1 | P a g e FACULTY OF SCIENCE, Department of Food and Resource Economics Thesis Submitted for the Award of a Master’s degree of MSc in Agricultural Economics Title: PATTERNS AND DETERMINANTS OF VEGETABLE INTAKE IN BABATI DISTRICT, TANZANIA Author: Victor William Jape (fkn608) Supervisors: 1. Christian Elleby (PhD), Assistant Professor at the Department of Food and Resource Economics—Copenhagen University, Denmark 2. Justus Ochieng (PhD), Agricultural Economist at the World Vegetable Center (Worldveg)—Eastern and Southern Africa, Arusha, Tanzania Submission Date: 22nd June 2017 2 | P a g e ABSTRACT The vegetable has potential to expand the diversity of rural and urban diets thereby improve human nutrition and health. The World Health Organization (WHO) recommends an intake of a minimum of 240 grams of vegetable per person per day in order to supply the body with the required micronutrients that are necessary to improve human health and reduces the risk of developing chronic diseases and disorders such as obesity and malnutrition. Despite the nutritional benefits, consumption levels, particularly in developing countries, are generally reported to be below the recommended level. This study presents the current consumption status and analyses the factors influencing household vegetable intake in Babati District, Tanzania. The study used cross-sectional data from 257 farm households and applied a binary logistic regression model to estimate determinants of vegetable intake. Results show that the mean daily intake of vegetable per person is 205.9 grams. Thirty-two percent of the sampled households had mean daily intake per person below the minimum recommended level. Education level, income, household size, having vegetable home garden, gender of the head and perception of the safety of vegetables sold in the market were found to significantly contribute to the vegetable consumption. The policy implication of the findings is that strategies that encourage households to grow vegetables at home, improve access to education and knowledge about healthy eating and build the capacity of women in making food-related decisions are likely to foster more consumption of vegetables. 3 | P a g e ACKNOWLEDGEMENT This work has been successful because of the support from different individuals and organisations. I would like to specifically thank my two supervisors—Christian Elleby, Assistant Professor at the Department of Food and Resource Economics, the University of Copenhagen and Justus Ochieng, Agricultural Economist at the World Vegetable Center, Arusha, Tanzania for their guidance and constant review and comments on this work. I would also like to share my sincere appreciation to the Danish Agency for Development Assistance (DANIDA) through its Building Stronger Universities (BSU) program for the scholarship opportunity that enabled me to study my Master Degree at the University of Copenhagen. Special thanks also go to the research organisation— World Vegetable Center, Eastern and Southern Africa through the Africa Research in Sustainable Intensification for the Next Generation (Africa RISING) project for financing this research study in Tanzania. Last but never least, to my lovely wife and son who accepted to stay alone in Tanzania and allowed me to spend two years in Denmark in their absence. 4 | P a g e TABLE OF CONTENTS Abstract ............................................................................................................................................................. 2 ACKNOWLEDGEMENT ................................................................................................................................. 3 LIST OF FIGURE AND TABLES ................................................................................................................... 5 ABBREVIATIONS AND ACRONYMS .......................................................................................................... 6 CHAPTER ONE ................................................................................................................................................ 7 1.0 INTRODUCTION ....................................................................................................................................... 7 1.1 Background.............................................................................................................................................. 7 1.2 Research Objectives ................................................................................................................................ 9 1.2.1 Specific objectives ............................................................................................................................ 9 1.2.2 Research Questions........................................................................................................................... 9 CHAPTER TWO ............................................................................................................................................. 10 2.0 LITERATURE REVIEW .......................................................................................................................... 10 2.1 Consumer theory and the household economic model .......................................................................... 10 2.2 Review of the factors influencing dietary diversity with a specific focus to vegetable ........................ 11 2.3 Measuring vegetable intake and dietary assessment tools ..................................................................... 13 CHAPTER THREE ......................................................................................................................................... 16 3.0 METHODOLOGY .................................................................................................................................... 16 3.1 Study area .............................................................................................................................................. 16 3.2 Sample design and selection .................................................................................................................. 17 3.3 Data collection ....................................................................................................................................... 18 3.4 Data analysis .......................................................................................................................................... 20 3.5 Study Limitations .................................................................................................................................. 24 CHAPTER FOUR ........................................................................................................................................... 26 4.0 RESULTS .................................................................................................................................................. 26 4.1 Descriptive results ................................................................................................................................. 26 4.2 Empirical results .................................................................................................................................... 30 CHAPTER FIVE ............................................................................................................................................. 34 5.0 CONCLUSION AND RECOMMENDATIONS ...................................................................................... 34 References ....................................................................................................................................................... 37 Appendix: Questionnaire ................................................................................................................................. 45 5 | P a g e LIST OF FIGURE AND TABLES Figure 1: Map of Babati District showing some of the study villages ............................................... 23 Figure 2: Vegetable consumption patterns for different households ................................................. 35 Figure 3: Different types of vegetables consumed and their frequencies .......................................... 36 Figure 4: Average quantities for the different types of vegetables .................................................... 37 Table 1: Variables used in the analysis .............................................................................................. 28 Table 2: Mean daily intake of vegetables for the different categories of households ....................... 34 Table 3: List of types of vegetables consumed by the sampled household ....................................... 35 Table 4: Regression results ................................................................................................................ 38 6 | P a g e ABBREVIATIONS AND ACRONYMS NCDs—Non-communicable diseases WHO—World Health Organization FAO—Food and Agriculture Organization of the United Nations UN—United Nations SES—Social Economic Status UNICEF—United Nations International Children’s Emergency Fund TNNS—Tanzania National Nutritional Survey EUFIC—European Food Information Council STEPS—WHO STEPwise approach to chronic disease risk factor surveillance FFQ—Food Frequency Questionnaire URT—United Republic of Tanzania RISING—Research In Sustainable Intensification for the Next Generation PHC—Population and Housing Census TZS—Tanzanian Shilling Df.—Degree of freedom DANIDA—Danish Agency for International Development Assistance BSU—Building Stronger Universities 7 | P a g e CHAPTER ONE 1.0 INTRODUCTION 1.1 Background Low intake of vegetable and fruit is one of the leading contributors to the rising burden of chronic diseases globally and is a major cause of deaths worldwide (WHO, 2004; Hall, et al., 2009). In low and middle-income countries, it is estimated that 80% of deaths are attributable to chronic diseases (WHO, 2013). Adequate intake of vegetables ensures to meet body demand for important nutrients such as vitamin A, vitamin C, iron calcium and magnesium, dietary fibres, antioxidant and phytochemicals (Uusiku, et al. 2010). Deficiencies of micronutrients, in particular, Vitamin A can cause night blindness for adults and may reduce bone growth for children (TNNS, 2014; UNICEF, 2017). In Africa, the vegetable has potential to expand human’s source of food and nutrition, thereby improving availability and dietary diversity. It also acts as a reliable source of income especially for the poor (Afari-Sefa, et al., 2012; Jaarsveld, et al., 2013). In Sub-Saharan Africa, the vegetable can help to reduce the burden of malnutrition caused by the deficiency of micronutrients especially in rural communities (Olivier, Katinka and Martin 2010; Afari-Sefa, et al., 2012). In Tanzania, the rate of malnutrition is high in both rural and urban communities and is particularly common among the low-income groups (Ochieng, et al., 2017). The heavy reliance of staple diets rich in carbohydrate with limited intake of vegetables and fruits is the major challenge to tackle malnutrition and poor dietary patterns in Tanzania (Marie et al., 2005; Leach and Kilama 2009). Furthermore, the increasing trend in consumption of foods with high fat and energy contents particularly among the youth is alarming and food such as French fries chips and scrambled eggs become their main favourites. These may increase cardiovascular diseases and obesity. A substantial amount of studies have addressed the importance of vegetables and their role in supplying human body with essential nutrients to prevent various deadly diseases and improve health and nutrition (Ritson and Mai 1998; Ezzati, et al., 2002; Yang and Keding 2009). Despite the nutritional facts, intake of vegetable is still low especially in developing countries including Tanzania (IARC, 2003; Marie, et al., 2005). A report on the Joint FAO/WHO Expert Consultation on diet, nutrition and the prevention of chronic diseases, recommends the intake of a minimum of 400 grams of fruit and vegetables per day (240 g for vegetable) for the prevention of chronic diseases and disorders such as cancer, diabetes and obesity (WHO, 2004). According to WHO, (2004) sufficient intake of fruits and vegetables can serve, approximately 2.7 million lives annually. 