Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
1
Protocol: Do gender-based inequities in agriculture affect food security and
nutrition outcomes? A mixed-methods systematic review
London, 08/05/2018
Hayaan Nur*, Helen Harris-Fry*, Bhavani Shankar**, Suneetha Kadiyala*
* London School of Hygiene and Tropical Medicine
** School of Oriental and African Studies
Contents
Background ............................................................................................................................................. 2
How increased gender equity might improve nutrition ....................................................................... 2
Why it is important to conduct this review ......................................................................................... 3
Objectives ............................................................................................................................................... 4
Methodology ........................................................................................................................................... 5
Inclusion and exclusion criteria .......................................................................................................... 5
Search Strategy ................................................................................................................................... 6
Description of methods used in primary research ............................................................................... 8
Criteria for determination of independence of findings ...................................................................... 8
Details of study coding categories ...................................................................................................... 8
Synthesis procedures and conventions ................................................................................................ 8
References ............................................................................................................................................. 10
Appendix 1: Sample Search Terms ....................................................................................................... 12
Appendix 2: Coding Form .................................................................................................................... 15
Appendix 3: Preliminary Time Frame .................................................................................................. 22
Appendix 4: Preliminary List of Experts .............................................................................................. 24
2
Background Emerging evidence shows that agricultural interventions can improve both agricultural and nutritional
outcomes (Ruel, Quisumbing, & Balagamwala, 2017), indicating the potential for improvements in
agriculture to have multiple co-benefits for the health and wellbeing of rural populations. Realisation
of these benefits may be constrained by pervasive but highly heterogeneous gender-based inequities in
agriculture that limit women’s participation in and control over agricultural processes and outputs
(FAO, 2011).
Gender equity is defined as the ‘fairness of treatment for women and men, according to their respective
needs and interests. This may include equal treatment or treatment that is different but considered
equivalent in terms of rights, benefits, obligations and opportunities’ (ILO, 2007, p. 92). Gender
inequity therefore is gender-comparative, taking both women and men into account.
Women’s participation in agriculture is highly heterogeneous across contexts, but the Food and
Agriculture Organization estimate that women comprise an average of 43 percent of the agricultural
workforce in developing countries (FAO, 2011).Yet, women experience gendered inequities in many
agricultural domains, including limited control and access to land, livestock and inputs (seeds,
fertilisers, equipment and technology), work burdens (Johnston, Stevano, Malapit, Hull, & Kadiyala,
2018), and earnings (ILO, 2015). In addition to the potentially large societal costs to human wellbeing,
these gender-based inequities may be responsible for gender differentials in agricultural productivity
(Peterman, Behrman, & Quisumbing, 2014) and sub-optimal yields (Udry, Hoddinott, Alderman, &
Haddad, 1995). Constrained access to inputs and markets, and inefficiently gendering tasks and
responsibilities, might prevent women from participating in the most economically productive way. The
FAO (2011) estimates that, if women had the same access to productive resources as men, total
agricultural output in developing countries could rise by between 2.5 and 4 percent. Gender inequities
in agriculture may also leave women particularly vulnerable to climatic shocks, having limited access
to coping mechanisms, such as irrigation, improved seeds, credit and insurance, and alterative
livelihood strategies (Hagos & Holden, 2013).
How increased gender equity might improve nutrition Gender equity in agriculture may present a major opportunity to improve agriculture, food security,
nutrition, and health outcomes. It has been estimated that the increase in agricultural production due to
gender equity in agriculture could reduce the number of hungry people in the world by 12 to 17 percent
(FAO, 2011). There may be additional pathways through which gender equity in agriculture could
improve nutrition (see figure 1). For example, there is evidence that women tend to invest more income
on nutrition than men (Allendorf, 2007). Furthermore, gender equity might lead to an intra-household
allocation of foods that is more in line with the household members’ nutritional needs.
3
Figure 1: The causal link between gender (in)equity and nutrition. On the gender-inequity side, structural inequities are
marked in orange, intra-household inequities in grey. The orange arrow and box relate to structural inequalities.
Sahn & Younger (2009), in their analysis of seven developing countries, find that most nutritional
inequality takes place at the intra-household level. Therefore, it is important to consider intra-household
inequities carefully. Nevertheless, wider, structural inequities may affect nutrition through the intra- or
extra-household level. For instance, an improvement in women’s land inheritance rights might lead to
an increased share of household land being held by women, and also redistribute land more equitably
between households.
Why it is important to conduct this review We aim to assess the potential for gender equity in agriculture to improve nutritional outcomes, and
identify the relative importance of different domains of gender inequity in agriculture in different low
and middle-income contexts. This analysis will help policymakers to: (1) prioritise policy responses to
address the gender gap in agriculture; and (2) increase the effectiveness of existing programs to improve
nutrition by targeting the most important domains of agricultural inequity. This information may also
further motivate and strengthen multi-sectoral approaches to improving agricultural productivity and
nutritional status of populations.
Currently, evidence on the impact of gender inequities in agriculture on food security, diets, and
nutrition outcomes is segregated across multiple disciplines and is not well synthesised. It is, therefore,
of limited use to policymakers. We identified six reviews that relate to gender inequities in agriculture
or nutrition.
1. Doss (2013) reviews evidence on intra-household bargaining and resource allocation in
developing countries. She finds that there is sufficient evidence to conclude that empowering
women affects outcomes in general. She does not follow a systematic methodology and does
not analyse effects on nutrition or food security, or in rural settings specifically.
2. Harris-Fry, Shrestha, Costello, & Saville (2017) conduct a ‘conceptual review’ of the
determinants of intra-household food allocation to identify factors that have been proven or
hypothesised to affect intra-household food allocation. Confining their scope to South Asia,
they find that food allocation may be determined by relative differences in relative earned
income, bargaining power, food behaviours, social status, preferences and relationships within
4
households. They find no high-quality evidence relating gender inequities in agriculture to the
intra-household allocation of foods.
3. Johnston et al. (2018) examine literature on time use in agriculture as an explanation for
undernutrition. They find a complex relationship between time use and nutrition and no
agreement regarding impacts. They find no studies that connect inequities in time-use with
nutrition.
