Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Household food production is positively associated with dietary diversity and intake of nutrient-dense foods for older preschool children in poorer families: Results from a nationally-representative survey in Nepal

  • Prajula Mulmi,

    Roles Conceptualization, Formal analysis, Methodology, Writing – original draft, Writing – review & editing

    Affiliation Friedman School of Nutrition Science & Policy, Tufts University, Boston, United States of America

  • William A. Masters ,

    Roles Conceptualization, Formal analysis, Investigation, Methodology, Supervision, Writing – original draft, Writing – review & editing

    Affiliation Friedman School of Nutrition Science & Policy, Tufts University, Boston, United States of America

  • Shibani Ghosh,

    Roles Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Supervision

    Affiliation Friedman School of Nutrition Science & Policy, Tufts University, Boston, United States of America

  • Grace Namirembe,

    Roles Data curation, Formal analysis

    Affiliation Friedman School of Nutrition Science & Policy, Tufts University, Boston, United States of America

  • Ruchita Rajbhandary,

    Roles Data curation, Investigation

    Affiliation Center for Human Nutrition, Department of International Health, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, United States of America

  • Swetha Manohar,

    Roles Data curation, Investigation, Methodology, Project administration, Supervision, Validation

    Affiliation Center for Human Nutrition, Department of International Health, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, United States of America

  • Binod Shrestha,

    Roles Data curation, Formal analysis, Investigation

    Affiliation Center for Human Nutrition, Department of International Health, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, United States of America

  • Keith P. West,

    Roles Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Writing – review & editing

    Affiliation Center for Human Nutrition, Department of International Health, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, United States of America

  • Patrick Webb

    Roles Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Writing – review & editing

    Affiliation Friedman School of Nutrition Science & Policy, Tufts University, Boston, United States of America



Nutrition-sensitive interventions supporting enhanced household food production have potential to improve child dietary quality. However, heterogeneity in market access may cause systematic differences in program effectiveness depending on household wealth and child age. Identifying these effect modifiers can help development agencies specify and target their interventions.


This study investigates mediating effects of household wealth and child age on links between farm production and child diets, as measured by production and intake of nutrient-dense food groups.


Two rounds (2013 and 2014) of nationally representative survey data (n = 5,978 observations) were used to measure production and children’s dietary intake, as well as a household wealth index and control variables, including breastfeeding. Novel steps used include measuring production diversity in terms of both species grown and food groups grown, as well as testing for mediating effects of family wealth and age of child.


We find significant associations between child dietary diversity and agricultural diversity in terms of diversity of food groups and of species grown, especially for older children in poorer households, and particularly for fruits and vegetables, dairy and eggs. With each additional food group produced, log-odds of meeting minimum dietary diversity score (≥4) increase by 0.25 (p = 0.01) for children aged 24–59 months. For younger children aged 18–23 months there is a similar effect size but only in the poorest two quintiles of household wealth, and for infants 6–18 months we find no correlation between production and intake in most models.


Child dietary intake is associated with the composition of farm production, most evident among older preschool children and in poorer households. To improve the nutrition of infants, other interventions are needed; and for relatively wealthier households, own farm production may displace market purchases, which could attenuate the impact of household production on child diets.


Children are particularly vulnerable to malnutrition [1,2,3], and stunted growth is especially widespread in poorer and more remote rural households [3,4]. Effects may include negative influences on cognition and health of individuals, and may more broadly adversely affect a national economy [2].

Optimal infant and young child feeding (IYCF) practices include ensuring intake of diverse food groups [5,6], because low diversity is closely linked to inadequate or poor quality of diets [7]. The World Health Organization (WHO) recommends that children under 24 months consume a minimum of 4 out of 7 food groups every day, to raise the probability that a child eats food from both plant and animal sources and promote intake of bioavailable iron, zinc, calcium, and vitamin A among other nutrients [4,5,6,8].

Intake below minimum dietary diversity is widespread in South Asia. In Nepal, for example, the share of children meeting minimum dietary diversity levels increases with age, but from only 12% of infants aged 6–8 months to 22% at 9–11 months, 33% for 12–17 months, and 37% for 18–23 months [9]. This age gradient could be related to differences in child feeding practices, such as older children relying on family meals while younger infants require more frequent feeding with nutrient-dense foods [5,6]. Maternal and child intake of diverse nutrient-dense food groups has been found to be associated with their production on the family’s farm [10,11]. As such, interventions to increase the diversity of farm production have been widely proposed as a path to improved nutrition [12,13,14,15], especially for more remote households who have less ability to use markets to complement their own production with purchased foods [16].


