of employment, stressing the link between health maintenance and productivity.

Children need to be a target of programs tailored to various age groups and possibly genders. Parents need to be educated about the importance of F&V intake by children beginning from an early age. Public service announcements using age-appropriate visual, printed, and broadcast media to raise healthier future generations as well as to prevent and reduce serious health problems that may arise later in life will help reach such goals. Hiring dietitians in family healthcare centers to offer vegetable and fruit diet guidelines should be considered to mitigate diseases that may result from insufficient F&V consumption. Efforts promoting F&V eating could be supplemented by the construction of sports fields, walking areas, and bicycle paths to encourage physical activity in residential neighborhoods leading to a healthier quality of life. Innovative interventions aiming at increasing F&V uptake should be a continuing focus in Turkey. Meanwhile, the differences in F&V intake across regions suggest that interventions to reach the recommended level of consumption should involve programs recognizing the regional specificity of low F&V intake.

The study has several limitations. The use of cross-sectional data limits the causal analysis of the decision to eat F&V. Multi-year panel data, if available, could permit a deeper analysis and provide knowledge for additional detailed actions encouraging F&V eating. A richer dataset will also allow the application of a different modeling framework overcoming the limitations of cross-sectional data use in the estimation of a latent dependent variable model. Also, the results obtained for Turkey, an emerging economy with a fast-growing population, may not be readily transferrable to every country. However, the results of the current study provide insights that can serve as a benchmark for assessing similar topics in other countries. Finally, the focus on the consumption of fruits and vegetables as the source of eliminating or reducing any vitamin and micronutrient deficiency in the current study implicitly recognizes that such deficiencies can have also other reasons such as the unaccounted household behavior including cooking and preparation methods of frhes produce, household storage methods of, or purchase of prepared or processed fruits and vegetables. Additional data could be collected in the future to examine such causes.

Availability of data and materials

Datasets used in the current study are available from the corresponding author upon reasonable request.


  1. Against the possibility of an endogeneity problem caused by depression in the system, the depression variable was run against a number of variables and then its residuals were calculated and used together in the probability of fruit and vegetable consumption. The insignificance of the variable confirms that there is no such endogeneity problem in the system.

  2. Since the data are on a national basis, representative weights were used in parameter estimation, thanks to a referee’s warning.

  3. All analyses were performed using NLOGIT-6 statistics software.

  4. Bivariate probit model and marginal effect results are given in the appendix of Tables 6 and 7, respectively.



Turkish Health Survey


Turkish Statistical Institute


Fruits and vegetable


Non-communicable diseases


Low fruit and vegetable intake


Blood pressure


Statistical Office of the European Union


Nomenclature of Territorial Units for Statistics


Variance inflation factor


Likelihood Ratio


Lagrangian Multiplier


Body mass index


Short Message Service


  1. WHO. The world health report 2003: reducing risks, promoting healthy life: overview. World Health Organization. 2003. (accessed May 2022).

  2. FAO. Promotion of Fruit and Vegetables for Health. Report of the Pacific Regional Workshop, The Food and Agriculture Organization (FAO), 94p, Rome, Italy. 2015. (accessed May 2022).

  3. Bertoia ML, Mukamal KJ, Cahill LE, Hou T, Ludwig DS, Mozaffarian D, et al. Changes in intake of fruits and vegetables and weight change in United States men and women followed for up to 24 years: analysis from three prospective cohort studies. PLoS Med. 2015;12(9):e1001878.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Poirier AE, Ruan Y, Hebert LA, Grevers X, Walter SD, Villeneuve P, et al. JEstimates of the current and future burden of cancer attributable to low fruit and vegetable consumption in Canada. Prev Med. 2019;122:20–30.

    Article  PubMed  Google Scholar 

  5. Micha R, Shulkin ML, Penalvo JL, Khatibzadeh S, Singh GM, Rao M, et al. Etiologic effects and optimal intakes of foods and nutrients for risk of cardiovascular diseases and diabetes: systematic reviews and meta-analyses from the Nutrition and Chronic Diseases Expert Group (NutriCoDE). PloS One. 2017;12(4):e0175149.

    Article  PubMed  PubMed Central  Google Scholar 

  6. Cavanagh M, Jurkowski J, Bozlak C, Hastings J, Klein A. Veggie Rx: an outcome evaluation of a healthy food incentive programme. Public Health Nutr. 2017;20(14):2636–41.