8 | P a g e In their study in Mozambique, Lunet et al., (2012) identified that 95 percent of the subject households (n=3323) had an intake of below the five servings of vegetable and fruits per day. In East African countries, the average consumption of vegetables and fruits also recorded to be below the minimum of 400 grams per person per day recommended by WHO. According to WHO, overall, Kenya, Tanzania and Uganda’s consumption stand at only 38 percent of the recommended amount (The East African, 2011). In Tanzania in 2003, it was estimated that average intake of fruit and vegetable was 164 g per person per day (WHO, 2004). A study conducted by Keller et al., (2012) in Zanzibar also found that a mean daily intake of vegetable in Zanzibar was only 56 g. Low fruit and vegetable intake are among contributor of the high deaths rates in East African countries. It is estimated that 27 percent of all deaths that occur in the East African region are attributable to low fruit and vegetable consumption (WHO, 2004). Consumption of food and in particular vegetable can be influenced by many factors. These factors can be categorized as (1) biological factors such as a need to satisfy hunger, appetite and taste; (2) psychological factors such as mood and attitude; (3) physiological factors such as health status; (4) socio-cultural factors such as belief and taboos; (5) physical factors such as education and skills and (6) socio-economic factors such as income, availability, cost, and access (EUFIC, 2006). While there are several studies carried out in Europe, America and other parts of the world to assess how these factors influence food consumption in general and vegetable in specific, similar studies in developing countries particularly, in Tanzania are lacking. An assessment of current patterns in consumption of vegetables and their determinants in low-income countries is of particular importance if adequate promotion measures are to be established. 9 | P a g e 1.2 Research Objectives The main objective of this study was to understand vegetable consumption patterns in Babati and assess factors influencing household vegetable intake. 1.2.1 Specific objectives 1. To describe the household-level vegetable consumption patterns in Babati District 2. To analyse determinants of vegetable intake in Babati District 1.2.2 Research Questions 1. What is the current situation of vegetable intake in Babati District? 2. Are the hypothesised determinants of household vegetable consumption significant? How do they influence the consumption of the recommended level of vegetable in the households? 10 | P a g e CHAPTER TWO 2.0 LITERATURE REVIEW 2.1 Consumer theory and the household economic model The basic framework for explaining economic behaviour and decision-making started in the “Utility Theory” first developed by Nicholas Bernoulli, John von Neumann and Oscar Morgenstern. This theory posits that consumer is a rational economic decision-maker who makes choices to satisfy their own needs (Deaton & Muellbauer, 1980). A set of rational preferences made by a consumer represents his utility function. The original idea of utility theory is mainly based on explaining consumer purchasing decision. However, contemporary research on Consumer Behaviour considers a wide range of factors influencing the consumption and acknowledges a broad range of consumption activities beyond purchasing (Deaton & Muellbauer, 1980; Marie, et al., 2005). These activities commonly include; identification of need, search for information about alternatives, evaluating the different alternatives to building purchase intent, purchasing, consumption and post-consumption activities (Deaton & Muellbauer, 1980). Two main household economic models are commonly applied in examining household economic decision-making process, resource allocation, income earning mechanisms and gender division of labour. The two models are the unitary and collective models of household behaviour (Matilla, 1999). The two models are based on the consumer’s choice theory. The unitary household model is the standard and considers a household as a single consumer unit. Furthermore, it assumes that decisions within a household are made jointly and that the household members have the same preferences, and thus, the household maximises a single set of objectives and have one utility function (Ellis, 1988; Matilla, 1999). Although the unitary economic model formed the simple and basic framework to explaining the household decision-making process, it does not reflect the social welfare within the household. The social welfare is attained when individual preferences are considered in determining the household preferences. In some households, decisions are made by a ruling household head since s/he is the one with full control over resources. These decisions may not reflect individual preferences. Other members may not have the same preference as those dictated by the household. Unitary models are not suitable to review individual preferences with the household and the intra-household inequalities. They should mainly be used in the studies where the preferences of individual members are not the main objective (Matilla, 1999). Individual preferences would have an impact on the overall household decision especially for key decisions such as a number of children (Martin, et al., 2006). 11 | P a g e Alternatives to the unitary model are suggested if one needs to account for the individuality in decision making within the household. These are called collective models and may range from the bargaining models to cooperative models (Browning, et al., 2004). According to Browning, et al., (2004), the collective model can be implemented under the assumption that the household’s welfare function is a weighted sum of the members’ own utility functions. Browning, et al., (2004), further added that the Pareto weight of this welfare function may depend on factors such as prices, total expenditure on all goods and on other variables such as household education, gender and income distribution which do not enter the individual preferences. Empirical studies suggest that selection of household economic model should be contextual and should reflect the focus of a research, its geographical coverage and purpose (Matilla, 1999; Browning, et al., 2004). If the main objective is to review outcomes of the collective decisions, the selection of the variables to include in the household model should be carefully considered to avoid variables that clearly postulates individual preferences. 2.2 Review of the factors influencing dietary diversity with a specific focus to vegetable Households’ food consumption decisions differ and are influenced by several factors. Socioeconomic status of the household: in this study is defined by several aspects including household income, the level of education, household size and composition, and household location. In other cases, the definition may be extended to consider health status of the respective household and participation in social events of the community. Generally, increase in income is associated with an increase in consumption of vegetables. A study conducted in Vietnam found that there is a positive correlation between household income and vegetable intake (Tan Van Bui et al., 2016). A household with higher income is also more likely to consume healthier foods such as whole grains, lean meats, fish, low-fat dairy products, fruit and vegetables (Mayen, et al., 2014). Low consumption of vegetables in low-income households is associated largely with the fact that the households have to prioritize the fulfilment of the basic energy requirements to avoid hunger, as such denser energy sources of food such as staples are preferred because are available at a relatively cheaper prices compared to vegetables and fruits (Minot, 1998; Marie, et al., 2005). Household size is also said to be correlated with vegetable consumption. However, larger households tend to allocate a lower share of their budget to vegetables (Marie, et al., 2005; Keller, et al., 2012). Education is also a key determinant of vegetable consumption. 12 | P a g e A study conducted by Marie et al. (2005) in developing countries identified that the higher education level is associated with higher consumption of vegetables. Place of residence could also affect consumption of vegetables. Households in urban areas normally have higher dietary options and this could affect vegetable intake. For example, in their study to review the socioeconomic determinants of dietary patterns in low and middle-income countries, Mayen and colleagues identified that urban households were generally consuming healthier diets with more vegetables than rural households (Mayen, et al., 2014). Marie, et al., (2005) further added that urban residence is significantly associated with a greater share of budget allocation for vegetables and fruits in some countries including Tanzania. Higher consumptions in urban areas are also associated with higher level of education. For example, in their study to assess the fruit and vegetable consumption in Mozambique, Padrao and colleagues (2012) identified that urban subjects with higher education level had higher consumptions of fruit and vegetables. However, no such correlation was found in rural population (Padrao, et al., 2012). A study in Vietnam further showed that provinces in urban population had the higher mean intake of fruit and vegetables (Tan Van Bui et al., 2016). Knowledge and awareness on health benefits of consuming fruits and vegetables can be a key factor in influencing vegetable consumption behaviour. Knowledge and awareness can be associated with education level. Awareness may create attention to seek knowledge which may influence change in behaviour. For example, in the United States, health awareness and knowledge of the number of vegetable servings recommended per day have been associated with greater vegetable intake (Nayga, 1995). Although it is difficult to measure actual behaviour change resulted from knowledge and awareness, it is important to include this aspect in analysing determinants of food intake. Knowledge and awareness are significant factors from a policy point of view when it comes to suggestions to influence consumer behaviour. Availability and accessibility are also key in influencing vegetable consumption. The fact that vegetables are perishable in nature, and given the limited infrastructure in many of the developing countries, makes their availability to consumers very seasonal. Literature suggests that post-harvest technologies to extend the harvest period and to facilitate storage are of particular importance (Ali and Tsou 1997). 