This observational study aims to explore associations between agricultural production diversity and preschool-aged children’s dietary diversity. The rich empirical data used allow for careful consideration of heterogeneity in household wealth and child’s age as potential confounders in such relationships [17]. Both child age and household purchasing power would be expected to modify effects of a household’s own agricultural production on their child’s intake. In economic models of farm households [18,19,20,21], families allocate their available time to caregiving within the household and food production on the farm, using markets to buy and sell whatever products can be exchanged to achieve consumption goals. An important finding from that literature is that food production can potentially be replaced by market purchases, especially in households with sufficient purchasing power and market access to do so [22]. Another significant point is that caregiving within the household, particularly for mothers and infants, requires time that cannot be produced on farm and bought or sold in the market [23]. This suggests that own agricultural production is likely to be associated with child dietary intake only in poorer households without the purchasing power to access markets, with no such association for the youngest children whose dietary intake depends on specialized caregiving efforts in addition to food access.

The hypothesis tests presented here add to the literature on agriculture-nutrition linkages. In Nepal, about two-thirds (68%) of households work in the agricultural sector [24], and isolation, poverty and social conflict may prevent households from using markets to improve dietary intake [1]. This study is also closely linked to recent findings about the high cost of nutrient-dense foods in Nepal [25], and the demonstrated interest of the national government and district officials to achieve the country’s multisector nutrition plan to improve diet quality [26].

Our distinctive focus on disaggregation by age, as recommended by the World Health Organization [5], is made possible by the large sample size and repeat visits of the Policy and Science for Health, Agriculture and Nutrition (PoSHAN) surveys conducted in Nepal in 2013 and 2014 [27, 28] as part of a large portfolio of Feed the Future Nutrition Innovation Lab research in that country. As observed in the PoSHAN data, dietary diversity for younger children (6–18 months) is lower than that for older children (≥18 months) (Fig 1). This is consistent with the observation that labor-intensive introduction and frequent feeding of solid foods to younger children is often inadequate [6], while the diets of older children converge to those of adults in the household as soon as they can consume family foods at household meal times.

Fig 1. Child dietary diversity by child age (in month).

Box plots show median, 25th and 75th percentile of distributions, and diamonds represent mean for each age range.

The aim of this paper is to test the hypothesis that household food production diversity is associated with preschool child dietary intake only for older children in poorer households. PoSHAN data were used to determine each household’s agricultural output in terms of the same food groups as those used to measure child diets—this is an important detail that distinguishes this analysis from other studies that have used overall crop/livestock counts as the metric of an agricultural diversity that was unrelated to foods consumed in the same household. We also used counts of individual species of crops and livestock to capture agronomic diversity (to be able to compare with other studies), within as well as a between dietary food groups.


Dataset and study variables

This study uses two waves of panel data obtained from PoSHAN survey conducted in Nepal in 2013 and 2014. Our analysis was reviewed by the Tufts University Institutional Review Board for Social, Behavioral and Educational Research as IRB Study Number 1606018 and excluded from further review on June 14, 2016. Ethical clearance for primary data collection was obtained from the Institutional Review Boards of Johns Hopkins University (USA) and the Nepal Health Research Council (Kathmandu). The survey used multistage stratified cluster sampling to be nationally representative as described in Dorsey et al. (2017) [28]. Interviews occurred during May, June and July of each year, which is typically a period of intermediate food security between harvests [29]. Strata include three agro-ecological zones: Mountains, Hills and Terai, and clusters include seven Village Development Committee (VDCs) per ecological zone [27]. The final sample used in our main results is a balanced panel (n = 2,989 children and 5,978 observations) comprised of children between the ages of 6–59 months and their mothers observed in both years. For this study we exclude caretakers other than mothers, typically grandmothers and fathers, that make up a small subset of the sample (n = 164).