4. Berti, Krasevec, & FitzGerald (2004) review the effectiveness of agriculture interventions in
improving nutrition outcomes. They find that most interventions increase food production, but
do not necessarily improve nutrition. Their discussion of gender inequities is limited to a
comparison of interventions that consider gender issues in some form with ones that do not.
They find that interventions that consider gender have a greater likelihood of being effective.
5. Meinzen-Dick, Quisumbing, Doss, & Theis (2017) review the effect of women’s land rights on
poverty reduction. Based mostly on observational studies and on limited analysis of causal
pathways, they find some evidence of a positive relationship. However, they do not discuss
effects on nutrition or food security.
6. Ruel, Quisumbing, & Balagamwala (2017), extending on Ruel & Alderman (2013), review the
evidence on nutrition-sensitive agriculture interventions on nutrition. They build on a tradition
of reviews generally finding mixed results (Harvey et al., 2014; Leroy & Frongillo, 2007;
Masset, Haddad, Cornelius, & Isaza-Castro, 2012; Webb, 2013; Webb Girard, Self, McAuliffe,
& Olude, 2012). While Ruel et al. (2017) find that nutrition-sensitive agriculture programs
improve a variety of nutrition outcomes, they and other reviews do not discuss gender in depth.
In summary, many reviews to date have not employed fully systematic approaches, and they have either
focused on the effects of gender inequities in agriculture on broader poverty-related outcomes, or they
have focused on the overall effects of agriculture interventions on nutrition without a gender equity
focus. We did not identify any reviews that investigated the effects of gender-based inequities in
agriculture on nutrition outcomes.
Objectives This systematic review is motivated by the question ‘does gender equity in agriculture improve
household food security and women and children’s nutrition outcomes?’ It assesses the plausibility of
this effect with evidence from correlational studies, and identifies direct evidence in experimental and
quasi-experimental studies. We will explore the mechanism behind this effect by extracting
intermediary outcomes and synthesising qualitative literature.
Primary objective:
To critically review the evidence of the impact of different gender inequities in agriculture on
women and child nutrition outcomes and household food security in low and middle-income
countries
Secondary objectives:
1. Review the mechanisms by which gender inequities may affect the above outcomes
2. Describe geographical differences in the impact of gender inequities in agriculture on women
and child nutrition outcomes and household food security
3. Identify any differences in the effects of agricultural inequities on nutritional outcomes between
boys and girls
4. Review the evidence of environmental effects on:
a. gender inequities in agriculture, and
b. the relationships between gender inequities in agriculture and food security and
nutrition outcomes
5
Methodology
Inclusion and exclusion criteria Types of participants We will include all studies referring to rural contexts in low- and middle-income countries LMICs, as
defined by World Bank, (2018). We will use the year of the baseline data collection of studies to
determine whether a country was a low- or middle-income country at that point in time, as some
countries graduated from LMIC status. Studies that had their baseline data collected before 1987 will
be categories according to the 1987 categorisation (the first published by the World Bank).
To be considered in this review, there has to be an indication that at least one household member in the
majority of participating households is engaged in agriculture (ISIC divisions 01-03, United Nations,
2008). This includes household members engaging in agricultural labour (including food cropping and
cash cropping), sharecropping, land owning, subsistence farming, livestock, fisheries and forestry. The
study should include or refer to both men and women.
Exposures We will include studies that document exposures to variation in:
1. Gender inequity in land and livestock. At the structural level, these can be gender inequitable
inheritance and ownership rights and norms. At the household level, gender inequity refers to
inequities in the usufruct and statutory ownership, access, or use of land or livestock between
men and women within the household. In the qualitative literature, a wider framing of natural
resources and livestock ownership is allowed for, to include work on a higher conceptual level.
2. Gender inequity in labour markets. At the structural level, these are inequitable differences in
remuneration for labour, goods and services delivered by men and women, as well as inequities
in labour market opportunities. At the household level, these are inequities in earned income
and workforce participation.
3. Gender inequity in work burdens. At the structural level, these are gendered tasks and duties.
At the household level, this is the share of women’s work burden in the household relative to
men’s. In the quantitative literature, gender inequity in the work burdens is measured as the
inequity in total hours worked, time spent doing strenuous labour, or total energy expenditure.
Qualitative literature that conceptualises equity in time use or workloads more broadly is
allowed for. In this case, equity is distinct from equality. We do not consider an equal spread
of tasks ‘fair in the treatment of women and men’, if the total time burden and heavy labour are
distributed equally. Rather, the needs of women and men have to be met equally. For instance,
we consider it more equitable if pregnant women completely abstain from strenuous labour.
Comparison Participants with high exposures to inequity in agricultural assets, labour markets and work burdens
will be compared to participants with low exposures to the former. We exclude comparisons between
female-headed households and male-headed households. We exclude studies that do not report an effect
of exposures to inequity alone. We will exclude studies that only report an effect of reduced inequities
coinciding with other exposures, for instance. Thus, an education intervention, for example, might
reduce gender inequities in earned incomes. We usually would still exclude a study on it, because the
effect on nutrition through earned incomes cannot easily be disentangled from the effect through child
feeding practices, for instance.
Outcomes To be included, studies must document associations with at least one of the four following outcomes:
6
1. Undernutrition of women of reproductive age (aged 15-49 years). For inclusion, quantitative
literature has to measure undernutrition as the body-mass-index (BMI; kg/m2) and may use
mean BMI or cut-offs. Qualitative literature that characterises undernutrition more widely is
included.
2. Undernutrition of the children under 5 years of age. This is measured as stunting, wasting or
underweight (height-for-age z-scores, weight-for-height z-score or weight-for-age z-scores).
Qualitative work can characterise child undernutrition more widely.
3. Dietary quality of women of reproductive age. In the quantitative literature, this is typically
measured using the Minimum Dietary Diversity for Women (MDD-W; FAO and FHI 360,
2016) score or earlier versions of this score, but can be any standardisable measure of nutritional
adequacy calculated using individual dietary assessment that accounts for needs in some way
(e.g. using adequacy ratios, the cut-point method, or probability-based approaches).
‘Standardisable’ means that a z-score or a similar standardised measure can be calculated.
Qualitative studies which examine dietary quality in broader terms are included.