13 | P a g e Household consumption patterns could also be influenced by what the members can produce at home. For example, Ali and Tsou (1997) reported that a program to promote vegetable home gardening in Bangladesh had significantly increased the volume of vegetables consumed for the respective households. However, the decision to produce vegetables for own consumption could be determined by whether the cost of home production is considered low and feasible compared to market transactions cost (Marie, et al., 2005). In addition to market transaction cost, other factors such as consumer’s perception could influence the decision to produce vegetable at home. This situation is particularly true when households do not trust vegetables sold in the market due to, for example, perceived poor safety due to heavy application of industrial chemicals such as pesticides in the production process. Home gardening would improve household’s access to vegetables thereby influencing more consumption, especially where low consumption is influenced by the lower per capita daily availability of vegetable products. Gender aspect is also key to consider in order to determine household food consumption patterns and their determinants. Some studies show that there are significant differences in vegetable consumption between female-headed and male-headed households. A study of household budget in Rwanda found that female-headed households allocate a larger share of the budget on vegetables and fruits (4.5 % compared to 3.1% than male-headed households (Minot, 1998). It is also believed that in the households where the female has higher social status and or more power in allocating resources, more priority is placed on child health and nutrition. For example, Smith et al., (2003) showed that women’s status has a statistically positive significant effect on the nutritional status of children. Some studies found no statistically significant difference in food consumption between male-headed and female- headed households. For example, a study in Vietnam found no statistically significant differences between female-headed and male-headed households in terms of vegetable consumption (Tan Van Bui et al., 2016). These mixed results entail the need for further research on the gender role in making decisions particularly those related to food consumption. 2.3 Measuring vegetable intake and dietary assessment tools Attempt to measure vegetable consumption so far has, in many studies relied on the World Health Organization—WHO STEPwise approach to chronic disease risk factor surveillance (Hall, et al., 2009; Padrao, et al., 2012; Tan Van Bui, et al., 2016; Kjøllesdal, et al., 2016). The WHO STEPwise approach to Surveillance (STEPS) is a regarded as a simple, standardised method for collecting and analysing data on consumption and chronic disease risk. 14 | P a g e It also provides a guide to designing survey questions to be included in collecting data related to food consumption. In the WHO STEPwise approach, a single portion/serving of fruit and vegetable is considered to be equivalent to 80 grams. Therefore, five servings per day make a minimum amount of 400 grams recommended by WHO (60:40 vegetable to fruit ratio). The WHO STEPwise methodology proposes two core questions to be included in the questionnaire to capture vegetable intake; (1) in a typical week, on how many days do you eat vegetable? (2) How many servings of vegetables do you eat on one of those days? When asking these questions the STEPS proposes to show respondent a “nutrition card" that represents an example of local vegetables. In answering the questions respondent is asked to think in a typical week and recall the last year. The WHO aims to promote an increase in vegetables and fruit consumption, so that the sufficient amount become part of daily diet of every person. It further emphasises that the mean intake goal should be expressed in numerical terms in order to evaluate its potential health benefit (Agudo Antonio, 2005). Selection of appropriate measurement method to be used in determining the numerical term of the average intake is therefore important in order to test the reliability of numerical values presented. Measuring vegetable intake accurately is not easy. Countries may not define vegetables and portion sizes in the same way. A portion of leafy vegetable such as spinach may have a different weight as a portion of fruit vegetables such as tomato or eggplant. Different tools have been proposed in assessing vegetable intakes at the household or individual level. Commonly used instruments now days are the food frequency questionnaire (FFQ) and the 24-Hour dietary recall. The FFQ is mostly preferred because of great flexibility and ease of application. It contains a structured list of individual foods or groups of foods, and the respondent is asked to estimate the frequency of consumption, indicating the number of times the food is consumed over a given period. In some cases, the FFQ is semi-quantitative as it specifies a standard serving or portion for each item listed. The quality of the estimates, however, is highly dependent on specification, i.e. whether the vegetables are expressed as groups or single items in the questionnaire and the number of items included (Agudo Antonio, 2005). The 24-hour dietary recall used to assess vegetable intake over a short period of time, normally within twenty-four hours prior to the interview. It is particularly well suited in the assessment of the group mean intake assuming the representativeness of population sample and a well-balanced distribution of 24-hour surveys by season and weekdays (Agudo Antonio, 2005). 15 | P a g e It is advised that selection of the appropriate tool to use in estimating vegetable intake should take into account, a number of factors such as the purpose of the study, the need for group data versus individual data, the population characteristics, the time frame of the interest and the available resources (Agudo Antonio, 2005). The purpose of the study is one of the main issues to consider when selecting a tool to assess vegetable intake. Estimates of vegetables may be needed for different purposes such as designing interventions, nutritional surveillance, screening for nutritional programmes or general nutrition assessment (WHO, 2004). Most analytical studies focus on investigating a relationship between disease risk and level of intake (see Padrao. et al., 2012; Tan Van Bui, et al., 2016; Kjøllesdal, et al., 2016). They tend to rank individuals according to their relative intake rather than providing the absolute measurement of intakes. Therefore, in most cases, they report vegetable intake in relative terms rather than the absolute amount consumed (for example; the number of servings of vegetables instead of average quantities). Precise estimation of vegetable intake and determination of the proportion of population eating below or above a given level of recommended intake is needed if we need to understand the real situation for nutritional and policy recommendations. However, it should be clear that obtaining precise estimates is almost impossible especially if we consider heterogeneity in vegetable consumption. Several measurement errors could occur in the assessment process. The fact that we use the recalling method to measure intake, the dependency of respondent memory may also bring unreliable estimates. In addition, characteristics of the respondent such as cultural background, perception, education, age, gender, attitude and knowledge may all affect the responses. Other problems related to the design of the tool may also result into assessment errors. Such design aspects as the structure of the questionnaire, order of questions and the time frame recalled should all be taken into account in order to minimise the difference between measured intake and the true intake (Field, et al., 1998). 16 | P a g e CHAPTER THREE 3.0 METHODOLOGY 3.1 Study area This study was conducted in Babati district, Tanzania. Babati is one of the six districts found in Manyara region, northern Tanzania. It is located at 04° 13′S 035° 45′E. Babati district covers an area of 6,069 km2, and a large proportion (640km2) is under water bodies of Lake Babati, Lake Burunge and Lake Manyara which is favourable for agricultural activities. The district is bordered to the South by Dodoma Region, to the North by Arusha Region, South West by Hanang District, North West by Mbulu District, and South East by Simanjiro District. Within the Babati district, there are two district councils; Babati Rural District Council and Babati Urban District Council. Administratively, the district has 29 Wards and 96 villages and its capital are Babati Town, 172 kilometres south of Arusha. The map of the study area is shown in figure 1 bellow. Figure 1: Map of Babati District showing some of the study areas Source: Bekunda, (2014) 17 | P a g e According to the 2012 Tanzania Population and Housing Census, Babati has an estimated total population of 405,500. Population distribution by place of residence shows that 312,392 people (77 percent) are in rural and 93,108 (23 percent) are in urban areas. Furthermore, there are approximately 80,629 private households in Babati, 74.2 percent of which are in rural area and 25.8 percent are in urban. Babati is one of 28 districts in Tanzania whose population identified in 2010 as having a high level of poor nutritional status caused by low intake of nutritious foods (Integrated Food Security and Nutrition Assessment Report, URT, 2010). The extent of poverty in the district was also reported to be high. There are a number of programs in Babati that work to improve people livelihood. Some are strategically targeting to improve nutrition and dietary diversity by promoting increased production and consumption of vegetables. 3.2 Sample design and selection The target sample size (n=250) was obtained from the population unit of 80,629 private households living in Babati. A multi-stage sampling technique was employed to obtain the target sample size. The first stage involved a purposive selection of Babati district due to agro-ecological conditions that favour production of a diverse number of crops including vegetables and existence of initiatives to promote vegetable production and dietary diversity. The second stage involved dividing the population into rural and urban units. Then, a sample size was obtained from each unit using a proportional sample determination method as proposed by Anderson, et al, (2007) and Kothari, (2004). The next stage was to locate the 250 households from rural and urban areas. Five villages from Babati Rural and three streets/villages from Babati Town were purposively selected. Selection of villages and streets was based on the recommendation from researchers who had experience with the study area and also representativeness in terms of population distribution and coverage of initiatives promoting vegetable production and intake. The last stage involved random picking of respondents in each study area. The rural and urban representative samples of the study area were obtained as shown below. Let N represent total population of private households (N=80,629) and n represent proposed total sample size (n=250) Let r and u represent rural and urban stratum. Let Nr be a stratum of rural private households, (Nr=59,853) 18 | P a g e Let Nu be a stratum of urban private households, (Nu=20,776) Let Pr be a proportional allocation of population in stratum Nr Let Pu be a proportional allocation of population in stratum Nu Let nr be the sample size of rural stratum and nu be sample size of urban stratum It then follows that sample sizes (nr and nu) in each stratum is obtained by the formula; nr= n*Pr and nu= n*Pu but, Pr=Nr /N, and Pu=Nu/N Thus, Total sample size (n)= n*( Nr /N) + n*( Nu/N) = 250(59,853/80,629) + 250(20,776/80629) =185+65 Therefore from the 250 sampled households, 74 percent expected to be from Babati rural areas and 26 percent from urban. 3.3 Data collection Data were collected for ten days in February 2017. Two experienced enumerators were recruited and trained to support the researcher (me) in data collection. A total of two hundred and fifty-seven (n=257) households were interviewed using a structured questionnaire. Out of the 257 respondents 185 are from rural villages (Seloto: n=38; Shaurimoyo: n=37; Bermi: n=36, Galapo: n=37 and Matufa: n=37) and 72 are from Babati Town streets (Himiti: n=25; Majengo: n=24 and Managha: n=23). The questionnaire (see attachment in Appendix 1) comprised of the following sections: (1) demography and socio-economic information (location, gender, age, education, household size, and self-reported income and expenditure); (2) Vegetable intake (reported on weekly basis); (3) Household preference, perception and experience, knowledge and awareness on nutrition and health, vegetable access and availability and home gardening. The questionnaire was pretested to determine the amount of time spent per respondent, the convenience of getting information and validate common vegetable types consumed in the study areas. 