The primary outcome variable of the study is whether children (6–59 months) reach a minimum dietary diversity threshold of four out of seven food groups consumed in the previous week. A secondary set of outcome variables is whether children consume any food from each of seven WHO-defined food groups when that food is produced on the family’s farm. The WHO food groups are (1) grains, roots, and tubers; (2) vitamin-A rich fruits and vegetables; (3) other fruits and vegetables; (4) flesh foods including meat, fish, poultry; (5) eggs; (6) dairy; (7) legumes, nuts and seeds. The WHO’s recommendation is specifically for children under the age of two over the previous 24 hours [5], but that same indicator is used here for all children over 7 days so as to compare dietary trajectories as children age. All dietary data was obtained from a standardized 7-day food-frequency questionnaire, listing over 50 distinct foods that were known from formative research to be commonly or occasionally consumed in Nepal.

Our measures of agricultural production begin with a dichotomous indicator of whether households produced any food at all, and then the total number of food groups grown ranging from 1 to 7. This measure provides a direct counterpart to diversity of dietary intake. To capture agro-ecological diversity within as well as across food groups, we also count the number of distinct crop or livestock species grown. The distribution of that variable is such that many households raise just one or a few of them, while a few have highly diversified farms reporting more than 30 different species of crops and livestock. To transform this measure of agricultural diversity into a variable that might plausibly be associated with child dietary diversity we divide the sample of farmers into quintiles, from 1 to 5. All the variables representing agricultural production decisions are constructed using crops grown during both rainy and dry seasons and livestock raised year-round.

Wealth is measured by an asset index following the same method used by the Demographic and Health Surveys (DHS) [9]. This index enters our hypothesis tests linearly in quintiles of the national distribution. The mediating variable, child age, is used to split the sample to permit full flexibility in all estimated coefficients. Since the optimal IYCF practices only apply to children under two [6], we first divide the sample to compare those aged 6–23 months against those aged 24–59 months. The youngest are further divided into three groups as recommended by the WHO to compare 6–11, 12–17 and 18–23 months [5].

All results control for standard confounding, socio-demographic variables found in the literature. They include caste and ethnicity, religion, maternal education and BMI, mother and child age, whether a child was breastfed, ecological zones, and a total amount of land owned and rented [2,9,30,31,32]. Geographic differences such as distance to markets and disease environments are controlled using VDC fixed effects, and temporal differences such as climatic and market conditions are controlled using year fixed effects. The fixed effects model absorbs all village-specific, year-specific, altitude and climate variation. This approach enables all coefficients to be interpreted as differences among children within villages in any given year, and all analyses are performed using Stata/SE, version 14.

Estimation strategy

To quantify associations between agricultural production diversity and an indicator of minimum dietary diversity for children (MDDC), logit regression models are estimated using the following base specification: (1) (2) (3) where MDDCi indicates whether a child, i, in each of our age groups from 6 to 59 months achieves the minimum dietary diversity score, farmih is a dummy variable equal to 1 when h = household of a child i produces any food and zero otherwise, fgroupih is a variable ranging from 0–7 indicating number of food groups produced by that household, and fquintih represents quintile of number of crop and livestock species produced (0–5) by that household.

For all the three equations, the study tests the null hypothesis that β1, which represents coefficient on the three primary predictor variables, is not statistically significant across the five child age categories. β2 represents wealth quintile of each child i in household h. Preliminary analysis indicates that wealth quintile, food group, and agricultural production diversity behave in a linear fashion. These variables are, hence, included in the models in a linear form. β3 represents an interaction between the predictor variables and wealth, to test hypotheses regarding heterogeneity by the household’s purchasing power.

Zi accounts for a vector of control variables at the child, household, and VDC levels as shown in results (Tables 19). All the standard determinants of child dietary intake are controlled for in the regressions [33,34]. Block (2006) showed that maternal nutritional knowledge substitutes for maternal schooling in achieving better child nutrition outcomes [35]. Hence, the regression model assumes and accounts for the non-linear relationship between maternal schooling (in years) and child nutrition outcomes. All the logit regressions above include VDC fixed effects and year fixed effects. For each child, i, the error term is represented by μi. The unbalanced panel data from PoSHAN comprising a larger sample size (Tables 1011) is used to check the robustness of the results for Eqs 13.

Table 1. Diet diversity and number of food groups produced by child's age.

Table 2. Diet diversity and quintile of food species produced by child's age.

Table 3. Dietary diversity in farm vs. non-farm households by child's age.

Table 4. Household vitamin A-rich fruits and vegetables consumption and production.

Table 5. Household fruit and vegetable production and consumption.