4. Dietary quality of the children under 5 years of age. This is typically measured using the
‘Minimum Dietary Diversity’ or ‘Minimum Adequate Diet’ indicators (WHO, 2010), but can
be any standardisable measure of nutritional adequacy. The same constraints to measurement
apply here, as for dietary quality of women.
5. Household food security. In the quantitative literature, food security is measured as (1)
household food expenditure or consumption in monetary terms (typically mean, per capita), (2)
the household share of food expenditures, i.e. proportion of total household expenditures spent
on food, (3) share of staple food expenditures, i.e. the proportion of total food expenditures
spent on staples; (4) household dietary diversity measured as any standardisable dietary
diversity score (typically HDDS;Swindale & Bilinsky, 2006), (4) the Household Food
Insecurity Access Scale (HFIAS; Coates, Swindale, & Bilinsky, 2007),. Qualitative literature
that measures food security in different terms is allowed for, as well.
Studies with at least one of these exposures and outcomes will be included. After these studies have
been selected, we will extract information on the following intermediate outcomes: agricultural
production, household income, women’s empowerment, household poverty, and economic inequity
between households. We will also extract information on climatic or environmental factors influencing
our exposures and the relationship between exposures and outcomes. We will also record other
intermediate outcomes on the impact pathways that emerge during the data extraction phase.
Types of studies Studies qualify for this review if they report original empirical evidence. This empirical evidence can
be qualitative or quantitative. The empirical evidence can document the association between exposures
and outcomes and causal impacts. We will exclude historical analyses. Studies written in English will
be included, and efforts will be made to include studies in other languages, based on the language
proficiencies of the investigators. There will be no exclusion based on publication date or the length of
follow-up. Published literature will be included, as well as unpublished studies that can be identified.
Search Strategy The search strategy will encompass a comprehensive literature and trial database search, the hand search
of selected journals, the backward tracking of citations and the consultation with experts.
The following databases and websites will be searched:
1. Bill and Melinda Gates Foundation Grants Database
2. BRIDGE
3. CAB Abstracts
4. Cochrane Central
7
5. DfID Research for Development Outputs
6. EBSCO Discovery Service (including EconLit, IDEAS Repec and the World Bank e-library)
7. Eldis
8. Evaluation at FAO Library
9. FAO AGRIS
10. Gender Studies Database (EBSCO)
11. OpenTrials (which includes information from ClinicalTrials.gov, EU CTR, HRA, WHO
ICTRP, and PubMed)
12. Popline
13. Medline (Ovid) P
14. Scopus
15. The 3ie impact evaluation database
16. The AEA RCT registry
17. The IFPRI library
18. The World Bank IEG evaluations
19. UNICEF Evaluation Database
20. USAID Development Experience Clearinghouse
21. Web of Science
The following journals will be hand-searched:
1. American Journal of Agricultural Economics
2. Food Policy
3. Journal of Development Economics
4. Maternal and Child Health Journal
5. Public Health Nutrition
6. The Journal of Development Studies
7. The Journal of Nutrition
8. The Lancet
9. The Lancet Global Health
10. The proceedings of the Argiculture, Nutrition and Health Academy conference
11. The proceedings of the CSAE Conference
12. The proceedings of the NEUDC Conference
13. The World Bank Economic Review
14. World Development
We will perform the hand-search for papers published after 1995, since a trial search did not find a
substantive number of papers that would be included before this date. In case more than 20% of the
articles identified through the database search is published before 1995, we will revise this cut-off. In
case we do not identify any articles that are generally relevant to this systematic review within the last
five years of publication, we will discontinue the hand-search of this source.
We will use the following publications for backward-citation-tracking:
1. All included articles
2. Berti, Krasevec, & FitzGerald, 2004
3. Doss, 2013
4. FAO, 2011
5. Girard et al., 2012
6. Johnston et al., 2018
7. Leroy & Frongillo, 2007
8. Masset et al., 2012
9. Meinzen-Dick, Quisumbing, Doss, & Theis, 2017
8
10. Ruel & Alderman, 2013
11. Ruel, Quisumbing, & Balagamwala, 2017
12. Webb, 2013
We will contact experts to identify any other relevant studies. A preliminary list of experts can be found
in Appendix 4 (Preliminary List of Experts).
We will search databases using the set of terms documented in Appendix 1 (Sample Search Terms).
Formatting will be adapted to the respective database.
Description of methods used in primary research We expect a mix of methods used in primary research. Since the field is multidisciplinary, we assume
that both qualitative and quantitative methods will be used. We anticipate that we will find both,
correlational, as well as causal evidence, and that most literature will not be experimental or quasi-
experimental studies.
Criteria for determination of independence of findings Multiple measures of the same exposure-outcome pair within one study or based on the same data will
not be regarded as independent. If we identify more than one publication or report on a single study, we
will choose the most prominent publication or report. This means, for instance, that we will prefer peer-
reviewed reports, edited and published reports, and reports with a higher citation count. If multiple
comparison groups are reported on, an equally weighted average will be created. If multiple estimates
of the same relationship are reported, we will choose the one preferred by the author of the paper. If
this is not clear, we will choose the one with the lowest risk of bias. If two papers report on two studies
of the same relationship based on the same data, we will choose the estimate with the lower risk of bias.
Details of study coding categories Details on the coding categories can be found in Appendix 2 (Coding Form).
Data Management EPPI reviewer will be used to manage references. Statistical analyses will be performed in Stata
(College Station, TX: StataCorp LP). Screening and risk of bias and study quality coding will be done
by Helen Harris-Fry and Hayaan Nur simultaneously (double screening and coding). Any conflicts in
their assessment will be resolved by the vote of Suneetha Kadiyala.
Missing data We will try to impute missing data. If data are missing, we will try to impute. In case the data cannot
be imputed, the authors will be contacted, with an answer period of at least one week.
Synthesis procedures and conventions Critical Appraisal We will report the results of the risk of bias assessment, both as an overall judgement per study, as well
as per bias domain or quality domain in an overview. The critical appraisal of quantitative studies
assesses risk of bias and employs a slightly modified version of the ROBINS I tool (Sterne et al., 2016).