19 | P a g e Data was collected in private households. Definition of private household in this study followed that one given by United Nations, Statistics Division and the 2012 Tanzanian PHC, where, private household is defined as a person or group of persons living together in the same homestead or compound (regardless of whether they possess family relationship), who make provision for their own food or other essentials for living (United Nations website, 2016; 2012 PHC Survey report, 2016). In that essence, groups such as schools and orphan centres were not included in the study. Before the interview, each potential interviewee was briefly introduced to the study purpose and the consent obtained in advance. To increase the accuracy of response, we interviewed only those household members responsible for purchasing and/or preparing food. So, the respondents were mostly women. Household vegetable intake per week was obtained and used to calculate the mean daily intake per person. Consumption in the previous week (seven days before the interview date) was used as a reference period. The mean daily intake per person was measured in grams and was obtained as follows: First; we conducted a market survey to identify common vegetable packages and sizes in order to calculate their average weights. We took different samples from different local markets in each village, weigh the samples using a scale and record the weights in grams. The common vegetable packaging in all the markets called "bunch" was identified and the average weight of a single bunch for each vegetable type was calculated and recorded. Second; during the interview, the respondent was presented with the list of different common types of vegetables and asked to recall and identify all types of vegetables consumed by their household in the last week. For each vegetable type consumed, the respondent was asked to recall and mention the number of intakes per week. A common plan of three meals per day (breakfast, lunch and dinner) was used as a reference to help respondent identify the correct number of intakes. The total number of intakes per week for that particular vegetable was then recorded. Then, we asked the respondent to mention the estimated number of market-size bunches they normally prepare per single meal. The total quantity consumed in by the household in a week was then obtained by multiplying the average weight of single bunch (as determined from the market survey) first with the average number of bunches per intake/meal and then with the total number of intakes in the whole week. After obtaining the total amount (in grams) consumed by the household per week, we computed the average intake per person per week by dividing the total amount by the number of members (excluding children  2 years old). Lastly, the mean daily intake per person was obtained by dividing the average intake per person per week by seven days. 20 | P a g e 3.4 Data analysis Collected data was organised and coded for the analysis. Specific objective 1: Descriptive statistics were used to present and describe the sample characteristics and household socio-economic characteristics. Household-level vegetable intake was computed and compared across different groups such as rural versus urban households, vegetable producers versus non- producer, female-headed versus male headed households, and low-income versus middle-income households. Vegetable diversities as defined by the different types of vegetables consumed and frequencies are also presented. Specific objective 2: A logit model was used to assess the determinants of vegetable intake among households in the study area. The dependent variable, mean daily intake of vegetables is coded as (rec=1 if the mean daily intake per person is ≥ 240 g; rec=0 if is < 240 g) and modelled against the set of explanatory variables. Table 1 presents and describes the explanatory variables hypothesised to influence the vegetable intake in the household. Marginal effects were computed in order to allow for easy interpretation of the regression results from the logit model. The marginal effects show the change in probability when the predictor or independent variable change by a given unit. 21 | P a g e Table 1: Description of variables used in the analysis Variable code Variable name and description Variable type HH_size Household size: Number of members living together for at least 6 months Continuous kids5 Number of children < 5years old in the household Continuous kids14 Number of children 5-14 years old in the household Continuous secondary Household education defined by availability of members with at least secondary education Binary f_head Household headship by female Binary location Living status: Household is living in urban or rural area Binary organic Household is concerned about use of inorganic fertilisers and pesticide in vegetable farming Binary income Household monthly per capita income Continuous training Household participation in training or awareness campaign about nutrition and health eating Binary access The access to vegetable Binary garden Having vegetable garden Binary 22 | P a g e Model specification As shown in Wooldridge (2013), binary logistic models can be specified as: 0 1 1( 1| ) ( ..... )k kP y x G x x      …………………Equation (1) Note that G is a function taking on values strictly between zero and one: 0 ( ) 1G z  for all real numbers z . For a logit model, G is the cumulative density function for the standard logistic distribution and can be expressed as: ( ) exp( ) / [1 exp( )]G z z z  . The model can also be derived from an underlining latent (unobserved) variable *y . Consider the following specification: * * 0 , 1[ 0]y x e y y      . .................................................Equation (2) Note that, the observed value y =1 if the function *[ 0]y  is true. It means that when * 0y  then, the observed variable y is zero. Furthermore, the error term e is assumed to be independent of X and that it follows a standard logistic distribution. Given the assumption and considering equation (2), the response probability of y for the logistic can be derived as: * 0 0 0 ( 1| ) ( 0 | ) [ ( ) | ] 1 [ ( )] ( ) P y x P y x P e x x G x G x                    .....................................Equation (3) Equation (3) is similar to equation (1), only that the latent variable formulation in equation (3) gives the impression that our primary interest is on effects of each independent variable x on unobserved variable *y . Now drawing the concept from equation (1) and (3) the logit model is specified as follows: 0 1 2 3 4 5 6 7 8 9 10 11 ( 1| ) ( 5 14 sec ( ) _ ) P rec x G HHsize kids kids ondary income urban f head organic garden training acces e                          23 | P a g e Computation of the marginal effects Since the primary interest is to estimate the effect of each independent variable on the probability that the mean daily intake of vegetable is in recommended amount, the value of betas of each independent variable in the logit model are by themselves, not interpretable until we compute their partial effects. In order to interpret the partial effect of the explanatory variables on the response probability, we computed the marginal effects. Marginal effects are the partial effects and can be expressed as the percentage of changes each explanatory variable has on the probability that the observed dependent variable Y equals 1 (Richard, 2017). With binary independent variables, marginal effects measure how predicted probabilities change as the binary independent variable changes from 0 to 1, and for continuous independent variables, the marginal effect measures the instantaneous rate of change (Richard, 2017). That is, it provides an approximation to the amount of change in the dependent variable Y that will be generated by a 1-unit change in explanatory variable X (Wooldridge, 2013). Wooldridge (2013) showed that for a binary explanatory variable (example, being a vegetable producer or not), the partial effect of changing 1x from zero to one (i.e. from being non-producer to being a producer of vegetable) to the probability ( 1| )P rec x holding all other factors fixed can be expressed as: 0 1 2 2 0 2 2( ... ) ( ... )k k k kG x x G x x              . ...................................Equation (4) Therefore, when we consider our example of household participation in vegetable production, equation (4) implies the change in the probability of consuming the recommended daily amount per person, given household's involvement in the vegetable production. Note that the sign of 1 will help to determine if, for example, participation in vegetable gardening positively or negatively affect the consumption of the recommended level. Wooldridge (2013) further explained that for a continuous explanatory variable, (example in our case, per capita income) the partial effect on response probability can be expressed as: 0 ( 1| ) ( ) ,j j P rec x g x x         where, ( ) ( ). dG g z z dz  Because G is the cumulative density function (CDF) of a continuous random variable, g is a probability density function. The CDF is strictly increasing and therefore ( ) 0g z  for all z . 24 | P a g e The partial effect of jx on the response probability hence depends on all explanatory variables ( x ) through the positive value 0( )g x  . After specifying a logit model, a Wald statistic is used to test whether the significant parameters do really have the explanatory power. The null hypothesis was tested at 5% significance level. Null hypothesis ( 0H ): 0 1 11..... 0      Alternative (H0): At least one parameter is different from zero McFadden pseudo-R-squared was used as the measure of goodness-of-fit of the logit model. Since the McFadden pseudo-R-squared is based on the log-likelihood functions for the unrestricted and restricted models, the two models were first estimated and the log-likelihood values computed. The McFadden pseudo-R-squared was obtained by the formula: ( ) 1 ( ) ur r Loglik MF Loglik     Where, ( )urLoglik  is the log-likelihood value from the unrestricted logit model and ( )rLoglik  is the log-likelihood value from the restricted logit model (with only the intercept). 3.5 Study Limitations The WHO STEPwise approach to Surveillance (STEPS) has formulated the basis for carrying out survey studies that intend to measure food consumption in general and for vegetable intake in particular. STEPs approach suggests to a researcher to show the respondent, a nutrition card that reflects a size of serving and assume each serving as an equivalent to 80 g in metric weight. The assumption of the eighty grams per portion/serving size is useful and provide an easy way to calculate the amount consumed per day. However, intake of vegetables may not be the same for different households. Definition of portion size may not be equal for all consumers and vegetable types. The suggestion to consider a specific context of the study area is key in order to accurately estimate the amount of vegetable consumed. However, it should be cautioned that an attempt to disregard the recommendation of the 80 grams portion size and consider the specific study area context in calculating the average amount of vegetable consumed may lead to unreliable results that can mislead the general conclusion regarding the average intakes. This study considered a specific context of the study area to estimate the average intake of vegetable. 25 | P a g e While we are confident that the quantities reported are representative of the sampled households, we allow a room for criticism and suggestions for better approaches to measuring vegetable intake. Different challenges of measuring food intake especially in developing countries need to be addressed in order to have more accurate data. These may include: the diversity of packages and lack of uniform unit quantities to support the determination of average consumption amount, the possibility of false responses that may lead to outliers, and measurement errors associated with the use of scales. 