Table 10. Diet diversity and number of food groups produced by child's age (unbalanced panel).

Table 11. Diet diversity and agricultural diversity quintile produced by child's age (unbalanced panel).

The mediating effect of child age on the association between production of individual food groups and its consumption is quantified using logit regression models. Almost all children consume starchy staples, so results are reported for the remaining six nutrient-dense food groups using the following base specification: (4)

The base specification is run for each of the six WHO-determined food groups, where Cnsmptni and prdctnih indicate binary variables. Cnsmptni indicates whether child i eats a food group and prdctnih represents whether household h of that child grows the same food group. Note that production and consumption may refer to different foods within that group, to allow for the possibility of exchanging one food for another within that group (for example, trading carrots for tomatoes) as well as self-provisioning of each food. As noted above, β1 is the primary coefficient of interest. β1 tests the null hypothesis that there is no difference in child’s intake of that food group between households that do and do not produce it. β3 shows the coefficient of the interaction between a food group produced and wealth quintile of household h of a child i. Coefficients are intended for hypothesis testing rather than estimation, but when significantly different from zero the magnitude of these log-odds ratios, when exponentiated, provide an estimate of the probability that a child’s intake includes that food group or meets the MDDS minimum. The remainder of the notations, functional forms and control variables are the same as those described in Eqs 13.


Agricultural production diversity and minimum dietary diversity score for children (MDDC)

Our principal finding is a significant association (p<0.001) between children reaching the minimum dietary diversity score (≥4 in the previous week) and the number of food groups grown in the household for all children over 24 months, and for children in the poorest households from 18–23 months of age, but not for children aged 6–23 months or in the 6–11 or 12–18 month groups (Table 1).

All else equal, for each additional food group that households produce, log-odds of meeting the minimum dietary diversity score (≥4) increases 0.253 (p = 0.01) among children between the ages of 24–59 months. In the smaller sample of younger children aged 18–23 months, there is also a significant effect but only for the poorest households: in the lowest quintile of wealth the effect is 0.43 minus 0.137 (both p = 0.01). At higher levels of household wealth and hence purchasing power, for children 18–23 months the association between achieving minimum dietary diversity and number of food groups grown on the family farm goes to zero—and it is zero at all levels of wealth for children under 18 of age, whose diets require more labor-intensive specialized feeding practices [9]. Each of the narrow age subgroups have relatively small sample sizes: n = 396 for 6–11 month-olds, and n = 399 for 12–17 month-olds. This is why we perform the same test for the larger 6–23 month group (n = 1,635) and also for the unbalanced sample that includes children observed in only one of the two years (n = 1,034 for 6–11 month-olds, and n = 934 for 12–17 month-olds) (Table 10).

Our alternative measure of production diversity counts each crop or livestock species separately to consider agroecological diversity even within food groups, and then transform the number of species grown into quintiles of the distribution for comparability with number of food groups or quintiles of wealth. Results are similar to our main findings, with agricultural diversity significantly associated with achieving minimum dietary diversity only after two years of age, and below that only in the poorer households (Table 2). For every unit increase in agricultural diversity quintile, log-odds of meeting minimum dietary diversity score (≥4) increases by 0.361 (p = 0.01) for children aged 24–59 months, while for children aged 18–23 months or 6–23 months the effect is significant only in the poorer households. The coefficient estimate for 6–11 months is statistically significant as well, but that sample is relatively small (n = 396). Further to our main results in Tables 1 and 2, we find that farming as such is not significantly associated with meeting minimum dietary diversity in any age category (Table 3). We also compare these results to the larger unbalanced panel, with attrition and replacement between rounds of the survey, using the household’s number of food groups produced (Table 10) and their quintile of agricultural diversity (Table 11).

Among control variables in addition to household wealth, the most consistently significant factors associated with achieving minimum dietary diversity are whether the child is still breastfeeding (-), the mother’s educational attainment (+), being from an excluded caste (-), and living in the Terai region (+), with additional significance in some regressions of the mother’s age (+), BMI (+) and having a female head of household (-). These findings are broadly similar to previous studies [36,37,38,39].