The main changes we made are: the omission of the section ‘bias due to deviations from intended
interventions’, since this is not a review of interventions; and the addition of a section of bias in
instrumental variables study designs. The critical appraisal of qualitative studies assesses study quality
and is based on Lockwood, Munn, & Porritt (2015). We will exclude qualitative evidence that we deem
to be critically low quality. Further details can be found in Appendix 2 (Coding Form).
9
Narrative synthesis The synthesis will be structured by exposure. We will use the correlational quantitative literature to
make an assessment of the plausibility of a relationship between gender inequities and nutrition. We
will then investigate whether causal quantitative literature supports this relationship. In the next step,
we will use the qualitative literature to explain the mechanisms behind the relationship, and intermediate
outcomes reported in the causal quantitative literature to triangulate these findings. We will then
integrate the results from each part of the synthesis. In this step, we will also integrate findings about
the structural-level, societal inequities that might undergird intra-household gender inequities.
Meta-analysis If the number of standardisable quantitative studies allows (minimum of two), we will conduct a meta-
analysis. The meta-analysis will follow the outline of the narrative synthesis. We will perform the meta-
analysis using random effects. We will only convert effect sizes if necessary, meaning that there is no
valid comparison possible without standardisation. All analyses will account for the nature of the
distribution of the relevant outcome, and we will present results as appropriate effects sizes with a
measure of precision.
Analysis of heterogeneity If the data warrants meta-analysis, we will report tests for heterogeneity of effects across studies, using
the I2 statistic and τ2 for reporting the between studies variance component.
Analysis of publication bias We will give provide a qualitative assessment of the likelihood of publication bias, comparing the
methods and results in unpublished included studies with the ones in published ones. If feasible, we
will assess potential publication bias and small-study effects using funnel plots and Egger tests.
Subgroup analysis We will analyse the heterogeneity of findings across geographical regions (according to the World
Bank-defined regions). For child nutritional and dietary outcomes, we will test for differences in effects
between boys and girls. Furthermore, if the data allow, we will assess the heterogeneity of findings
across different climatic and environmental conditions in a narrative style. We will analyse the
sensitivity of findings to the risk of bias or study quality.
10
References Allendorf, K. (2007). Do Women’s Land Rights Promote Empowerment and Child Health in Nepal?
World Development, 35(11), 1975–1988. https://doi.org/10.1016/j.worlddev.2006.12.005
Berti, P. R., Krasevec, J., & FitzGerald, S. (2004). A review of the effectiveness of agriculture
interventions in improving nutrition outcomes. Public Health Nutrition, 7(5), 599–609.
https://doi.org/10.1079/PHN2003595
Coates, J., Swindale, A., & Bilinsky, P. (2007). Household Food Insecurity Access Scale (HFIAS) for
measurement of food access: indicator guide. Washington DC: FHI 360/FANTA.
https://doi.org/10.1007/s13398-014-0173-7.2
Doss, C. (2013). Intrahousehold bargaining and resource allocation in developing countries. World
Bank Research Observer, 28(1), 52–78. https://doi.org/10.1093/wbro/lkt001
FAO. (2011). Women in Agriculture. Closing the Gender Gap for Development. The State of Food and
Agriculture. Rome.
FAO and FHI 360. (2016). Minimum Dietary Diversity for Women. A Guide to Measurement. Food
And Nutrition Technical Assistance Ⅲ. Rome: Food and Agriculture Organization of the United
Nations.
Hagos, H. G., & Holden, S. (2013). Links Between Tenure Security and Food Security : Evidence from
Ethiopia (ESSP Working Paper No. 59). ESSP Working Paper 59. Addis Ababa.
https://doi.org/10.2139/ssrn.2343158
Harris-Fry, H., Shrestha, N., Costello, A., & Saville, N. M. (2017). Determinants of intra-household
food allocation between adults in South Asia - A systematic review. International Journal for
Equity in Health, 16(1), 1–21. https://doi.org/10.1186/s12939-017-0603-1
Harvey, M., Dangour, A., Lambert, R., Alemu, M., Ashton-Griffiths, E., Green, T., … Sutherland, J.
(2014). Can agriculture interventions promote nutrition? (Agriculture and nutrition evidence
paper). London.
ILO. (2007). ABC of women workers’ rights and gender equality (2nd ed.). Geneva: International
Labour Office. Retrieved from http://www.ilo.org/wcmsp5/groups/public/---dgreports/---
gender/documents/publication/wcms_087314.pdf
ILO. (2015). Global Wage Report 2014/15. Wages and Income Inequality. Geneva: International
Labour Office.
Johnston, D., Stevano, S., Malapit, H. J., Hull, E., & Kadiyala, S. (2018). Review: Time Use as an
Explanation for the Agri-Nutrition Disconnect: Evidence from Rural Areas in Low and Middle-
Income Countries. Food Policy, 76(January 2016), 8–18.
https://doi.org/10.1016/j.foodpol.2017.12.011
Leroy, J. L., & Frongillo, E. a. (2007). Can interventions to promote animal production ameliorate
undernutrition? The Journal of Nutrition, 137(10), 2311–2316. Retrieved from
jn.nutrition.org/content/137/10/2311.
Lockwood, C., Munn, Z., & Porritt, K. (2015). Qualitative research synthesis: Methodological guidance
for systematic reviewers utilizing meta-aggregation. International Journal of Evidence-Based
Healthcare, 13(3), 179–187. https://doi.org/10.1097/XEB.0000000000000062
Masset, E., Haddad, L., Cornelius, A., & Isaza-Castro, J. (2012). Effectiveness of agricultural
interventions that aim to improve nutritional status of children: Systematic review. BMJ (Online),
344(7843), 1–7. https://doi.org/10.1136/bmj.d8222
Meinzen-Dick, R., Quisumbing, A., Doss, C., & Theis, S. (2017). Women’s Land Rights as a Pathway
to Poverty Reduction. A Framework and Review of Available Evidence (IFPRI Discussion Paper
11
No. 1663). Washington DC.
Peterman, A., Behrman, J., & Quisumbing, A. R. (2014). A review of empirical evidence on gender
differences in nonland agricultural inputs, technology, and services in developing countries. In A.
R. Quisumbing, R. Meinzen-Dick, T. L. Raney, A. Croppenstedt, J. A. Behrman, & A. Peterman
(Eds.), Gender in Agriculture. Closing the Knowledge Gap (pp. 145–186). Dordrecht: Springer
Netherlands.
