Abstract

This data is from a quantitative survey administered in 2023 to 2,000 married Nepali women and men from 4 provinces in the country about their own beliefs regarding norms-related behaviors, their expectations of how common it is for others in their social group to engage in those behaviors, and the expected social consequences surrounding those behaviors. It is the primary dataset used to author the working paper titled "Women’s Labor Force Participation in Nepal: An Exploration of The Role of Social Norms" - which presents rigorous evidence on whether and the extent to which social norms matter for women's labor force participation in Nepal.

Geographic coverage

The survey data includes a representative sample of households from 4 out of 7 provinces in Nepal: 1. Bagmati Province 2. Sudurpashchim Province 3. Madhesh Province 4. Gandaki Province

Analysis unit

Individual

Universe

The sampling frame is a list of all wards within each selected province.

Kind of data

Sample survey data [ssd]

Sampling procedure

Ward (cluster) selection: The sampling frame consisted of the list of all wards within each selected province. Each province comprises districts and within each district are municipalities (urban and rural municipalities) which are further broken down into wards – the smallest administrative units. The list of wards and their population figures were taken from the latest available 2021 Census. First, the universe of all districts was stratified by urban and rural to ensure greater statistical power for detecting differences between the 2 localities. The stratification by urban-rural proportionate to the population proportion of each group within a province resulted in a self-weighted sample, allowing for analysis of data at the province level and further at locality level within each province. To select the wards, a random start point was generated to negate any bias in the list and to provide an independent chance of selection from the list. The sampling method used here, probability proportionate to size (PPS), gives an independent chance of selection to each ward as per its population size, i.e., a higher chance of selection to wards with a higher population size.38 As a first step of random selection of wards, the cumulative frequency (CF) of the population of households in a ward was calculated. Since the unit of analysis for our study purpose was households having certain criteria and we expected the main outcome variables (social norms) to vary at household levels (as opposed to at an individual level), the household population figures served as the basis for sampling purpose (as opposed to the population size of individuals for a ward). Applying PPS, in the first step, the required number of wards were selected for Categories 1 and 2 households (households with working and non-working females respectively). Following this, the clusters allocated for Category 3 (households with migrant population) households were taken as a subset of the wards selected for Categories 1 and 2.

Selection of the random starting point within each ward during in-field random sampling of households: The selection of the random starting point within a PSU was done by the survey supervisors. For every ward, a predefined landmark for the starting point was chosen. The predefined landmark consisted of i) school, ii) health post, iii) central marketplace, or iv) ward office. The selection of a predefined landmark was the basis of the starting point which was made at the central office. The chosen landmark for every cluster was rotated to account for randomization and to avoid interviewer bias. Once the landmark was chosen, each enumerator used the spin-the-bottle method to randomize the direction in which the survey took place. After starting with a household, enumerators used a skip interval to survey every third household in rural and every fifth household in urban areas. Once the household was chosen, the interviewer used the screener to ascertain the eligibility as per the category quota set aside for them.

Respondent selection: The respondents were selected based on a screener instrument that surveyed the following factors: 1. Gender: Since the views about social norms and labor market outcomes vary by gender, both males and females within a household were interviewed. However, for households with migrant men, only the women were interviewed. 2. Age group: For all women, the screener was applied so as to ensure that only women within the economically active age range, i.e., between the ages of 18-59 years were interviewed. For spouses of female respondents, they had to be at least 18 years of age with no maximum age limit set. 3. Ethnicity: Nepal has more than a hundred ethnic groups residing across the country, and thus the major 8-10 groups are captured in the sample. The other objective of applying a screener for monitoring ethnic composition was to ensure that marginalized ethnic groups such as Dalits were sufficiently represented in the survey. 4. Marital Status: Only married men and women were interviewed since marriage and the responsibilities that come with are sown to impose greater social barriers and restrictions on mobility and work of females. 5. Location: The survey was carried out in both rural and urban locations in a total of 4 provinces. 6. General demographic factors include: • Perceived economic situation: Low to middle-income • It was ensured that both the respondents (male and female for Categories 1 and 2) and female respondent for Category 3 belonged to the second generation of the selected household (for example, not the in-laws residing in a household but their son and his wife.

Mode of data collection

Computer Assisted Personal Interview [capi]