The negative association between breastfeeding and solid food intake arises primarily in older infancy, e.g. for the 18–23 month olds. In that age range, breastmilk and solid foods are likely to substitute for each other as sources of energy and nutrients [39,40,41], with diversity in IYCF practices due to differences in attitude, behaviors and perceptions on infant feeding [42,43].The positive association we find with maternal schooling is typical of previous work [9,44], as is the negative association with being from a socially excluded caste such as Dalit, Janajati, and Terai groups [31,45], except for those located in the Terai region which, controlling for the other factors mentioned above, is positively associated with meeting the minimum level of child dietary diversity. Without those controls, there is usually higher dietary diversity among children in Hill followed by Mountain and Terai [9,36].

Individual food group production and consumption

Our tests for associations between children’s intake of each food group and their production on their family’s farm are presented in Tables 49. Of the six food groups considered to be nutrient dense, other than starchy staples, we find no such associations for legumes (Table 9) or meat (Table 6), implying that these can more readily be purchased if consumed (which occurs frequently for legumes, and rarely for meat). The strongest association is for eggs (Table 7), whose log odds of consumption is 0.98–1.6 points higher (p<0.01) when produced at home, for children both above and below two years of age, with no significance in the small samples of 6–11 and 12–17 months of age, and no significant interaction with household wealth.

For fruits and vegetables considered to be vitamin-A-rich (i.e., sources of provitamin A carotenoids), other fruits and vegetables, and dairy, there are significantly higher odds of intake in households that produce these foods on the family farm, but only at lower levels of wealth (Tables 4, 5 and 8). These could be the incremental food groups that contribute to those households reaching the minimum dietary diversity level reported in previous tables. These results hold primarily for older children. For those between the ages of 6–11 months and 12–17 months, the coefficients on fruits and vegetables (Tables 4 and 5) and eggs (Table 7) are positive, but not statistically significant. For the small sample of children aged 6–11 months we find a negative association between intake and dairy production on the farm (Table 8), which could be an anomaly due to the small sample involved (n = 391).

Discussion and conclusions

Our principal finding is a strong association between the diversity of a household food production and dietary diversity among older children (18 or 24 months of age) in poorer households (the lowest one or two quintiles of wealth). Each additional food group produced is associated with 0.25 (p = 0.01) higher log-odds of meeting minimum dietary diversity score (≥4) by children aged 24–59 months. For younger children aged 18–23 months there is a similar effect but only in the poorest two quintiles of household wealth, and for infants 6–18 months we find little or no correlation between dietary intake and household production. Among the older children, the specific food groups for which intake is associated with production are eggs (at all wealth levels), and vitamin-A-rich fruits and vegetables, other fruits and vegetables, and dairy (in poorer households).

These findings control for geographic and climatic factors using village and year fixed effects, and also control for other variables that are significantly associated with intake notably the child’s breastfeeding status, the mother’s age, caste, educational attainment and BMI, and having a female head of household. By identifying differences among food groups and disaggregating by age we aim to help inform future interventions that could fill age-specific inadequacies in major food group dietary diversity across the preschool years in rural Nepal, to meet WHO recommendations regarding complementary feeding with foods such as dairy and eggs, as well as with nutrient-rich fruits and vegetables.

These findings suggest that interventions aimed at diversifying households’ farm production are likely to improve children’s diet quality only after they reach 18 months of age. To address poor nutrition among younger children may therefore require other actions such as targeted nutrition-specific interventions as well as improved market access to purchased complementary foods [46]. Even for older children, investments made in agricultural diversification are likely to be most effective if they target the poorest households least connected to food markets. We find that the upper three or four quintiles of Nepal’s income distribution, on average, have sufficient purchasing power for their market transactions to replace differences in what they grow. The exception in these data is eggs, for which the association between intake and production is not mediated by wealth.

Importantly, agricultural production data used here refer to crop and livestock raised at any time of year, while the dietary data refer to the previous seven days before the survey interview. As a result, many of the foods produced may be available only in particular seasons. This temporal limitation of household food production can sometimes be overcome by interventions that help households produce each food throughout the year, but local climate and resource constraints often make that prohibitively difficult. Since the timing of harvests usually differs over space, a more feasible approach to making intake more uniform over time is to facilitate market access. The effectiveness of food markets in overcoming local climate variation over time in Nepal is shown in Mulmi et al. (2016) [47]; similar effects of market access in smoothing nutritional outcomes has been shown for Africa in Darrouzet-Nardi and Masters (2017) [48].