Ruel, M. T., & Alderman, H. (2013). Nutrition-sensitive interventions and programmes: How can they
help to accelerate progress in improving maternal and child nutrition? The Lancet, 382(9891),
536–551. https://doi.org/10.1016/S0140-6736(13)60843-0
Ruel, M. T., Quisumbing, A. R., & Balagamwala, M. (2017). Nutrition-Sensitive Agriculture. What
have we learned and where do we go from here ? (IFPRI Discussion Papers No. 1681).
https://doi.org/10.1056/NEJMoa1511939.2.
Sahn, D. E., & Younger, S. D. (2009). Measuring intra-household health inequality. Explorations using
the body mass index. Health Economics, 18, S13–S36. https://doi.org/10.1002/hec.1459
Sterne, J. A., Hernán, M. A., Reeves, B. C., Savović, J., Berkman, N. D., Viswanathan, M., …
Higgins, J. P. (2016). ROBINS-I: A tool for assessing risk of bias in non-randomised studies of
interventions. BMJ (Online), 355, 4–10. https://doi.org/10.1136/bmj.i4919
Swindale, A., & Bilinsky, P. (2006). Household Dietary Diversity Score (HDDS) for measurement of
household food access: Indicator guide (v.2). Washington DC: FHI 360/FANTA.
https://doi.org/10.1017/CBO9781107415324.004
Udry, C., Hoddinott, J., Alderman, H., & Haddad, L. (1995). Gender differentials in farm productivity:
implications for household efficiency and agricultural policy. Food Policy, 20(5), 407–423.
https://doi.org/10.1016/0306-9192(95)00035-D
United Nations. (2008). International standard Industrial Classification of All Economic Activities
(ISIC), Rev. 4. (Department of Economic and Social Affairs & Statisics Division, Eds.), Statistical
Papers (Series M N). New York: United Nations Publication.
Webb, P. (2013). Impact Pathways from Agricultural Research to Improved Nutrition and Health:
Literature Analysis and Research Priorities. Geneva.
Webb Girard, A., Self, J. L., McAuliffe, C., & Olude, O. (2012). The effects of household food
production strategies on the health and nutrition outcomes of women and young children: A
systematic review. Paediatric and Perinatal Epidemiology, 26(SUPPL. 1), 205–222.
https://doi.org/10.1111/j.1365-3016.2012.01282.x
WHO. (2010). Indicators for assessing infant and young child feeding practices. Part 2, Measurement.
Geneva: World Health Organization. https://doi.org/10.3945/ajcn.111.020099
World Bank. (2018). World Bank Country and Lending Groups. Retrieved February 27, 2018, from
https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-and-
lending-groups
12
Appendix 1: Sample Search Terms Formatted for MEDLINE (Ovid)
1
( ( land adj5 right* ) OR landownership OR ( land adj5 own* ) OR ( land adj5 tenure ) OR
( property adj5 right* ) OR ( land adj5 tit* ) OR (land adj3 use*) OR ( asset* adj5 own* ) OR
( asset* adj5 right* ) OR ( livestock adj5 own* ) OR ( livestock adj5 right* ) OR ( land adj5
access ) OR ( property adj5 access ) OR ( poultry adj5 own* ) OR ( control* adj5 land ) OR
( control* adj5 asset* ) OR ( control* adj5 livestock ) OR ( inherit* adj5 right* ) OR ( inherit*
adj5 practic* ) OR ( farmland adj5 right* ) OR ( farmland adj5 own* ) OR ( farmland adj5
tenure ) OR ( farmland adj5 tit* ) OR ( farmland adj5 access ) OR ( farmland adj5 control )
OR (inherit* adj3 (asset* OR land OR farmland)) OR ((decision* OR power) adj3 ( land OR
livestock OR farmland OR asset* OR poultry ))).ti,ab,kw,sh.
2
( wage* OR remunerat* OR pay* OR paid OR salar* OR income* OR labor* OR labour*
OR employ* OR workforce OR ( ( vocation* OR job* OR occupation* OR profession* ) adj3
( opportunit* OR participa* ) ) OR (livelihood* adj3 opportunit*) OR ( ( intrahousehold OR "intra
household" ) adj5 bargain* ) OR ( ( intrahousehold OR "intra household" ) adj5 allocat* ) OR
( ( intrahousehold OR "intra household" ) adj5 ( decision* adj3 ( make* OR making OR made
) ) ) OR ( within adj3 ( household adj5 ( allocation OR bargain* ) ) ) ) .ti,ab,kw,sh.
3
( task* OR ( time adj3 use* ) OR activit* OR ( time adj3 allocat* ) OR workload* OR (
energ* adj3 expend* ) ).ti,ab,kw,sh.
4
(nutr* OR nourish* OR undernutrition OR undernourish* OR underweight OR bmi OR body
mass index OR anthropom* OR stunt* OR haz OR height-for-age OR length-for-age OR
wasting OR wasted OR WHZ OR WLZ OR WAZ diet* OR mdds OR mdd-w OR mddw OR dds
OR diversity score OR malnutrition OR caloric deficienc* OR malnourish* OR food security
OR food insecurity OR ( food adj5 expenditure* ) OR ( food adj5 consumption ) OR ( food
adj3 share* ) OR ( staple* adj5 share) OR hfias OR ( ( food OR calorie ) adj3 intake ) OR (
( intrahousehold OR intra household ) adj3 welfare ) ).ti,ab,kw,sh.
5
(( ( sex OR gender* OR female* OR woman OR women ) AND ( equit* OR equalit* OR
inequit* OR inequal* OR unequal* OR discriminat* OR power OR bargain* OR empower*
OR disempower*) ) OR (( sex OR gender* OR female* OR woman OR women ) adj5
power)).ti,ab,kw,sh.
6
(( ( sex OR gender* OR female* OR woman OR women ) adj3 (bias* OR parity) ) OR (effects
adj3 gender)).ti,ab,kw,sh.