    Article  PubMed  Google Scholar 

  7. Richie KA. The Impact of A Fruit and Vegetable Farmers' market Voucher Prescription Program on A Low-Income Rural Population. Theses and Dissertations-Dietetics and Human Nutrition 65. 2019.

  8. Glabska D, Guzek D, Groele B, Gutkowska K. Fruit and vegetable intake and mental health in adults: a systematic review. Nutrients. 2020;12(1):115.

    Article  PubMed  PubMed Central  Google Scholar 

  9. Conner TS, Brookie KL, Carr AC, Mainvil LA, Vissers MC. Let them eat fruit! The effect of fruit and vegetable consumption on psychological well-being in young adults: a randomized controlled trial. PLoS ONE. 2017;12(2):e0171206.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Aune D, Giovannucci E, Boffetta P, Fadnes LT, Keum N, Norat T, et al. Fruit and vegetable intake and the risk of cardiovascular disease, total cancer and all-cause mortality—a systematic review and dose-response meta-analysis of prospective studies. Int J Epidemiol. 2017;46(3):1029–56.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Chai W, Fan JX, Wen M. Association of individual and neighborhood factors with home food availability: evidence from the national health and nutrition examination survey. J Acad Nutr Diet. 2018;118(5):815–23.

    Article  PubMed  PubMed Central  Google Scholar 

  12. Terin M, Birinci A, Bilgiç A, Urak F. Determinants of fresh and frozen fruit and vegetable expenditures in turkish households: a bivariate tobit model approach. J Food Prod Mark. 2018;25(2):137–58.

    Article  Google Scholar 

  13. FAO. The Food and Agriculture Organization (FAO). 2019. (accessed April 2022).

  14. Meram I, Bozkurt AI, Kilincer S, Ozcirpici B, Ozgur S. Vitamin A and beta-carotene levels during pregnancy in Gaziantep, Turkey. Acta Med (Hradec Kralove). 2004;47(3):189–93.

    Article  CAS  Google Scholar 

  15. NationMaster: Global Industry Market Sizing. 2022. (accessed May 2022).

  16. Baykan Z, Öztürk A, Poyrazoğlu S, Gün İ. Awareness, knowledge, and use of folic acid among women: a study from Turkey. Arch Gynecol Obstet. 2011;283(6):1249–53.

    Article  CAS  PubMed  Google Scholar 

  17. Tunçbilek E. Türkiye’deki yüksek nöral tüp defekti sıklığı ve önlemek için yapılabilecekler. Çocuk sağlığı ve hastalıkları dergisi. 2004;47(2):79–84.

    Google Scholar 

  18. Saydam BK, Genc RE, Sarac F, Turfan EC. Prevalence of anemia and related factors among women in Turke. Pak J Med Sci. 2017;33(2):433.

    PubMed  PubMed Central  Google Scholar 

  19. Sun H, Weaver CM. Rise in potassium deficiency in the us population linked to agriculture practices and dietary potassium deficits. J Agric Food Chem. 2020;68(40):11121–7.

    Article  CAS  PubMed  Google Scholar 

  20. Maps and Statistics of the World and Regions. 2022. (accessed March 2022).

  21. Yalçin A, Keleş H, Kahraman T, Bozkurt M, Aydin H. Protective effects of ellagic acid against chemotherapy-induced hepatotoxicity. Duzce Med J. 2020;22(2):124–30.

    Article  Google Scholar 

  22. Çimke S, Gürkan DY. Determination of interest in vitamin use during COVID-19 pandemic using Google trends data: infodemiology study. Nutrition. 2021;85:111138.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Yang P, Wang H, Li L, Ma Y. Determination and evaluation of bioavailability of vitamins from different multivitamin supplements using a pig model. Agriculture. 2021;11(5):418.

    Article  CAS  Google Scholar 

  24. Pressman P, Clemens RA, Hayes AW. Bioavailability of micronutrients obtained from supplements and food: a survey and case study of the polyphenols. Toxicol Res Appl. 2017;1:2397847317696366.

    Google Scholar 

  25. Pudney S. Modelling individual choice: the econometrics of corners, kinks, and holes. Cambridge: Blackwell Publishers. 1989. Available At:

  26. Yen ST. A multivariate sample–selection model: estimating cigarette and alcohol demands with zero observations. Am J Agric Econ. 2005;87(2):453–66.