A limitation of the study is the relatively small sample sizes for narrow age ranges of interest such as 6–11 months and 12–18 months. A potential extension of this work could focus on larger surveys with more children at each age. Another limitation is that our work addresses only the qualitative question of whether or not foods from a given group were consumed and produced. Future research should address the more complex question of how much of each food is consumed or produced. Another possible extension related to survey design is the potential to repeat surveys during the year to capture seasonal fluctuations, and to continue surveys for enough years for household and child fixed effects to be feasible, both of which would more fully exploit the panel nature of the survey.

Despite these limitations, the study provides clear insights regarding the heterogeneity in effects of home production on children’s intake of nutrient-dense foods, contributing to the rapidly growing evidence base on how and for whom agricultural interventions can best be used to improve child nutrition outcomes.


The authors thank all contributors to the Nutrition Innovation Lab including Liz Marino-Costello of Tufts University and Gerald Shively from Purdue University; Raman Shrestha, Abhighyna Bhattarai, Hari Krishna Shah, Dev Raj Gautam and Diplav Sapkota of the PoSHAN study team in Nepal; Deepak Thapa from the Nepal Technical Assistance Group and Hari Koirala from USAID/Nepal. We are particularly grateful for the work of more than 80 interviewers and anthropometrists, field supervisors, quality control officers, data entry and management staff and research supervisors as well as Sidhartha M Tuladhar, Executive Director of New Era (P.) Ltd., and support received from the Logistics Management Division at the Child Health Division and UNICEF. We also thank the administrative leadership of Rajkumar Pokharel, Chief, Nutrition Section, at the Child Health Division of the Department of Health Services, Ministry of Health and Population in the Government of Nepal. Ethical review in Nepal was conducted by the Nepal Health Research Council (Kathmandu).