7
( ( sex OR gender* OR female* OR woman OR women ) AND ( ( share OR gap OR distrib*)
adj3 ( land OR landownership OR property OR asset* OR livestock OR poultry OR inherit*
OR farmland OR wage* OR remunerat* OR pay* OR paid OR salar* OR income* OR
labor* OR labour* OR employ* OR workforce OR ( ( vocation* OR job* OR occupation*
OR profession* ) adj3 ( opportunit* OR participa* ) ) OR task* OR activit* OR workload* ) )
).ti,ab,kw,sh.
8
13
( ( ( intrahousehold OR intra household ) adj5 bargain* ) OR ( ( intrahousehold OR intra
household ) adj5 allocat* ) OR ( ( intrahousehold OR intra household ) adj5 ( decision* adj3 (
make* OR making OR made ) ) ) OR ( within adj3 ( household adj5 ( allocation OR bargain*
) ) ) AND ( sex OR gender* OR female* OR woman OR women )).ti,ab,kw,sh.
9
(1 or 2 or 3) and 4 and (5 or 6 or 7 or 8)
10
( afghanistan OR albania OR algeria OR angola OR argentina OR armenia OR armenian OR
azerbaijan OR bangladesh OR benin OR byelarus OR byelorussian OR belarus OR belorussian
OR belorussia OR belize OR bhutan OR bolivia OR bosnia OR herzegovina OR hercegovina
OR botswana OR brazil OR bulgaria OR Burkina Faso OR Burkina Fasso OR Upper Volta
OR burundi OR urundi OR cambodia OR Khmer Republic OR kampuchea OR cameroon OR
cameroons OR cameron OR camerons OR Cape Verde OR Central African Republic OR chad
OR china OR colombia OR comoros OR comoro islands OR comores OR mayotte OR congo
OR zaire OR Costa Rica OR Cote d'Ivoire OR Ivory Coast OR cuba OR djibouti OR
somaliland OR dominica OR Dominican Republic OR East Timor OR East Timur OR Timor
Leste OR ecuador OR egypt OR United Arab Republic OR El Salvador OR eritrea OR ethiopia
OR fiji OR gabon OR Gabonese Republic OR gambia OR gaza OR Georgia Republic OR
Georgian Republic OR ghana OR grenada OR guatemala OR guinea OR guiana OR guyana
OR haiti OR honduras OR india OR maldives OR indonesia OR iran OR iraq OR jamaica
OR jordan OR kazakhstan OR kazakh OR kenya OR kiribati OR korea OR kosovo OR
kyrgyzstan OR kirghizia OR Kyrgyz Republic OR kirghiz OR kirgizstan OR Lao PDR OR
laos OR lebanon OR lesotho OR basutoland OR liberia OR libya OR macedonia OR
madagascar OR Malagasy Republic OR malaysia OR malaya OR malay OR sabah OR sarawak
OR malawi OR mali OR Marshall Islands OR mauritania OR mauritius OR Agalega Islands
OR mexico OR micronesia OR Middle East OR moldova OR moldovia OR moldovian OR
mongolia OR montenegro OR morocco OR ifni OR mozambique OR myanmar OR myanma
OR burma OR namibia OR nepal OR Netherlands Antilles OR nicaragua OR niger OR nigeria
OR muscat OR pakistan OR palau OR palestine OR panama OR paraguay OR peru OR
philippines OR philipines OR phillipines OR phillippines OR Papua New Guinea OR romania
OR rumania OR roumania OR rwanda OR ruanda OR Saint Lucia OR St Lucia OR Saint
Vincent OR St Vincent OR grenadines OR samoa OR Samoan Islands OR Navigator Island
OR Navigator Islands OR Sao Tome OR senegal OR serbia OR montenegro OR seychelles
OR Sierra Leone OR Sri Lanka OR Solomon Islands OR somalia OR sudan OR suriname OR
surinam OR swaziland OR South Africa OR syria OR tajikistan OR tadzhikistan OR tadjikistan
OR tadzhik OR tanzania OR thailand OR togo OR Togolese Republic OR tonga OR tunisia
OR turkey OR turkmenistan OR turkmen OR uganda OR ukraine OR uzbekistan OR uzbek
OR vanuatu OR New Hebrides OR venezuela OR vietnam OR Viet Nam OR West Bank OR
yemen OR zambia OR zimbabwe ).ti,ab,kw,sh.
11
( ( ( developing OR less* developed OR under developed OR underdeveloped OR middle
income OR low* income OR underserved OR under served OR deprived OR poor* ) adj2 (
countr* OR nation* OR population* OR world OR state* ) ) OR ( ( developing OR less*
developed OR under developed OR underdeveloped OR middle income OR low* income ) adj2
( economy OR economies ) ) OR ( low* adj2 ( gdp OR gnp OR gross domestic OR gross
national ) ) OR ( low adj3 middle adj3 countr* ) OR ( lmic OR lmics OR ( third adj2 world
) OR lami countr* OR transitional countr* ) OR Africa OR (Africa* NOT African American)
OR asia* OR Latin America* OR South America* OR Central America* OR caribbean OR
oceania* OR Middle East* OR mena ).ti,ab,kw,sh.
12
14
10 OR 11
13
9 AND 12
15
Appendix 2: Coding Form
Bibliographic information
Author
Title
Year
Publication type Journal Article/Other peer-reviewed
material/Other (specify type and
published/unpublished)
Series Name
Language
General Information
Country of study
Funding source of study
Potential conflict of interest from funding?
(Based on own judgement if no declaration is
provided)
Yes/No/Unclear
Study Eligibility
Types of participants
Low or middle income country? Yes/No (Exclude)/Unclear
Geographic region East Asia and Pacific/Europe and Central Asia/
Latin America & the Caribbean/MENA/South
Asia/Sub-Saharan Africa
One household member in primary sector in
majority of cases? (If assumptions have to be
made, describe them)
Yes/No (Exclude)/Unclear
Exposures
Gender inequity in land or livestock Yes/No/Unclear
Gender inequity in work burden Yes/No/Unclear
Gender inequity in labour markets (wage gap,
opportunities, earned income or participation)
Yes/No/Unclear
Exclude if all three “No”
Is the study an interventionist study? Yes/No/Unclear
Outcomes
What are the outcomes documented in the study
(including those that are reported but not
included in our study)?
Is at least one outcome: nutrition or dietary
quality of women, nutrition or dietary quality of
children, or household food security?
Yes/No(Exclude)/Unclear
If the study is quantitative: Is an eligible metric
used to assess the outcome?