26 | P a g e CHAPTER FOUR 4.0 RESULTS 4.1 Descriptive results Socio-economic and demographic characteristics of the sample Seventy-two percent of total households interviewed are located in geographically defined rural areas and twenty-eight percent were from urban. The average household size in the sample is 4.9 persons (national household size: 4.8 persons). Every household had on average, 0.6 kids below five years old and 1.5 children who are 5-14 years old. Composition by gender shows that Babati households have on average, 0.2 more females than males (mean= 2.5). About 36 percent of households have at least one member with secondary education. Household headship by gender shows that 25 percent (51 households) are headed by females. The average age of a female head is 49 years while that of male head is 45 years. Furthermore, female-headed households have on average smaller household size (4.7 vs. 5.2 persons) compared to male-headed households. The household average income is 256,463 Tanzanian Shillings (TZS) per month (standard deviation=205,475 TZS). Income per capita is approximately 60,000 TZS (~27 USD). The World Bank classifies Tanzania as the low-income economy because the income per capita, calculated using the World Bank Atlas method is ≤ 85 USD per month (World Bank website, 2017). About 97 percent of the sampled households in Babati had an income per capita less than the 85 USD. The poverty headcount ratio shows that about 92 percent of the sampled households have an income per capita per day below the 2 USD. Furthermore, on average, 62 percent of household monthly income is dedicated to food. Households whose average income per member fall below the 2 USD dedicate a higher percentage (78.8 versus 58.6 percent) of income to food compared to their colleagues. Expenditures on vegetables show that households in Babati spend about 10 percent of their total food budget for vegetables. Female-headed households earn on average, 18 percent less per month compared to male-headed households. 27 | P a g e Vegetable consumption patterns in Babati district The mean daily intake of vegetable per person in Babati was identified as 205.9 g (standard deviation=110.4 g; median=182.7 g). National average reported in 2005 was 164 g (Marie T, et al, 2005). Thirty-one percent (31.5%) of all sampled households had mean daily intake per person below the minimum level recommended by WHO. Table 2 summarises the vegetable intake identified for different categories of households. Table 2: Mean daily intake of vegetables for the different categories of households Category Mean daily intake per person (grams) Standard deviation Urban households 184.9 112.1 Rural households 214.11 108.9 Households with vegetable gardens 246.4 109.9 Households without vegetable gardens 126.8 54.2 Households headed by a female 351 97.9 Households headed by male 170 80 Low-income group Middle-income group Participation in nutritional education program Never participated in any nutritional education program 222 202.8 252.9 114.1 117.5 109 102.9 50.6 The average intakes for different household groups can also be visualised by a graph as shown in figure 2. 0 50 100 150 200 250 300 350 400 Living in urban Living in rural Have vegetable home garden No home garden Participated in nutrition program Never participated in any nutrition program Female headed Male headed Low income Middle income middle income Average intake per household (grams) per day Fig.2: Vegetable intake for different household categories 28 | P a g e Table 3 shows different types of vegetables identified to be consumed by the sampled households and average quantity intake per household per day. Table 3: Different vegetables consumed and average intake per household Vegetable type Percentage of households consumed (n=257) Average intake (grams) per household per day Amaranth 58 239.4 Jute Mallow 21 86.7 Tomato 95 430 African NightShade 33 120.2 Ethiopian Mustad 56 104.4 Carrots 28 355.7 Sweet potato leaves 8 128.4 Chinese leafy 56 111 African eggplant 19 129.2 Cabbage 4 142.9 Pumpkin leaves 10 115.5 Okra 16 95.1 Cowpea leaves 11 164 Spider plant 7 134 As observed in Table 3, the tomato was consumed by the most households. It also recorded the highest average intake per household per day (430 g). In Tanzania, tomato and carrots are mostly used as spice mixed together with other vegetables, legumes or meat and eaten with steamed rice, stiff porridge commonly known as ugali or banana. Cabbage observed to be consumed by fewest households. It is one of the exotic varieties that is not common especially in rural areas in Babati and other districts. Amaranth, Jute Mallow, Ethiopian Mustad, African nightshade, Sweet potato leaves, Pumpkin leaves, Cowpea leaves and Spider plant are all traditional African leafy vegetables. 29 | P a g e Okra and tomato are considered as fruit vegetables while carrot is categorised into root vegetable. The different varieties of vegetables and percentage of households consumed can also be visualised by the graph in figure 3. Figure 4 visualises the average intake per household per day of all the identified vegetable types/varieties. 58 21 95 33 56 28 8 56 19 4 10 16 11 7 0 10 20 30 40 50 60 70 80 90 100 P er ce n ta g e o f h o u se h o ld s Vegetable varieties Fig.3: Different types of vegetables and Percentage of households consumed 0 50 100 150 200 250 300 350 400 450 500 A v er ag e in at k e (g ra m s) p er h o u se h o ld p er d ay Vegetable varieties Fig. 4: Average consumption per household per day for each vegetable type 30 | P a g e 4.2 Empirical results The McFadden's goodness-of-fit measure showed that the logit model well fitted the data. Furthermore, the results of the Wald test clearly rejected the null hypothesis of no statistical significance of the explanatory variables to the response variable (Table 4). It is fair to conclude that the hypothesised explanatory variables have statistical power in explaining the probability of consuming the recommended level of vegetable. Table 4 presents the results of the logit model and the calculated marginal effects. The coefficients of the estimates only give the signs of the partial effects of each explanatory variable on the response probability of the dependent variable- ( 1| )P rec x . Table 4: Regression results of the logit model and the marginal effects Estimate Marginal Effects Std. Error z value Pr(> | z| ) (Intercept) -0.716 2.779 -0.258 0.797 Household size -0.577 -0.032 0.260 -2.218 0.027 Children between 5-14 years 0.173 0.006 0.365 0.475 0.635 Children less than 5 years -0.431 -0.015 0.433 -0.995 0.320 Secondary education 2.022 0.103 0.652 3.101 0.002 Female-headed household 3.889 0.444 0.943 4.122 0.000 Living in urban -0.692 -0.029 0.749 0.923 0.356 Concerns over vegetable safety -3.318 -0.080 1.421 -2.335 0.020 Log (Income per capita) -1.444 -0.079 0.454 -3.179 0.001 Easy and regular access to vegetable 1.909 0.080 0.779 2.450 0.014 Participation in nutrition training/campaign 1.122 0.035 1.160 0.968 0.333 Having a vegetable home garden 5.047 0.172 1.473 3.427 0.001 Sample size: n=257 Wald test: Pr (>F) = 5.415e-05 *** (Df. =11) McFadden's pseudo-R-squared: 0.7033079 (Df. =12) 31 | P a g e While the household size is statistically significant, the sizes of children below 14 years old in the household were found to have no significant effect in determining vegetable intake. Controlling for other factors, households with more members were less likely to consume the recommended quantity of vegetables. This is because they may not have adequate income to purchase enough vegetables to feed a large family. Marie et al., (2005) also found that increase in household size is negatively associated with demand for vegetables in Tanzania. They further showed that larger households allocate a lower share of their budget to purchases of vegetables (Marie, et al., 2005). Food consumption in most households in Tanzania is shared on the same dish. The increase in household size may reduce the portion consumed by a single member. Furthermore, poor households do not put much focus on increasing vegetable intake when the household grows, but rather, focus more on increasing the main staple dish. The priority is on fulfilling the basic energy requirements. Having at least one member with secondary school education or beyond in the household significantly contributed to the consumption of more vegetables. Households with higher education were significantly more likely to consume the recommended amount of vegetables. Education reflects the level of literacy and a higher level of literacy may influence individuals to seek knowledge about healthy diets and how to live a healthier lifestyle. These may in turn, foster more consumption of healthier foods in general and vegetable in particular. Table 4 shows that chances of consuming the recommended amount of vegetable are higher when the household is headed by a female. The marginal effects are positive and significant. A recent study by Marie, et al., (2005) also found a significant difference in vegetable intake between male-headed and female-headed households in Tanzania with the latter consuming more vegetables. These results support the arguments that females are more concerned with the health of the household members particularly children and usually place more emphasis on consumption of healthier foods. Smith et al., (2003) argued that quality of diets can be improved when female have full control over household resources. In Africa, women are the one responsible for household food intake and general well- being. However, ownership and decision over resources such as income remain mostly to men. Living in the urban or rural area had no significant effect in influencing consumption amount. Field observations show no much differences in socio-economic characteristics between rural and urban households in Babati. 32 | P a g e The main difference identified is in the sourcing of vegetables whereby, most urban households source their vegetables mainly from local markets while rural households sources vegetable mainly from own gardens or free access from the bush. Marie, et al. (2005) found that in Tanzania, urban households had a lower intake of vegetables than rural households although the difference was not statistically significance (Marie, et al., 2005 pg. 33). This study also found similar results in Babati District. Higher consumptions in rural areas of Babati may be due to the high availability of vegetables contributed by the feasibility of vegetable production. Feasibility of production may be linked to the existence of several livelihoods and the nutritional program currently existing in rural areas that promote production and consumption of vegetables. Concern over the safety of the vegetable is another factor identified to influence consumption of vegetable. About 27 percent of the sampled households reported being concerned about the increasing trend in the heavy application of pesticides and inorganic fertilisers in vegetable farming. Empirical results show that these households were significantly less likely to consume the recommended amount, controlling for other variables. Their mean daily intake per person was 100.6 grams per day. Lower consumption average in this group may be contributed by the lack of trust over the vegetable they founded at the market due to perceived poor safety, as one respondent explained; “There were many times when I didn’t buy vegetables just because I realised that the vegetable had fresh smell of pesticide”. Another respondent stated; “If I had good access to water I would have grown my own vegetables at home and stop to buy from other people”. Concerns over the safety can be influenced by other unobserved factors such general attitude and beliefs in the society. For example, it was observed in the study area that some villagers believe that consumption of vegetables like Chinese cabbage can affect the reproductive health of men and may lead to impotence. Household income is also found to significantly influence consumption of vegetable. The likelihood of consuming the recommended amount significantly decreases when the percentage of income increases. However, the decrease in the likelihood of consuming recommended amount is small. For example, controlling for other factors, a 10 percent increase in household income would only reduce the probability of consuming the recommended amount by .7 point. Some studies support that income is a significant factor in determining consumption of healthier foods. As income increases, people generally tend to shift to healthier diets (Marie, et al., 2005; Mayen, et al., 2014). Healthier foods are generally more costly and that many consumers especially the poor cannot afford. In the presence of alternatives that are relatively cheaper, poor households can easily switch to those alternatives. 