    Article  Google Scholar 

  27. Prais SJ, Houthakker HS. The analysis of family budgets, with an application to two British surveys conducted in 1937–39 and their detailed results. New York: Cambridge University Press; 1955.

    Google Scholar 

  28. Kano S. Like Husband, Like Wife: A Bivariate Dynamic Probit Analysis of Spousal Obesities, mimeo, Osaka Prefecture University. 2008.

  29. Mulkay B. Bivariate Probit Estimation for Panel Data: a two-step Gauss–Hermite Quadrature Approach with an application to product and process innovations for France. UniversitÈ de Montpellier—MRE. Working Paper. 2017.

  30. Powell LM, Zhao Z, Wang Y. Food prices and fruit and vegetable consumption among young American adults. Health Place. 2009;15(4):1064–70.

    Article  PubMed  Google Scholar 

  31. Yen ST, Tan AKG, Feisul MI. Consumption of fruits and vegetables in Malaysia: profiling the daily and nondaily consumers. Asia Pac J Public Health. 2015;27(2):635–2650.

    Article  Google Scholar 

  32. Choudhury S, Shankar B, Aleksandrowicz L, Tak M, Green R, Harris F, et al. What underlies inadequate and unequal fruit and vegetable consumption in India? An exploratory amalysis. Global Food Secur. 2020;24:100332.

    Article  Google Scholar 

  33. Kabwama SN, Bahendeka SK, Wesonga R, Mutungi G, Guwatudde D. Low consumption of fruits and vegetables among adults in Uganda: findings from a countrywide cross-sectional survey. Archives of Public Health. 2019;77(1):1–8.

    Article  Google Scholar 

  34. Vinther JL, Conklin AI, Wareham NJ, Monsivais P. Marital transitions and associated changes in fruit and vegetable intake: findings from the population-based prospective EPIC-Norfolk cohort, UK. Soc Sci Med. 2016;157:120–6.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Umberson D. Gender, marital status and the social control of health behavior. Soc Sci Med. 1992;34(8):907–17.

    Article  CAS  PubMed  Google Scholar 

  36. Pollard J, Greenwood D, Kirk S, Cade J. Lifestyle factors affecting fruit and vegetable consumption in the UK Women’s Cohort Study. Appetite. 2001;37(1):71–9.

    Article  CAS  PubMed  Google Scholar 

  37. Riediger ND, Moghadasian MH. Patterns of fruit and vegetable consumption and the influence of sex, age, and socio-demographics factors among Canadian elderly. J Am Coll Nutr. 2008;27(2):306–13.

    Article  PubMed  Google Scholar 

  38. Chong KH, Lee ST, Ng SA, Khouw I, Poh BK. Fruit and vegetable intake patterns and their associations with sociodemographic characteristics, anthropometric status and nutrient intake profiles among Malaysian children aged 1–6 years. Nutrients. 2017;9(8):723.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Craveiro D, Marques S, Zverinova I, Máca V, Ščasný M, Chiabai A, et al. Explaining inequalities in fruit and vegetable intake in Europe: the role of capabilities, opportunities and motivations. Appetite. 2021;165:105283.

    Article  PubMed  Google Scholar 

  40. Heo M, Kim RS, Wylie-Rosett J, Allison DB, Heymsfield SB, Faith MS. Inverse association between fruit and vegetable intake and BMI even after controlling for demographic, socioeconomic and lifestyle factors. Obes Facts. 2011;4(6):449–55.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Meng H, Hu W, Chen Z, Shen Y. Fruit and vegetable intake and prostate cancer risk: a meta-analysis. Asia Pac J Clin Oncol. 2014;10(2):133–40.

    Article  PubMed  Google Scholar 

  42. Tohill BC, Seymour J, Serdula M, Kettel-Khan L, Rolls BJ. What epidemiologic studies tell us about the relationship between fruit and vegetable consumption and body weight. Nutr Rev. 2004;62(10):365–74.

    Article  PubMed  Google Scholar 

  43. Mohajeri M, Hoojeghani S, Izadi A, Ghahremanzadeh M, Pourfarzi F, Nemati A, et al. Food choice motivations among Ardabil-Iran adults? Nutr Food Sci. 2019;50(4):641–52.