(women: BMI, dietary diversity, dietary
adequacy;
children: HAZ, WHZ. WAZ, dietary diversity,
dietary adequacy;
Yes/No(Exclude)/Unclear
16
HH: food expenditure, food share, staples share,
HDDS, HFIAS)
Type of Study
Is the study empirical? (reports original data, i.e.
qualitative or quantitative data directly on
women/ children/ HHs).
No literature research/reviews
Yes/No(Exclude)/Unclear
Is the study quantitative? Yes/No/Unclear
Is the study qualitative? Yes/No/Unclear
Does the study try to establish a causal or
correlational link between exposures and
outcomes?
Causal/Correlational/Both/Unclear/Other
(describe)
If the study is excluded: Should references be
tracked?
Yes/No/Unclear
Study details
Description of the exposure measure
Description of any evidence presented that the
exposure is (driven by) a structural phenomenon
Description of the outcome measure Maternal BMI, maternal dietary
diversity, maternal dietary adequacy;
HAZ, WHZ, WAZ, child dietary
diversity, child dietary adequacy;
HH food expenditure, HH food share,
HH staples share, HDDS, HFIAS)
Measure of effect Treatment effect: ITT/ATT/ATE/Other(specify)
Correlation: Conditional/Unconditional (specify)
Description of the measure of effect
Start date of the study (baseline)
If intervention-based: Description of the
intervention
Start date of the intervention/ change in exposure
Description of the method of analysis
Sampling methodology
Sample size
If a representative sample: Of which population
was the sample representative?
If non-representative sample: Describe the
participants
% missing values/ attrition
Results
(Statistics will be extracted with a separate table)
Describe the results for the association between
the exposure and outcome. In case there is more
than one pair, report separately.
Is there any evidence that this relationship is
influenced by climatic or environmental change?
In case there is more than one pair, report
separately.
17
Describe the results for the association of the
exposure on any intermediary outcome
(agricultural production, household income,
women’s empowerment, household poverty,
economic inequity between households, or
other).
In case is more than one pair, report separately.
Are there any adverse effects of the intervention/
change in exposure reported in the study?
Describe.
Risk of bias
For quantitative evidence:
The following confounders should be addressed in most studies
Household income, poverty or other appropriate measure of economic status,
Land ownership
Household size / composition,
Caste/ ethnicity/ religion/ gender attitudes
Specify a target randomised trial
Unit level of randomisation Individual/community/country/other(describe)
Participants
Experimental intervention
Comparator
Specify the outcome which is being
assessed for risk of bias
In case of multiple analyses being
presented, specify the numeric result
and/or reference (e.g. to a table, figure
or paragraph) that uniquely defines the
result being assessed
Complete the following item for each confounding domain (i) listed above, or (ii) relevant to
the setting or this particular study, or (iii) which the study authors identified as potentially
important
Confounding domain
Measure of confounder
Is there evidence that controlling for this
variable was unnecessary?
Is the confounding domain measured
validly and reliably by this variable?
Bias due to confounding Description Response options
Y = Yes
PY = Probably Yes
PN = Probably No
N = No
NI = No Information
NA = Not applicable
1.1 Is there potential for confounding in
this study?
Y/PY/PN/N
If Y/PY: determine whether there is a need to assess time-varying confounding
18
Bias due to confounding Description Response options
Y = Yes
PY = Probably Yes
PN = Probably No
N = No
NI = No Information
NA = Not applicable
1.2 Was the analysis based on
panel/time-series data?
(If N/PN, answer question 1.4 to 1.6.
If Y/PY, go to question 1.3)
NA/Y/PY/PN/N/NI
1.3 Were changes in the exposure likely
to be related to factors that are
prognostic for the outcome?
(If N/PN, answer questions 1.4 to 1.6.
If Y/PY, answer questions 1.4 to 1.8)
NA/Y/PY/PN/N/NI
Relating to baseline confounding
1.4 Did the authors use an appropriate
analysis method that controlled for all
important confounding domains?
NA/Y/PY/PN/N/NI
1.5 If Y/PY to 1.4:
Were confounding domains that were
controlled for measured validly and
reliably by the variables available in this
study?
NA/Y/PY/PN/N/NI
1.6 Did the authors control for any post-
exposure variables that could have been
affected by the intervention/exposure?
NA/Y/PY/PN/N/NI
Relating to baseline and time-varying confounding
1.7 Did the authors use an appropriate
analysis method that controlled for all
important confounding domains and for
time-varying confounding?
NA/Y/PY/PN/N/NI
1.8 If Y/PY to 1.7:
Were confounding domains that were
controlled for measured validly and
reliably by the variables available in this
study?
NA/Y/PY/PN/N/NI
Risk of bias due to confounding
judgement
Low/Moderate/Serious/Critical/NI
Selection Bias
2.1 Was the inclusion of participants
into the study based on participants’
characteristics?
Y/PY/PN/N/NI
2.2 If Y/PY to 2.1:
Were these characteristics likely to be
associated with the exposure?
NA/Y/PY/PN/N/NI
2.3 If Y/PY to 2.1 and 2.2:
Were the post-exposure variables that
influenced selection likely to be
influenced by nutrition/ food security or
a cause of nutrition/ food security?
NA/Y/PY/PN/N/NI
Risk of selection bias judgement Low/Moderate/Serious/Critical/NI
Bias in the classification of exposures
19
Bias due to confounding Description Response options
Y = Yes
PY = Probably Yes
PN = Probably No
N = No
NI = No Information
NA = Not applicable
3.1 Were the groups exposed to gender
inequity clearly defined?
Y/PY/PN/N/NI
3.2 Could the classification of gender
inequity have been affected by
knowledge of the outcome or risk of the
outcome?
Y/PY/PN/N/NI
Risk of bias due to the classification of
exposures judgement
Low/Moderate/Serious/Critical/NI
Bias due to missing data
4.1 Were outcome data available for all,
or nearly all, participants?
Y/PY/PN/N/NI
4.2 Were participants excluded due to
missing data on the exposure?
Y/PY/PN/N/NI
4.3 Were participants excluded due to
missing data on other variables needed
for the analysis?
Y/PY/PN/N/NI
4.4 If PN/N to 4.1 or Y/PY to 4.2 or 4.3:
Are the proportion of participants and
reasons for missing data similar across
all degrees of exposure?
NA/Y/PY/PN/N/NI
4.5 If PN/N to 4.1 or Y/PY to 4.2 or 4.3:
Is there evidence that results were robust
to the presence of missing data?
NA/Y/PY/PN/N/NI
Risk of bias due to missing data
judgement
Low/Moderate/Serious/Critical/NI
Bias in measurement of outcomes
5.1 Could the outcome measure have
been influenced by the knowledge of the
exposure to gender inequity?
Y/PY/PN/N/NI
5.2 Were the outcome assessors aware
of the study goal?
Y/PY/PN/N/NI
5.3 Were the methods of outcome
assessment comparable across different
exposures to gender inequity?
Y/PY/PN/N/NI
5.4 Were any systematic errors in
measurement of the outcome related to
exposures to gender inequity?
Y/PY/PN/N/NI
Risk of bias in measurement of
outcomes judgement
Low/Moderate/Serious/Critical/NI
Bias in selection of the reported results
Is the reported effect estimate likely to be selected on the basis of the results, from…
7.1 … multiple outcome measurements
within the outcome domain?
Y/PY/PN/N/NI
7.1 … multiple analyses of the
exposure-outcome relationship?
Y/PY/PN/N/NI
7.3 … different subgroups? Y/PY/PN/N/NI
20
Bias due to confounding Description Response options
Y = Yes
PY = Probably Yes
PN = Probably No
N = No
NI = No Information
NA = Not applicable
Risk of bias due to the selection of
reported results judgement
Low/Moderate/Serious/Critical/NI
IV Methods Bias
7.1 If an IV estimate was used
(including fuzzy RDDs), is there
evidence that the exclusion restriction
did not hold?
NA/Y/PY/PN/N/NI
7.2 If an IV estimate was used
(including fuzzy RDDs), is there
evidence that the instrument was
relevant (F-statistic >10)?
NA/Y/PY/PN/N/NI
Risk of IV-bias judgement Low/Moderate/Serious/Critical/NI
Overall bias
Risk of bias judgement Low/Moderate/Serious/Critical/NI
For qualitative evidence:
Description Response
Y = Yes
PY = Probably Yes
PN = Probably No
N = No
NI = No Information
NA = Not applicable
1. Is there congruity between the stated
philosophical perspective and the
research methodology? Consider if the
researcher has justified the study
design.
NA/Y/PY/PN/N/NI
2. Is there congruity between the research
methodology and the research question
or objectives?
NA/Y/PY/PN/N/NI
3. Is there congruity between the research
methodology and the sampling
methods? Can tell if we know who was
selected, why, and if any people
refused to participate
NA/Y/PY/PN/N/NI
4. Is there congruity between the research
methodology and the methods used to
collect data? Can tell if methods are
described, justified, modifications
made clear, form of data made clear
(tape recordings or notes?), and data
saturation mentioned.
NA/Y/PY/PN/N/NI
5. Is there congruity between the research
methodology and the representation
and analysis of data? Can tell if there
NA/Y/PY/PN/N/NI
21
is an in-depth description of analysis
methods, including how themes were
derived from the data if thematic
analysis, whether the researcher
explains how data presented were
selected from the original sample, if
data are sufficient
6. Is there congruity between the research
methodology and the interpretation of
results? Can tell if findings are
explicit, there is adequate discussion of
the evidence for and against
researcher’s arguments, if the
researcher has discussed the credibility
of their findings (e.g.
triangulation/multiple analyses) and if
findings are discussed in relation to
original research question
NA/Y/PY/PN/N/NI
7. Is there a statement locating the
researcher culturally or theoretically?
NA/Y/PY/PN/N/NI
8. Is the influence of the researcher on the
research, and vice- versa, addressed?
Can tell if researcher has examined
their own role and potential bias in
formulation of questions, data
collection, analysis, and response to
events during the study
NA/Y/PY/PN/N/NI
9. Are participants, and their voices,
adequately represented?
NA/Y/PY/PN/N/NI
10. Is the research ethical according to
current criteria or, for recent studies,
and is there evidence of ethical
approval by an appropriate body?
NA/Y/PY/PN/N/NI
11. Do the conclusions follow from the
analysis, or interpretation, of the data?
NA/Y/PY/PN/N/NI
Overall quality appraisal High quality/Medium quality/Low
quality/Critical/NI
Extractor information
Name
Date
Place
Extracted references for screening
22
Appendix 3: Preliminary Time Frame
12-Feb 19-Feb 26-Feb 05-Mar 12-Mar 19-Mar 26-Mar 02-Apr 09-Apr 16-Apr 23-Apr 30-Apr 07-May 14-May
Review activities
Develop protocol Feedback and
finalisation of protocol Run preliminary
searches, refine as
needed, and run the
search
Filter results
Reconcile discrepancies
Data extraction
Quality review Write summary results
for Lancet series Incorporate feedback
from inequities group Incorporate feedback
from wider Lancet
group Full paper - including
meta-analysis if
appropriate
(Continued on next page)
23
21-May 28-May 04-Jun 11-Jun 18-Jun 25-Jun 02-Jul 09-Jul 16-Jul 23-Jul 30-Jul 06-Aug
Review activities
Develop protocol Feedback and finalisation of
protocol Run preliminary searches,
refine as needed, and run the
search
Filter results
Reconcile discrepancies
Data extraction
Quality review Write summary results for
Lancet series Incorporate feedback from
inequities group Incorporate feedback from
wider Lancet group Full paper - including meta-
analysis if appropriate
24
Appendix 4: Preliminary List of Experts
1. Agnes Quisumbing
2. Amy Webb-Girard
3. Cheryl Doss
4. Chrisopher Udry
5. Deborah Johnston
6. Eileen Kennedy
7. Esther Duflo
8. Keera Allendorf
9. Lawrence Haddad
10. Marie Ruel
11. Naila Kabeer
12. Patrick Webb
13. Paulina Rossi
14. Peter Berti
15. Ruth Meinzen-Dick
16. Somebody at DFID
17. Somebody at the FAO
18. Somebody at the World Bank
19. Somebody at WHO
20. Virginie Le Masson