  1. 1. Bageant E, Liu Y, Diao X. Agriculture-nutrition linkages and child health in the presence of conflict in Nepal. IFPRI Discussion Paper 1515 Washington, DC: International Food Policy Research Institute; 2016.
  2. 2. Black RE, Victora CG, Walker SP, Bhutta ZA, Christian P, de Onis M, et al. Maternal and child undernutrition and overweight in low-income and middle-income countries. The Lancet. 2013 Aug; 382(9890): 427–51.
  3. 3. UNICEF. Improving child nutrition: the achievable imperative for global progress. New York: United Nations Children’s Fund; 2013.
  4. 4. UNICEF. From the first hour of life. New York: United Nations Children’s Fund; 2016.
  5. 5. WHO. Complementary feeding: report of the global consultation. Geneva: World Health Organization; 2001.
  6. 6. WHO and UNICEF. Global strategy for infant and young child feeding. Geneva: World Health Organization; 2003.
  7. 7. Arimond M, Ruel MT. Dietary diversity is associated with child nutritional status: evidence from 11 Demographic and Health Surveys. J Nutr. 2004; 134(10): 2579–85. pmid:15465751
  8. 8. Byrd-Bredbenner C, Moe G, Beshgetoor D, Berning J. Wardlaw’s Perspectives in Nutrition. 9 edition. New York: McGraw-Hill; 2012. 976 p.
  9. 9. Ministry of Health and Population, New Era, ICF International Inc. Nepal Demographic and Health Survey 2011. Kathmandu, Nepal; 2012.
  10. 10. Malapit HJL, Kadiyala S, Quisumbing AR, Cunningham K, Tyagi P. Women’s Empowerment in Agriculture, Production Diversity, and Nutrition. IFPRI Discussion Paper 1313. Washington, DC: International Food Policy Research Institute; 2013.
  11. 11. Bellon MR, Ntandou-Bouzitou GD, Caracciolo F. On-Farm Diversity and Market Participation Are Positively Associated with Dietary Diversity of Rural Mothers in Southern Benin, West Africa. PLOS ONE. 2016 Sep 8; 11(9): e0162535. pmid:27606782
  12. 12. Alaofe H, Burney J, Naylor R, Taren D. Solar-Powered Drip Irrigation Impacts on Crops Production Diversity and Dietary Diversity in Northern Benin. Food Nutr Bull. 2016 Jun 1; 37(2): 164–75. pmid:27009089
  13. 13. Osei A, Pandey P, Nielsen J, Pries A, Spiro D, Davis D, et al. Combining Home Garden, Poultry, and Nutrition Education Program Targeted to Families With Young Children Improved Anemia Among Children and Anemia and Underweight Among Nonpregnant Women in Nepal. Food Nutr Bull. 2017 Mar; 38(1):49–64.
  14. 14. Darrouzet-Nardi AF, Miller LC, Joshi N, Mahato S, Lohani M, Rogers BL. Child dietary quality in rural Nepal: Effectiveness of a community-level development intervention. Food Policy. 2016 May; 61: 185–97.
  15. 15. Miller LC, Joshi N, Lohani M, Rogers B, Loraditch M, Houser R, et al. Community development and livestock promotion in rural Nepal: effects on child growth and health. Food Nutr Bull. 2014 Sep; 35(3): 312–26. pmid:25902591
  16. 16. Sibhatu KT, Krishna VV, Qaim M. Farm production diversity and dietary diversity in developing countries. Proceedings of the National Academy of Sciences (PNAS). 2015 112: 10657–10662.
  17. 17. Hjelm L, Mathiassen A, Wadhwa A. Measuring Poverty for Food Security Analysis: Consumption- Versus Asset-Based Approaches. Food Nutr Bull. 2016 Sep; 37(3):275–289.
  18. 18. Singh I, Squire L, Strauss J. Agricultural household models: extensions, applications, and policy. Baltimore: John Hopkins University Press;1986.
  19. 19. de Janvry A, Fafchamps M, Sadoulet E. Peasant Household Behaviour with Missing Markets: Some Paradoxes Explained. Econ J. 1991 Nov; 101(409): 1400.
  20. 20. Taylor EJ, Adelman I. Agricultural household models: genesis, evolution, and extensions. Rev Econ Househ. 2003; 1(1): 33–58.
  21. 21. Mendola M. Farm household production theories: a review of institutional and behavioral responses. Asian Dev Rev. 2007; 24(1): 49–68.
  22. 22. Eswaran M, Kotwal A. Credit as insurance in agrarian economies. J Dev Econ. 1989; 31(1): 37–53.
  23. 23. Bardhan P, Udry C. Development microeconomics. Oxford University Press; 1999.
  24. 24. Central Bureau of Statistics. Population monograph of Nepal. Government of Nepal, National Planning Comission Secretariat; 2014.
  25. 25. Biehl E, Klemm RDW, Manohar S, Webb P, Gauchan D, West KP. What Does It Cost to Improve Household Diets in Nepal? Using the Cost of the Diet Method to Model Lowest Cost Dietary Changes. Food Nutr Bull. 2016 Sep;37(3):247–60.
  26. 26. Pomeroy-Stevens A, Shrestha MB, Biradavolu M, Hachhethu K, Houston R, Sharma I, et al. Prioritizing and Funding Nepal’s Multisector Nutrition Plan. Food Nutr Bull. 2016 Dec; 37(4_suppl): S151–69. pmid:27909260
  27. 27. Manohar S, Klemm R, Rajbhandary R, Adhikari R, Gauchan D, Shrestha K, et al. PoSHAN Community Studies Baseline Report. Baltimore, MD: Nutrition Innovation Lab, Johns Hopkins University; 2014 Aug.