    Article  Google Scholar 

  44. Charlton K, Kowal P, Soriano MM, Williams S, Banks E, Vo K, et al. Fruit and vegetable intake and body mass index in a large sample of middle-aged Australian men and women. Nutrients. 2014;6(6):2305–19.

    Article  PubMed  PubMed Central  Google Scholar 

  45. Marcinkevage J, Auvinen A, Nambuthiri S. Washington state’s fruit and vegetable prescription program: improving affordability of healthy foods for low-income patients. Prev Chronic Dis. 2019;16:180617.

    Article  Google Scholar 

  46. Lim S, Beauchamp A, Dodson S, O’Hara J, McPhee C, Fulton A, et al. Health literacy and fruit and vegetable intake in rural Australia. Public Health Nutr. 2017;20(15):680–2684.

    Article  Google Scholar 

  47. Borelli RC. Sustainability of well-being: Are we really choosing fruit and vegetables for our health? Agric Agric Sci Procedia. 2016;8:419–25.

    Google Scholar 

  48. Ocean N, Howley P, Ensor J. Lettuce be happy: a longitudinal UK study on the relationship between fruit and vegetable consumption and well-being. Soc Sci Med. 2019;222:335–45.

    Article  PubMed  PubMed Central  Google Scholar 

  49. Correaleita ML, Nicolosi A, Cristina S, et al. Dietary and nutritional patterns in an elderly rural population in northern and southern Italy: a cluster analysis of food consumption. Eur J Clin Nutr. 2003;57(14):1514–21.

    Article  Google Scholar 

  50. Dibsdall LA, Lambert N, Bobbin RF, Frewer LJ. Low-income consumers attitudes and behavior towards access, availability and motivation to eat fruit and vegetables. Public Health Nutr. 2003;6(2):159–68.

    Article  CAS  PubMed  Google Scholar 

  51. Choi A, Ha K, Joung H, Song Y. Frequency of consumption of whole fruit, not fruit juice, is associated with reduced prevalence of obesity in Korean adults. J Acad Nutr Diet. 2019;119(11):1842–51.

    Article  PubMed  Google Scholar 

  52. Mirmiran P, Noori N, Zavareh MB, Azizi F. Fruit and vegetable consumption and risk factors for cardiovascular disease. Metabolism. 2009;58(4):460–8.

    Article  CAS  PubMed  Google Scholar 

  53. Dehghan M, Akhtar-Danesh N, Merchant AT. Factors associated with fruit and vegetable consumption among adults. J Hum Nutr Diet. 2011;24(2):128–34.

    Article  CAS  PubMed  Google Scholar 

Download references


The authors express their thanks to Laura Alfonso and Tydaisha White for their help in the preparation of the manuscript.


This study did not receive any financial support.

Author information

Authors and Affiliations



Conceptualization: AB, WJF, FU, NK, AKK, FNÖ; Data curation: NK, AB, WJF, FU; Formal analysis: AB, WJF, FU; Funding acquisition: NK, AB, WJF, FU; Investigation: NK, AB, WJF, FU; Methodology: AB, WJF, FU; Project administration: AB, WJF, FU; Software: AB, WJF, FU; Supervision: AB, WJF, FU, NK, AKK, FNÖ; Visualization: NK, AB, WJF, FU; Writing—original draft: AB, WJF, FU, NK, AKK, FNÖ; Writing—review and editing: AB, WJF, FU. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Faruk Urak.

Ethics declarations

Ethics approval and consent to participate

The data were obtained through the joint teamwork of both the Turkish Statistical Institute (TSI) and the European Union Statistical Office (SOEU). We obtained this data from TSI in return for a contract without the need for an ethics committee document and used it in our study.

Consent for publication

Not applicable.

Competing interests

Authors declared that they have no competing interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.



See Tables 4, 5, 6 and 7.

Table 4 Vitamin and micronutrient content of commonly eaten fruits in Turkey
Table 5 Vitamin and micronutrient content of commonly eaten vegetables in Turkey
Table 6 Maximum likelihood estimates from the bivariate probit model
Table 7 Marginal effects of explanatory variables on eating one or more servings of fruits and vegetables in Turkey using the bivariate probit model

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Küçük, N., Urak, F., Bilgic, A. et al. Fruit and vegetable consumption across population segments: evidence from a national household survey. J Health Popul Nutr 42, 54 (2023).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: