Table 2 Qualitative studies: Geographical and methodological characteristics of selected articles (n=15)

 Abdullah [36]

1983

Unstructured interviews

40 HH

One rural Muslim village in central-west Bangladesh. Mostly male respondents. Particularly in poor households, women also participated.

Notes recorded on paper, and results analysed by wealth group.

• Economic contributions

• Norms relating to receiving a ‘fair share

• Food security and scarcity

• Household structure (allocation to women in parental vs marital homes)

Food allocation

 Mukherjee [75]

2002

Seasonal calendar

1 group

Not reported

Method of quantifying discrimination not specified

• Season

Discrimination in consumption of food items and types

 Naved [35]

2000

Focus group discussions, case studies, and other methods.

Case studies of 22 women; 19 men

Three villages participating in an agricultural program.

Male and female beneficiaries of the program.

Triangulation of multiple qualitative techniques

• Physically strenuous labour contributions

• Bargaining power

• Individuals tastes and preferences (women eating more less-preferred foods than men)

• Food availability (Seemingly contradictory anecdotes that increased food availability did not change food allocation patterns much, but food scarcity led to men being less likely to have sufficient food than women.)

Allocation of specific food items

 Rohner and Chaki-Sircar [71]

1988

Observation? (limited detail)

1 village

Not reported

Not reported

• Caste - High caste men and boys had the best quality food, especially eggs, milk and fish. Implied that this is less the case with lower caste households.

Food quality

 Caldwell, Reddy and Caldwell [51]

1983

In-depth questions and case studies

50% of 4773 population (n = 2387)

One large village and eight smaller villages in rural area of southern Karnataka.

Individual respondent characteristics not reported.

Daily scrutiny of findings and on-going modification of questions to identify behavioural patterns

• Beliefs about equity - Respondents were reluctant to talk about food allocation. This “demonstrates the existence of some belief in equitable distribution”. Inequity was “as much a matter of poor communication as of deliberate intent”.

• Interpersonal relationships – Differential food allocation was in decline due to the strengthening bond between husband and wife.

Food allocation

 Daivadanam et al. [59]

2014

Interviews and focus group discussions

17 individuals;

3 groups

Rural areas (one coastal and one non-coastal)

Men and women aged between 23 and 75 years, of different religions and socio-economic status. Mostly female heads of household and others involved in dietary decision-making. One group mostly comprised of men.

Modified framework analysis using inductive and deductive reasoning – did not try to fit the data into pre-identified themes.

• Tastes and preferences - women prioritised their own food preferences the least

Allocation of preferred foods

 Katona-Apte [39]

1977

In-depth interviews

62 pregnant women or mothers

Two districts from Tamil Nadu. All households had a total income of <200 Indian Rupees per month.

All female respondents, and most were pregnant, lactating, or a mother of child under two years.

Analysis method not reported.

• Cultural beliefs about foods – pregnant and lactating women avoided certain foods, and this caused them to have less adequate diets, particularly if there was lack of variety or budget to replace avoided foods with nutritious alternatives

Allocation of specific food types that have different properties

 Khan et al. [33]

1987

In-depth interviews and participant observations

20 individuals

One study village from western Uttar Pradesh

20 main female informants (age not reported) from different caste and class groups, and extensive discussions with other villagers, including men.

Analysis method not reported.

• Economic contributions - Respondents said that men should eat more because they earn and provide for the family. The belief that men should be given more food was rarer (3 / 6 respondents) when women earned an income. Some women ate less because they did not have time to eat. Women had less appetite due to fatigue after cooking and serving her family members.

• Religious and cultural beliefs - Women “enjoy this spirit of sacrifice for the family”. There was also a belief that pregnant women should eat ghee (clarified butter) to give lubrication during birth. The cultural norm of the female cook eating last meant that women eat less.

• Status - Women had a religious obligation to fast for the family and for men to have superior status and allocation of food.

• Household income - In poor households, the eating order negatively affected women; in wealthy landowning families it did not.

Allocation of food generally, and also of specific food items

 Miller [58]

1981

Review of ethnographies

58 studies

Review of many studies from across India.

Meta-analysis

• Interpersonal relationships - Serving food was a way that women show love and affection to their men. Similarly, refusing to eat food was a common method for a man to punish his wife or mother.

Food allocation

 Nichols [74]

2016

Semi-structured interviews, and informal conversations and participant observation

81 individuals

Four villages in sub-Himalayan district.