  28. 28. Dorsey JL, Manohar S, Neupane S, Shrestha B, Klemm RDW, West KP. Individual, household, and community level risk factors of stunting in children younger than 5 years: Findings from a national surveillance system in Nepal. Matern Child Nutr. 2017; e12434.
  29. 29. Krishnamurthy KP, Lewis K, Choularton RJ. Climate impacts on food security and nutrition. A review of existing knowledge. World Food Program; 2014.
  30. 30. Bennett L. Gender, caste and ethnic exclusion in Nepal: Following the policy process from analysis to action. In: New Frontiers of Social Policy, Arusha, Tanzania. 2005.
  31. 31. Bennett L, Dahal DR, Govindasamy P. Caste, ethnic and regional identity in Nepal: Further analysis of the 2006 Nepal Demographic and Health Survey. Calverton, Maryland, USA: Macro International Inc.; 2008.
  32. 32. Pandey JP, Dhakal MR, Karki S, Poudel P, Pradhan MS. Maternal and child health in Nepal: the effects of caste, ethnicity, and regional identity. Calverton, Maryland, USA: Nepal Ministry of Health and Population, New Era, and ICF International; 2013.
  33. 33. Lamichhane DK, Leem JH, Kim HC, Park MS, Lee JY, Moon SH, et al. Association of infant and young child feeding practices with under-nutrition: evidence from the Nepal Demographic and Health Survey. Paediatr Int Child Health. 2016 Oct; 36(4): 260–9. pmid:26863233
  34. 34. Martorell R, Leslie J, Moock P. Characteristics and determinants of child nutritional status in Nepal. Am J Clin Nutr. 1984; 39(1): 74–86. pmid:6691295
  35. 35. Block SA. Maternal nutrition knowledge versus schooling as determinants of child micronutrient status. Oxf Econ Pap. 2006 Nov 28; 59(2): 330–53.
  36. 36. Bhutta ZA, Das JK, Rizvi A, Gaffey MF, Walker N, Horton S, et al. Evidence-based interventions for improvement of maternal and child nutrition: what can be done and at what cost? The Lancet. 2013 Aug; 382(9890): 452–77.
  37. 37. Lartey A. Maternal and child nutrition in Sub-Saharan Africa: challenges and interventions. Proc Nutr Soc. 2008 Feb; 67(01): 105–8.
  38. 38. Rahman MM, Abe SK, Kanda M, Narita S, Rahman MS, Bilano V, et al. Maternal body mass index and risk of birth and maternal health outcomes in low- and middle-income countries: a systematic review and meta-analysis: Body mass index and pregnancy and health outcomes. Obes Rev. 2015 Sep; 16(9): 758–70. pmid:26094567
  39. 39. Simondon KB. Children’s height, health and appetite influence mothers’ weaning decisions in rural Senegal. Int J Epidemiol. 2001 Jun 1; 30(3): 476–81. pmid:11416068
  40. 40. Islam MM, Khatun M, Peerson JM, Ahmed T, Mollah MAH, Dewey KG, et al. Effects of energy density and feeding frequency of complementary foods on total daily energy intakes and consumption of breast milk by healthy breastfed Bangladeshi children. Am J Clin Nutr. 2008 Jul; 88(1): 84–94. pmid:18614728
  41. 41. Singh PK, Sachdev HP, Nagpal J, Bajaj M, Dubey AP. Short-term effect of complementary feeding frequency on total ad libitum consumption in 6-to 10-month-old breastfed Indian infants. J Pediatr Gastroenterol Nutr. 2005; 41(4): 422–425. pmid:16205509
  42. 42. Lindsay AC, Machado MT, Sussner KM, Hardwick CK, Peterson KE. Infant-Feeding Practices and Beliefs about Complementary Feeding among Low-Income Brazilian Mothers: A Qualitative Study. Food Nutr Bull. 2008 Mar; 29(1): 15–24. pmid:18510201
  43. 43. Rasheed S, Haider R, Hassan N, Pachón H, Islam S, Jalal CSB, et al. Why Does Nutrition Deteriorate Rapidly among Children under 2 Years of Age? Using Qualitative Methods to Understand Community Perspectives on Complementary Feeding Practices in Bangladesh. Food Nutr Bull. 2011 Sep; 32(3): 192–200. pmid:22073792
  44. 44. Nguyen PH, Avula R, Ruel MT, Saha KK, Ali D, Tran LM, et al. Maternal and Child Dietary Diversity Are Associated in Bangladesh, Vietnam, and Ethiopia. J Nutr. 2013 Jul 1; 143 (7): 1176–83. pmid:23658424
  45. 45. DFID, The World Bank. Unequal citizens: gender, caste and ethnic exclusion in Nepal. Kathmandu: World Bank; 2006.
  46. 46. Masters WA, Nene MD, Bell W. Nutrient composition of premixed and packaged complementary foods for sale in low-and middle-income countries: Lack of standards threatens infant growth. Maternal & child nutrition. 2017 Oct; 13(4): 13:e12421.
  47. 47. Mulmi P, Block SA, Shively GE, Masters WA. Climatic conditions and child height: Sex-specific vulnerability and the protective effects of sanitation and food markets in Nepal. Econ Hum Biol. 2016 Dec; 23: 63–75. pmid:27494247
  48. 48. Darrouzet-Nardi AF, Masters WA. Nutrition Smoothing: Can Proximity to Towns and Cities Protect Rural Children against Seasonal Variation in Agroclimatic Conditions at Birth? PLOS ONE. 2017 Jan 3; 12(1): e0168759. pmid:28045998