33 | P a g e The weak negative association of income and vegetable intake observed in the sampled households may imply that vegetable is considered as inferior good and that increase in income may not results into the increased consumption. Participation in food and nutrition training and awareness programs did not show any significant impact on consumption of vegetable. Although the difference was not statistically significant in the model, households with members participated in the training programs had higher mean daily intake per person (259.9 g vs. 114.1 g) compared to their counterparts. Training and awareness program can contribute to new knowledge and hence lead to change in consumer behaviour. However, a change in behaviour may be contributed by other factors such as culture and experience. A lack of significance in this study may imply that training and awareness program is correlated with other factors such as education and gender of the household member who received the training. Production of vegetable significantly increased the likelihood of consuming the recommended amount. Households that grow vegetable had on average, 119.6 g more in their mean daily intake compared to non-growers and the difference is statistically significant at a smaller p-value (Table 4). The decision to grow vegetables at home can be influenced by several factors including feasibility in terms of cost infrastructure and availability of water, availability and access to vegetable markets, trust to the quality of vegetables sold in marketplaces and knowledge about benefits of consuming vegetables. 34 | P a g e CHAPTER FIVE 5.0 CONCLUSION AND RECOMMENDATIONS Summary of the key findings The mean daily intake of vegetables per person for the sampled households in Babati is 205.9 g. Almost 32% of sampled households had mean daily intake below the minimum level of 240 g recommended by WHO. Rural households had higher mean daily intake than urban households (214.11 g vs. 184.9 g). Households that produce vegetables for own consumption had on average, higher mean daily intake compared to non-producer households (246.4 g vs. 126.8 g) and the difference is statistically significant. In addition, households headed by a female had on average, higher mean daily intake of vegetables (351 g vs. 170 g) than male-headed households. Furthermore, households with nutrition knowledge had also higher mean daily intake compared to households without nutrition education (252.9 g vs. 114.1 g). Having secondary education, household headship by a female, having a vegetable garden for home consumption and high and regular access to vegetables were all found to be statistically significant and positive in influencing the probability of consuming the recommended amount of vegetables. On the other hand, increase in household size, increase in household income, and sensitivity of chemical application in vegetable production were all found to significantly reduce the likelihood of consuming the recommended amount. Furthermore, living in the urban or rural area, having participated in nutritional training or awareness program had no statistically significant effect in the logit model. Implication and recommendations Results from this study show that there are differences in vegetable intake among different socio- economic, demographic and physical characteristics of the households, influenced by different factors. Gender and education: It is clearly observed that there is a significant difference in vegetable intake between female-headed households and male headed households, the former seemingly consume more vegetables than the latter. When women have more power on household decisions such as allocation of resources to the food they put more emphasis on sourcing healthy foods such as vegetables. Policies to empower women can have a positive impact on household health improvement. 35 | P a g e The study suggests that more programs strategically designed to build the capacity of women in different aspects such as financial independence, enterprises development, farming and agribusiness, health and nutrition are needed. Furthermore, as women place more value on healthy diets, building and strengthening their health and nutrition knowledge and awareness can help the households to shift their diets to healthier food components such as vegetable. However, programs to empower women must consider the gender issues in the society and the intra-household relationships between male and female to avoid gender conflicts. Education is another important factor that can influence human behaviour in seeking knowledge about a healthy lifestyle. Higher education levels can influence healthy lifestyle and the demand for healthy foods such as vegetables. Results indicate that households which had members with higher education had a higher intake of vegetables compared with the households that had no member with secondary education or higher. Therefore, programs aiming to improve education status among poor households should be strengthened to give more households chance to access further education. In schools, the curriculum should also be expanded to include lessons about nutrition, healthy eating. Intervention projects such as vegetable gardening in schools can help to create awareness among youth and develop a behaviour to consume vegetables from childhood. Promotion of home gardening and improvement of market infrastructure to increase vegetable access: This study suggests that production of vegetables is important to promote increased consumption at the household level. Interventions designed to promote home gardening and increased productivity are therefore necessary to improve household access to vegetables. To achieve even a more meaningful impact on vegetable consumption, these interventions should be coupled with effective education program and behaviour change strategies that will impact knowledge and awareness of healthy lifestyle and the benefits of consuming nutritious foods such as vegetables. At the market level, policies to improve market infrastructure are necessary to make vegetables available to consumers at all time. Cost-effective storage facilities are also needed since vegetable is a perishable good. Some sorts of processing such as vegetable drying and fresh vegetable juices can be introduced to create more forms of vegetables and give more options to consumers. Consumer knowledge and awareness of health and nutrition: Psychosocial studies in developed world show that knowledge and awareness of nutrition and health play an important role in determining intake of healthy foods such as vegetables. Although household food consumption in Tanzania is also based on cultural behaviour and past experience, educational and awareness programs on health and nutrition can help to bring changes in consumer behaviour and therefore foster more intake of healthy 36 | P a g e foods such as vegetables. Behavioural change communication programs should also target to move people from thinking that vegetables are just for poor people. Vegetable consumption and safety issues: Emphasis on higher vegetable production to increase its availability to consumers should be coupled with consideration of issues related to safety and standards. With the advancement of farming technology, farmers can now easily access inputs such as inorganic fertilisers and pesticides that can boost their productivity. However, the safety of the product resulted from the use of these chemicals should be taken into account to protect the health of consumers. Concerns of consumers that vegetable is subjected to heavy application of industrial chemicals such as pesticides in the farming process should not be ignored. As people are becoming more motivated to consume vegetables, there is a need to build strong trust over the safety of vegetable produced. Extension programs should be designed to equip farmers with necessary extension education on safe vegetable production and at the same time, bring to the market vegetables that are free from hazardous chemicals for the safety of consumers. Different farming methods such as the use of integrated pest management practices through biological control measures, application of organic manure and crop rotation can be used as alternatives to the inorganic fertilisers and pesticides. Furthermore, research on quality and safety for the vegetables is needed to determine the level of toxins in vegetables and their health effects on human being. 37 | P a g e REFERENCES Agudo, A. and Joint FAO (2005) ‘Measuring intake of fruit and vegetables’, World Health Organization, p. 40. doi: 10.1017/CBO9781107415324.004. Ali, M and Tsou, S. (1997) “Combating micronutrients deficiency through vegetables-a neglected food frontier in Asia.” Food policy, Feb 1997, Vol (1), pp.17-38. Available at: https://doi.org/10.1016/S0306-9192(96)00029-2 (Accessed: 12 June 2017) Anderson. D, R, Sweeny, J, D, Williams, T, A, Freeman, J and Shoesmith, E. (2007). Statistics for Business and Economics. Thomson Learning Arora, M., Kiran, B., Rani, S., Rani, A., Kaur, B. and Mittal, N. (2008) ‘Heavy metal accumulation in vegetables irrigated with water from different sources’, Food Chemistry. Available at: http://www.sciencedirect.com/science/article/pii/S0308814608005013 (Accessed: 9 May 2017). Assema, P. Van, Brug, J. and Ronda, G. (2002) ‘A short dutch questionnaire to measure fruit and vegetable intake: relative validity among adults and adolescents’, Nutrition and. Available at: http://journals.sagepub.com/doi/abs/10.1177/026010600201600203 (Accessed: 9 May 2017). Bahemuka, T. E. and Mubofu, E. B. (1999) ‘Heavy metals in edible green vegetables grown along the sides of the Sinza and Msimbazi rivers in Dar es Salaam, Tanzania’, Food Chemistry, 66(1), pp. 63–66. doi: 10.1016/S0308-8146(98)00213-1. Bekunda, M. (2014). Research in Sustainable Intensification in the sub-humid maize-based cropping systems of Babati: Providing alternative integrated technologies to improve food security and income; http://africa-rising.wikispaces.com/file/view/BabatiProposal2014-16(2).docx Blanchette, L. and Brug, J. (2005) ‘Determinants of fruit and vegetable consumption among 6–12‐ year‐old children and effective interventions to increase consumption’, Journal of Human Nutrition and Dietetics. Available at: http://onlinelibrary.wiley.com/doi/10.1111/j.1365- 277X.2005.00648.x/full (Accessed: 9 May 2017). Browning, M., Chiappori, P.-A. and Lechene, V. (2004) ‘Collective and Unitary Models: A Clarification’, Review of Economics of the Household. Kluwer Academic Publishers, 4(1), pp. 1–17. Available at: http://www.econ.ku.dk/cam/wp0910/wp0203/2004-15.pdf. ( Verified: 22 June 2017) 38 | P a g e Bui, T. Van, Blizzard, C. L., Luong, K. N., Truong, N. L. Van, Tran, B. Q., Otahal, P., Srikanth, V., Nelson, M. R., Au, T. B., Ha, S. T., Phung, H. N., Tran, M. H., Callisaya, M., Smith, K. and Gall, S. (2016) ‘Fruit and vegetable consumption in Vietnam, and the use of a “standard serving” size to measure intake’, British Journal of Nutrition, 116(1), pp. 149–157. doi: 10.1017/S0007114516001690. Christine Hotz and Rosalind S Gibson (2007). Traditional food processing and preparation practices to enhance the bioavailability of micronutrients in plant-based diets. The Journal of Nutrition vol. 137 issue 4 (2007) pp: 1097-100 Published by American Society for Nutrition. Available at: http://jn.nutrition.org/cgi/content/short/137/4/1097. (Accessed: 12 June 2017) Darmon N, Drewnowski, A (2008). Does social class predict diet quality? Am J Clinical Nutrition 2008. The American Journal of Clinical Nutrition, Vol.87 (5), p.1107- Peer Reviewed Journal Deaton, A. and Muellbauer, J. (1980) Economics and consumer behaviour. Cambridge University Press. Di Cesare M, Khang YH, Asaria P, Blakely T, Cowan MJ, Farzadfar F, Guerrero R, Ikeda N, Kyobutungi C, Msyamboza KP, Oum S, Lynch JW, Marmot MG, Ezzati M (2013) Inequalities in non-communicable diseases and effective responses. The Lancet, Feb 16, 2013, Vol.381 (9866), p.585 (13) - Peer Reviewed Journal. Available at: http://www.sciencedirect.com.ep.fjernadgang.kb.dk/science/article/pii/S0140673612618510. (Accessed: 12 June 2017) Ecker, O., Weinberger, K., and Qaim, M. (2010). Patterns and determinants of dietary micronutrient deficiencies in rural areas of East Africa. African Journal of Agricultural and Resource Economics, June 2010, Vol.04 (2) pp.175-194. Available at: https://www-cabdirect- org.ep.fjernadgang.kb.dk/cabdirect/FullTextPDF/2011/20113072999.pdf. (Accessed: 12 June 2017) European Food Information Council, EUFIC, (2006). Determinants of Food Choice, A Review. Available at: http://www.eufic.org/en/healthy-living/article/the-determinants-of-food-choice. (Accessed: 12 June 2017) Everaarts, A. P., Putter, H. De and Maerere, A. P. (2015) ‘Profitability, labour input, fertiliser application and crop protection in vegetable production in the Arusha region, Tanzania’, p. (No. 653). PPO-AGV. 39 | P a g e Ezzati, M., Lopez, A.D., Rodgers, A., van der Hoorn, S., Murray, C.J.L., (2002). Selected major risk factors and global and regional burden of disease. The Lancet, Volume 360, Issue 9343, 2 November 2002, Pages 1347-1360. Field, A. E., Colditz, G. A., Fox, M. K., Byers, T., Serdula, M., Bosch, R. J. and Peterson, K. E. (1998) ‘Comparison of 4 questionnaires for assessment of fruit and vegetable intake.’, American Journal of Public Health, 88(8), pp. 1216–1218. doi: 10.2105/AJPH.88.8.1216. Gibson, E., Wardle, J. and Watts, C. (1998) ‘Fruit and vegetable consumption, nutritional knowledge and beliefs in mothers and children’, Appetite. Available at: http://www.sciencedirect.com/science/article/pii/S0195666398901805 (Accessed: 9 May 2017). Hall, J. N., Moore, S., Harper, S. B. and Lynch, J. W. (2009) ‘Global Variability in Fruit and Vegetable Consumption’, American Journal of Preventive Medicine. American Journal of Preventive Medicine, 36(5), p. 402–409.e5. doi: 10.1016/j.amepre.2009.01.029. Hellevik, O. (2009) ‘Linear versus logistic regression when the dependent variable is a dichotomy’, Quality and Quantity. Springer Netherlands, 43(1), pp. 59–74. doi: 10.1007/s11135-007-9077-3. Hlavac, Marek (2015). Stargazer: Well-Formatted Regression and Summary Statistics Tables. R package version 5.2. http://CRAN.R-project.org/package=stargazer IARC-International Agency for Research on Cancer (IARC) Handbook of Cancer Prevention. Volume 8: Fruit and Vegetables. Lyon, France: IARC Press, 2003. Ihucha, A. (2011) ‘Low fruit and vegetable intake killing East Africans, study’, The EastAfrican, pp. 1–2. Available at: http://www.theeastafrican.co.ke/news/Low-fruit--veg-intake-killing-East- Africans--study-/-/2558/1264416/-/item/1/-/v4f2eqz/-/index.html. Jeffrey, M. Wooldridge (2013). Introductory Econometrics. A Modern Approach. 5th Edition. Library of Congress, ISBN-13: 978-1-111-53104-1 Keding, G., Msuya, J., Maass, B. and Krawinkel, M. (2012) ‘Relating dietary diversity and food variety scores to vegetable production and socio-economic status of women in rural Tanzania’, Food Security. Available at: http://link.springer.com/article/10.1007/s12571-011-0163-y (Accessed: 14 May 2017). 40 | P a g e Keller, A., de Courten, M. and Dræbel, T. A. (2012) ‘Fruit and vegetable consumption and prevalence of diet-related chronic non-communicable diseases in Zanzibar, Tanzania: a mixed methods study’, The Lancet. Elsevier Ltd, 380, p. S16. doi: 10.1016/S0140-6736(13)60302-5. Kjøllesdal, M., et al., (2016) ‘Consumption of fruits and vegetables and associations with risk factors for non-communicable diseases in the Yangon region of Myanmar: a cross-sectional study’, BMJ Open, 6(8), p. e011649. doi: 10.1136/bmjopen-2016-011649. Kothari, C. R. (2004) Research Methodology: Methods & Techniques, New Age International (P) Ltd. doi: 10.1017/CBO9781107415324.004. Kristjansdottir, A. and Thorsdottir, I. (2006) ‘Determinants of fruit and vegetable intake among 11- year-old schoolchildren in a country of traditionally low fruit and vegetable consumption’, International. Available at: https://ijbnpa.biomedcentral.com/articles/10.1186/1479-5868-3-41 (Accessed: 9 May 2017). Lee, A. (2016) ‘Affordability of fruits and vegetables and dietary quality worldwide’, The Lancet Global Health. The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY license, 4(10), pp. e664–e665. doi: 10.1016/S2214-109X(16)30206-6. Liu, R. H. (2013) ‘Health-Promoting Components of Fruits and Vegetables in the Diet’, Advances in Nutrition: An International Review Journal. American Society for Nutrition, 4(3), p. 384S–392S. doi: 10.3945/an.112.003517. Lock, K., Pomerleau, J. and Causer, L. (2005) ‘The global burden of disease attributable to low consumption of fruit and vegetables: implications for the global strategy on diet’, Bulletin of the World. Available at: http://www.scielosp.org/scielo.php?pid=S0042- 96862005000200010&script=sci_arttext (Accessed: 9 May 2017). Locke, E., Coronado, G. D., Thompson, B. and Kuniyuki, A. (2009) ‘Seasonal Variation in Fruit and Vegetable Consumption in a Rural Agricultural Community’, Journal of the American Dietetic Association, 109(1), pp. 45–51. doi: 10.1016/j.jada.2008.10.007. Mattila-wiro, P. (1999) ‘Economic theories of household : A critical review April 1999 Working Papers No . 159’. 41 | P a g e Mayen, A.-L., Marques-Vidal, P., Paccaud, F., Bovet, P. and Stringhini, S. (2014) ‘Socioeconomic determinants of dietary patterns in low- and middle-income countries: a systematic review’, American Journal of Clinical Nutrition. American Society for Nutrition, 100(6), pp. 1520–1531. doi: 10.3945/ajcn.114.089029. Mazengo, M. C., Simell, O., Lukmanji, Z., Shirima, R. and Karvetti, R. L. (1997) ‘Food consumption in rural and urban Tanzania.’, Acta Tropica, 68, pp. 313–326. doi: 10.1016/S0001-706X(97)00113- 7. Miller, V. et al., (2016) ‘Availability, affordability, and consumption of fruits and vegetables in 18 countries across income levels: findings from the Prospective Urban Rural Epidemiology (PURE) study’, The Lancet Global Health, 4(10), pp. e695–e703. doi: 10.1016/S2214-109X(16)30186-3. Minot, Nicholas W. (1998). Distributional and Nutritional Impact of Devaluation in Rwanda. Economic Development and Cultural Change January 1998, Vol.46(2), pp.379-402- Peer Reviewed Journal. doi:10.1086/452343. Available at: http://www.jstor.org.ep.fjernadgang.kb.dk/stable/10.1086/452343 Nayga, R.M. (1995) “Determinants of U.S household expenditures on fruit and vegetables: a note and update.” Journal of Agriculture and Applied Economics vol 27(2): pp. 588-04 Narayan, D. and Pritchett, L. (1999) ‘Cents and Sociability: Household Income and Social Capital in Rural Tanzania’, World Bank, 47(4), pp. 871–897. doi: 10.1086/452436. Ngowi, A. V. F., Mbise, T. J., Ijani, A. S. M., London, L. and Ajayi, O. C. (2007) ‘Smallholder vegetable farmers in Northern Tanzania: Pesticides use practices, perceptions, cost and health effects’, Crop Protection, 26(11), pp. 1617–1624. doi: 10.1016/j.cropro.2007.01.008. Njelekela, M., Sato, T., Nara, Y., Miki, T. and Kuga, S. (2003) ‘Nutritional variation and cardiovascular risk factors in Tanzania-rural-urban difference: original article’, South African Medical. Available at: https://journals.co.za/content/m_samj/93/4/EJC67910 (Accessed: 9 May 2017). Ochieng, J., Afari-Sefa, V., Karanja, D., Kessy, R., Rajendran, S., & Samali, S. (2017). How promoting consumption of traditional African vegetables affects household nutrition security in Tanzania. Renewable Agriculture and Food Systems, 1-11. https://doi.org/10.1017/S1742170516000508 42 | P a g e Othman, O. (2001) ‘Heavy metals in green vegetables and soils from vegetable gardens in Dar es Salaam, Tanzania.’, Tanzania Journal of Science. doi: 10.4314/tjs.v27i1.18334. Padrão, P., Laszczyńska, O., Silva-Matos, C., Damasceno, A. and Lunet, N. (2012) ‘Low fruit and vegetable consumption in Mozambique: results from a WHO STEPwise approach to chronic disease risk factor surveillance.’, The British journal of nutrition, 107(3), pp. 428–35. doi: 10.1017/S0007114511003023. Pomerleau, J., Lock, K., Knai, C. and McKee, M. (2005) ‘Interventions designed to increase adult fruit and vegetable intake can be effective: a systematic review of the literature.’, The Journal of nutrition. American Society for Nutrition, 135(10), pp. 2486–95. Available at: http://www.ncbi.nlm.nih.gov/pubmed/16177217 (Accessed: 9 May 2017). Pomerleau, J., Lock, K. and McKee, M. (2004) ‘The challenge of measuring global fruit and vegetable intake’, The Journal of. Available at: http://jn.nutrition.org/content/134/5/1175.short (Accessed: 9 May 2017). Rasmussen, M., Krølner, R., Klepp, K.-I., Lytle, L., Brug, J., Bere, E. and Due, P. (2006) ‘Determinants of fruit and vegetable consumption among children and adolescents: a review of the literature. Part I: Quantitative studies.’, The international journal of behavioural nutrition and physical activity, 3(1), p. 22. doi: 10.1186/1479-5868-3-22. (Accessed: 9 May 2017). Ruel, Marie T., M. N. and S. L. (2005) ‘Patterns and determinants of fruit and vegetable consumption in sub-Saharan Africa: a multicountry comparison’, Joint FAO/WHO workshop on fruit and vegetables for Health, p. 45. Available at: http://www.who.int/entity/dietphysicalactivity/publications/f&v_africa_economics.pdf?ua=1 (Accessed: 10 May 2017). Richard Williams, University of Notre Dame, (2017). Marginal Effects for Continuous Variables. Revised January 2017. Available at: https://www3.nd.edu/~rwilliam/stats3/Margins02.pdf. (Accessed: 22 June 2017) Satheannoppakao, W., Aekplakorn, W. and Pradipasen, M. (2009) ‘Fruit and vegetable consumption and its recommended intake associated with sociodemographic factors: Thailand National Health Examination Survey III’, Public Health Nutrition, 12(11), pp. 2192–2198. doi: 10.1017/S1368980009005837. (Accessed: 10 May 2017). 43 | P a g e Smith-Warner, S. A., Elmer, P. J., Fosdick, L., Tharp, T. M. and Randall, B. (1997) ‘Reliability and comparability of three dietary assessment methods for estimating fruit and vegetable intakes.’, Epidemiology (Cambridge, Mass.), 8(2), pp. 196–201. doi: 10.1097/00001648-199703000-00013. (Accessed: 10 May 2017). Smith, L. C., Ramakrishnan, U., Ndiaye, A., Haddad, L. and Martorell, R. (2003) The importance of women’s status for child nutrition in developing countries, Food and Nutrition Bulletin. TFNC (2014) ‘Ministry of Health and Social Welfare Tanzania National Nutrition Survey 2014’, (November). The East African (2011). Low Fruits and Vegetable Intake Killing East Africans, Study. Available at: http://www.theeastafrican.co.ke/news/Low-fruit--veg-intake-killing-East-Africans--study-/2558- 1264416-h737bjz/index.html. (Verified: 22 June 2017) Tserendejid, Z., Hwang, J., Lee, J. and Park, H. (2013) ‘The consumption of more vegetables and less meat is associated with higher levels of acculturation among Mongolians in South Korea’, Nutrition Research. Elsevier Inc., 33(12), pp. 1019–1025. doi: 10.1016/j.nutres.2013.09.001. (Accessed: 10 May 2017). The United Republic of Tanzania-URT, MUCHALI (2011) ‘Integrated Food Security and Nutrition Assessment Report of the 2009 / 10 Season for the Market Year 2010 / 2011. Prepared by "Mfumo wa Uchambuzi wa Uhakika wa Chakula na Lishe-MUCHALI". Available at: http://www.ipcinfo.org/fileadmin/user_upload/ipcinfo/docs/1_IPC_TZA_Nutrition%20Assessemen t_10_10_Report.pdf. (Accessed: 21 June 2017) UNICEF--Tanzania. (2017). Overview. Available at https://www.unicef.org/tanzania/nutrition.html (Verified: 22 June 2017) United Republic of Tanzania-URT (2008). National Road Map Strategic Plan 2008-2015to Accelerate Reduction of Maternal, Newborn and Child Deaths in Tanzania. Ministry of Health and Social Welfare. Available at: http://www.who.int/pmnch/countries/tanzaniamapstrategic.pdf. (Accessed: 12 June 2017) United Republic of Tanzania-URT (2017). Tanzania-Household Budget Survey 2011-2012. Key findings. Available at: http://www.nbs.go.tz/nbstz/index.php/english/statistics-by-subject/household- budget-survey-hbs/367-household-budget-survey-main-report-2011-12. (Accessed: 12 June 2017) 44 | P a g e Uusiku, N., Oelofse, A., Duodu, K., Bester, M., and Faber, M. (2010). Nutritional value of leafy vegetables of sub-Saharan Africa and their potential contribution to human health: A review. Journal of Food Composition and Analysis, 2010 Sep, Vol.23 (6), pp.499-509- Peer Reviewed Journal. Available at: https://doi-org.ep.fjernadgang.kb.dk/10.1016/j.jfca.2010.05.002. (Accessed: 12 June 2017) Valmórbida, J. L. and Vitolo, M. R. (2014) ‘Factors associated with low consumption of fruits and vegetables by preschoolers of low socio-economic level.’, Jornal of paediatric. Elsevier Editorial Ltd, 90(5), pp. 464–71. Available at: doi: 10.1016/j.jped.2014.02.002. (Accessed: 12 June 2017) Wagner, K. H. and Brath, H. (2012) ‘A global view on the development of non-communicable diseases’, Preventive Medicine, pp. S38–S41. doi: 10.1016/j.ypmed.2011.11.012. Weinberger, K. and Swai, I. (2006) ‘Consumption of traditional vegetables in Central and Northeastern Tanzania’, Ecology of Food and Nutrition. Available at: http://www.tandfonline.com/doi/abs/10.1080/03670240500530626 (Accessed: 9 May 2017). World Health Organization (2004) ‘Fruit and vegetable promotion initiative/a meeting report’, Report of the meeting, p. 29. doi: http://www.who.int/dietphysicalactivity/publications/f&v_promotion_initiative_report.pdf. World Health Organization. The WHO STEPwise approach to non-communicable disease risk factor surveillance (STEPS). Available at: http://www.who.int/chp/steps/instrument/STEPS_Instrument_V3.1.pdf?ua=1. (Accessed: 12 June 2017) World Health Organization (2013). 10 Facts on Non-communicable diseases. Available at: www.who.int/features/factfiles/noncommunicable_diseases/en/. (Accessed: 21 June 2017) Worsley, A., Wang, W. C. and Farragher, T. (2016) ‘The associations of vegetable consumption with food mavenism, personal values, food knowledge and demographic factors’, Appetite. Elsevier Ltd, 97, pp. 29–36. doi: 10.1016/j.appet.2015.11.005. Wright, J., Sherriff, J., Mamo, J., Scott, J. (2015) Validity of Two New Brief Instruments to Estimate Vegetable Intake in Adults. Nutrients 2015, 7(8), 6688-6699; doi:10.3390/nu7085305. (Accessed: 12 June 2017) 45 | P a g e Appendix: Questionnaire Enumerator: [Greetings!], my name is [enumerator name]. I am a researcher and we are conducting a study to better understand the consumption of vegetables in your household. This study is for research purpose and we are collaborating with the World Vegetable Research Center under Africa RISING Program being implemented in Babati District. The information you give is very important to make this research successful. I encourage you to answer the questions trustfully and to the best of your knowledge. I would like to assure you that your information will only be used for study purpose and it will be treated confidential and NEVER disclosed to the public. Are you ready to participate in this interview? Yes No PART A: General Information Enumerator: Make sure the respondent is the household head or household head partner/spouse. If both of them are unavailable, please proceed to another household. A1 Consent read and obtained Tick the box A2 Enumerator name A3 Date of interview A4 HH identification number A5 Village identification number 1. Matufa 2. Galapo 3. Shaurimoyo 4. Berni 5. Seloto 6. Himiti 7. Managha 8. Majengo A6 Village location 1. Rural 2. Urban A7 Name of the respondent A8 Position of the respondent in the household 1. Household head 2. Partner of the household head A9 Gender of respondent 1. Female 2. Male 46 | P a g e Part B: Demographic and Socioeconomic information of the household B1: Household size, composition, education and occupation Relationship codes: 1. Head 2. Household partner/spouse 3. Daughter/Son 5. Grandchildren 6. Nephew/Nice 7. House maid 8. Mother/Father in law 9. Sister/Brother in Law 10. No family relationship; Codes for Occupation: 1. Employed 2. Farmer 3. Self-employed 4. Student 5. Retired officer 6. Casual worker 7. Unemployed; Codes for highest education level completed: 1. Completed primary school education 2.Ordinary secondary school level 3. Advanced secondary school level 4. College 5 University degree 6. No formal schooling Membe r ID Name of HH member Relati onship to the house hold head See codes below Age (in years) Indica te 0 less than one year Sex? (1. Female 2. Male) Educatio n level in years The highest level of formal education complete d? See codes below Main occupation? See codes below 1 2 3 4 5 6 7 8 9 10 11 12 47 | P a g e B2 Household income B3: Household expenditure Please consider the last month income and expenditure: How much did your household spend on the items below from the monthly average income? Income ID Please consider the past year and Tick all that apply Reliable income sources Average monthly income(TZS) 1 Wages or salary from regular job 2 Wages from casual labour (farm or non-farm) 3 Selling crop products 4 Selling livestock products 5 Grants/Pension or subsidy of some sort 6 Support from friends or another family 7 Running business 8 Remittances within Tanzania 9 Remittances from abroad 10 Others(specify) [For the analysis use only] TOTAL AVERAGE MONTHLY INCOME ID Item Average monthly expenditure (TZS) Percentage of monthly income identified in B2 spent on this item (%) 1 Food-All types 2 All non-food items (housing rent, home appliances, school fees and stationaries, clothes, medical, transport, etc.) 48 | P a g e PART C: Household food expenditure: Consider food expenditure in the last week Codes for possible reasons of not consuming the food item: 1. was not available 2. Could not afford/did not have money 3. Not preferred by most of the household members. PART D: Vegetable consumption in the household: [Enumerator] Consider vegetables consumed in last week and let respondent identify all kinds of vegetables consumed) Food item ID Food group Did you consume? 1. Yes 2. No If consumed, how much you spent in the last week? (TZS) If No, what is the MAIN reason? See codes for possible reasons 1 Vegetables-All types 2 Meat-All types 3 Beans-All type 4 Cereal and Carbohydrate related-All type 5 Fruits 6 Drinks-All types Veg ID Identify all types of vegetable in your household in the last week [Enumerator ]: Tick all types of vegetables consumed Did young childre n 2 years old consum e this? 1. Y e s 2. N o How many household members above 2 years old consumed in the last week? How many times in the last week your household consumed this vegetable? [Enumerat or]: probe on the number of intakes/me als the household consumed vegetable in last week. On average, how many market- size bunches of this vegetable your househol d normally purchase/ source for a single meal/inta ke? Total amount (in grams) consumed in the last week: [Enumerato r multiply the number of market- size bunches consumed per intake by the average weight of a single bunch and then by the total number of intakes per week. 1 Amaranth (Mchicha) 2 African eggplant(Ngogwe) 49 | P a g e PART E: Consumption preference E1 Generally, which vegetable among the listed in part D is the MOST preferred in your household [Enumerator] Please indicate the ID number of the vegetable E2 Which dish does your household usually prefer as a companion for vegetables? [Enumerator] Please circle one that is preferred by the most members 1. Stiff porridge (Ugali) 2. Rice 3. Banana 4. Ugali and meat 5. Rice and meat 6. Others___________ E3 Do your household consume dried vegetables? 1. Yes 2. No E3b If Yes, which type of vegetables consumed dried? Enumerator: probe the vegetable and refer the list in PART D and write the ID/code of the respective vegetable. 3 Night shade (Mnavu) 4 Spider flower (Mgagani) 5 Cowpea leaves (Majani ya kunde) 6 Ethiopian Mustad (Sukuma wiki/Loshu) 7 Chinese 8 Jute Mallow (mlenda) 9 Cassava leaves (kisamvu) 10 Pumpkin leaves (Majani maboga) 11 Sweet potato leaves (matembele) 12 Okra (bamia) 13 Cabbage (Kabichi) 14 Eggplant(Biringanya) 15 Black jack (Shomanguo) 16 Broccoli 17 Spinach 18 Baobab leaves 19 Moringa leaves (Mlenda) 20 Tomato 21 Carrots 22 Others (specify) 50 | P a g e E4 Do your household members mostly prefer to eat vegetables prepared at home or away from home? 1. At home 2. Away from home E4b What are the reason (s) for the choice above? Circle all that apply 1. Convenience 2. Distrust over the other option 3. Lack of time to prepare at home 4. Household size 5. Cheaper option 6. Other: Please specify___________ E5 When do buying/sourcing vegetables what aspects you mostly look at? Circle all that apply 1. Quality of the vegetable 2. The type/variety of vegetable 3. Origin of the variety (tradditional vs imported) 4. Price 5. Other factors___________ E6 Why your household eat vegetables? Circle all that apply 1. Health and nutritional benefit 2. Vegetable is cheaper and we cannot afford other options 3. We like vegetables 4. Cannot easily access other options in our area 5. Others: please specify___________ E6b Which factor among listed in E6 is the MOST important? PART F: Awareness and knowledge related to consumption of healthier foods including vegetables F1 Has any member of your household participated in any training/awareness campaign about nutrition and healthy eating? 1. Yes 2. No F1b If Yes, what were the sources provided the training/sensitization? Tick all that applicable 1. AVRDC-Africa RISING project 2. Government's community health officers 3. Non-government health activist 4. Other NGOs 5. Health centres 6. School programs F2 Is any member of your household aware of AFRICA RISING project promoting vegetable production and consumption in Babati? 1. Yes 2. No 51 | P a g e F2b If YES to F2, from which source of information did you hear about the project? 1. Agricultural extension officer 2. One of the Africa RISING beneficiary 3. Field days/Agricultural shows(Nane Nane) 4. Other farmers/friends 5. Broadcast media (Radio/TV) 6. Print media (newspaper/brochures) 7. Project officer PART G: Vegetable Access, production and perception on use of inorganic methods in vegetable farming G1 From which sources your household normally access the vegetables? 1. Local market 2. Wild vegetables 3. From neighbours 4. Exchange for other commodities with neighbours 5. On vegetable garden G2 What is the distance (in kilometres) from where you live to the nearest source? G3 It is said that some vegetable sold in the markets are grown under heavy application of industrial fertilisers and pesticide. Is your household concerned about this? 1. Yes 2. No G3b If YES to G3, Was there a moment where you decided not to buy certain vegetable because of the feeling that it was not produced organically? 1. Yes 2. No G4 What is your opinion on the safety of vegetables produced under application of chemicals such as pesticides and industrial fertiliser? 1. It is safe 2. It is not safe 3. I don't know G5 Over the past month, did you grow vegetable at home? 1. Yes 2. No G6 If Yes to G5, is the most of the vegetable consumed in your household in the past month sourced from your own garden? 1. Yes 2. No 52 | P a g e G7 Generally, how would you rate the availability of vegetables in your area? 1. High but seasonal 2. High and throughout the year 3. Low and seasonal 4. Low but throughout the year 5. Very scarce THANK YOU FOR YOUR PARTICIPATION IN THIS SURVEY!