The (in)direct effect of a training program on women's employment -- experimental evidence from the health sector in Egypt

The goal of our study is to measure the impact of primary care training on labor outcomes and female empowerment, both for direct beneficiaries of the training as well as their social contacts. To that end, we collaborate with an Egyptian NGO---named Giving Without Limits (GLW)---that offers primary care training for young women in Southern Egypt. The training program targets approximately 650 young women (aged 18-29) in the two Governorates of Assiut and Sohag in Southern Egypt. The program prepares the beneficiaries for wage employment in private households of elderly people and geriatric care facilities, or to work as self-employed primary care specialists in rural areas. The training is conducted with the help of local medical schools in each province and runs for approximately 1.5 months, consisting of a combination of soft skills, technical skills, and on-the-job training. After graduation, trained participants are matched to potential employers.

2022-08-01

2023-05-31

1. Attitudes and decision-making index, consisting of the following elements:
– Attitudes towards female employment
– Respondent makes employment decisions herself
2. Job search and expectations index, consisting of the following elements:
– Days searched for job in past seven days
– Expected number of job offers in next four months
– Expected earnings in next four months
3. Employment index, consisting of the following elements:
– Currently being in paid employment
– Weekly work hours
– Weekly income

1. A variable indicating whether the main applicant has referred the social contact to a job
2. A variable indicating whether the main applicant and the social contact talked about the
training program
3. Frequency and type of social interactions between main applicant and social contact
4. Aspiration outcomes:
– Whether the respondent sees herself in employment in five years
– Whether the respondent would like her daughter to be employed
– Based on a question “Who are people from your village/community that you aspire to be like in your social network?”, we will define an indicator for whether the respondent names [name of original participant] among the people she aspires to be like.
– How much would you like your life to be similar to [name of original participant]? (scale 1-10)

We will evaluate the effects of the training intervention based on a clustered randomized controlled trial, where individuals applying to participate in the training program will be randomly assigned to a treatment or a control group. The unit of randomization will be the applicant's area of residence (hamlet in the local administrative language and referred to as `location' from now on), meaning that we randomize the location into treatment and control locations. Those individuals residing in treatment locations receive the opportunity to take-up the training; those residing in control locations will not get that opportunity. We stratify the randomization of the location by two variables: i) the number of applicants applying to the program in each location and ii) the tercile of the number of social contacts who were listed by the applicants in a given location (classified into bottom, middle, top tercile of number of social contacts listed). Stratifying by these two variables enforces more strongly a balance in both the number of applicants and the social contacts by whether a location is in the treatment or control group. In addition to analyzing the direct effect of the training program on the participants, we will further study the indirect impacts of the training program by following the close social contacts of program applicants.

Not available

randomization done in office by a computer

Location of residence of the applicant / her social contact (hamlet)

Yes

The expected sample size for the main sample of applicants is 1300 female applicants from 100-120 locations (hamlets). The expected sample size for the social network sample is 2500 female contacts from 100-120 locations (hamlets).

The expected sample size for the main sample of applicants is 1300 female applicants from 100-120 locations (hamlets). The expected sample size for the social network sample is 2500 female contacts from 100-120 locations (hamlets).

50-60 locations (hamlets) treated (primary care training), 50-60 locations (hamlets) control

See PAP