Respondents: Government workers, NGO employees, village men, women, and couples. Convenience sampling to include respondents from different class, caste, age, gender and household composition. Plus, national-level NGO representatives from Delhi.

Thematic analysis, by coding themes and intersections between themes

• Labour / physically strenuous economic contributions - women ate the least during planting and harvest seasons when they were working the hardest (and working harder than men) due to a lack of appetite from the exhaustion of the labour

Food allocation

 Palriwala [34]

1993

Participant observation

1 village

Sikar district, rural agricultural village with Hindu (85%) and Muslim (15%) castes.

Individual participant characteristics not reported.

Not reported

• Cultural beliefs / eating order - youngest daughter in law usually cooks and eats last, leading to less diverse diet as there may be no lentils or vegetables left.

• Food scarcity – eating order particularly affected the daughters-in-law during food scarcity.

• Economic contributions affect food allocation – income earners are given priority of delicacies and nutrient-rich items like ghee,

• Interpersonal relationships – food allocation affected by kinship status, particularly agnation.

Allocation of specific food items

 Gittelsohn, Thapa and Landman [41]

1997

Key informant interviews, participant observation, unstructured pilot observations, focus group discussions, and structured pile sorts

105 HH

Six rural villages, with a mixture of agricultural and non-agricultural occupations.

Men aged 18 to 50 years, and women aged 18 to 50 years, including menstruating, pregnant, lactating, and postpartum women.

Analysis method of qualitative results not reported

• Cultural beliefs - Men were considered the least vulnerable and therefore had the fewest dietary restrictions, unless they were ill. Older people considered vulnerable and believed to require strengthening foods. Some pregnant women mentioned preferentially eating animal products due to ‘craving’. Post-partum women avoided ‘cold’ foods and ‘indigestible’ foods like wheat bread, peanuts, soybeans and corn porridge. They preferentially ate certain ‘hot’ foods like fish and millet roti. Lactating women avoided fresh green leafy vegetables that were perceived as ‘cold’ and believed to cause arthritis, swelling and other illnesses.

• Status - Women’s status increased by having children. Before childbearing, young married women had low status and were subtly discouraged from eating special foods like animal products and certain fried foods. Men’s higher status was “recognised in many ways, including household food behaviour”

Allocation of ‘special’ foods

 Madjdian and Bras [40]

2016

In-depth interviews

30 individuals

Two Himalayan communities from Humla district.

Female respondents (15 Buddhist; 15 Hindu Dalit or Chhetri) of reproductive age (aged 15 to 49 years). Selected respondents who were pregnant or had been pregnant at least once before.

Inductive coding based on a conceptual framework, using bottom-up and top-down coding to allow new themes to emerge.

• Beliefs about ‘fair share’ / Religion - Buddhist households allocated food according to appetite; this was not reported in Hindu households.

• Cultural beliefs and food habits - Certain foods believed to cause skin allergies. Eating order was associated with eating less.

• Food security - Food insecure households did not adhere to food proscriptions due to a lack of food

Food allocation

 Morrison, J. et al. Formative research to inform the development of interventions to tackle low birth weight in the rural plans of Nepal. In preparation.

Unpublished observations

Interviews and focus group discussions

25 women, 2 groups.

One district in Terai.

25 young daughters-in-law from marginalised groups living in extended families, one focus group discussion with men, and one with Female Community Health Volunteers who were mothers-in-law. Most (90%) respondents were Hindu. Respondent age not reported

Descriptive content analysis. Data were copied from transcripts into columns of 15 descriptive emergent categories.

• Status - Respondents reported that men ate more because they had higher status and so deserved to.

• Interpersonal relationships - Husbands may hide food for their pregnant wives, disrespecting the mother-in-law.

• Household structure - Married women who visited or lived at their maternal homes had fewer food restrictions.

• Economic contributions – Manual labourers were perceived to deserve more

• Cultural food beliefs – pregnant women ate less (fear of full stomach harming the baby)

• Household income – no effect of food being bought vs grown on food decisions.

Allocation of food generally, and also of ‘special’ foods

 Panter-Brick and Eggerman [64]

1997

Semi-structured interviews

120 heads of household

Population of high and low caste Indo-Nepalese and Tibeto-Burmese ethnic groups from four Panchayats in two districts.

Sampled households to ensure proportional representation of large and small land-holding farmers

Age of respondents not reported.

Analysis method of qualitative results not reported.

• Food shortages / Ethnicity - Indo-Nepalese household used discrimination against women as a coping mechanism during food shortages whereas Tibeto-Burmese households did not.

Food allocation