Using Religion to Influence Female Labor Force Participation in Pakistan

The intervention aims at exogenously influencing the fathers’ (or any male decision-maker of the household) religiosity. The intervention is simple, providing religious information concerning positive attitude toward FLFP to arm1, that concerning potentially negative attitude toward FLFP to arm2, and no information to arm3. The religious information provided will follow the consultation by the religious scholar.

2024-01-19

2024-03-31

We have two data points, (i) right after the intervention (donation game response and follow-up survey response), (ii) endline survey (6 months after the intervention). We interpret that the "End Points" means (ii), however, for the sake of completeness, we describe key outcome variables at both (i) and (ii).

(i) Parental religious knowledge (survey response), Parental attitude toward FLFP (donation game outcome)

(ii) Young women's participation in work outside the home

(i) Primary outcomes will be measured by the questions in the follow-up survey right after the intervention.

1. Parental religious knowledge will be measured based on the answers to the following questions and answers.
R1. How would you rate your Islamic knowledge about female employment?
1=Expert, 2=Good, 3=Neither good nor bad, 4=Bad, 5=Never studied
R2. Do you think Islam allows female to participate in the Halal labour activities?
1=Yes, 2=No, 3=Unsure
R3. Do you know that the first wife of the Holy Prophet (peace be upon him) Hazrat Khadija (RA) was a very successful business woman?
1=Yes, 2=No

-> The outcome variable about the religious knowledge will be constructed (a) as a binary variable, i.e., proportion of the respondents who answered 1 or 2 to the question R1 and 1 to the questions R2 and R3. And we will construct (b) the sum of these answers (taking the values 0-3).

2. Parental attitude toward FLFP will be measured by the donation game outcome, i.e., how much the respondent is willing to donate to a NGO working for enhancement of FLFP.

(ii) Primary outcome at the endline will be the binary variable that takes the value of 1 if she works outside the home. Girls' paid-work participation will be measured by Yes/No household income questions in Section 2.

Parental attitude toward FLFP (based on the subjective questions), Girls' empowerment, Girls' attitude toward FLFP

Secondary outcomes will be measured by the questions in the endline survey ((ii) 6 months later) except for some which will be measured by the ones in the (i) follow-up survey.

1. Parental attitude toward FLFP will be measured based on the answers to the following questions.
(from (i) follow-up survey)
F1. Do you think it a good idea for women in general to work outside the home?
1= Yes, 2= Yes, only if it is white collar job, 3= No
F2. Would you allow your daughter to work outside the home?
1= Yes, 2= Yes, only if it is white collar job, 3= No

-> The outcome variable will be constructed (a) as ordinal variable, i.e., the ordered answers to the questions F1 and F2, respectively. We will construct (b) binary variables taking the value of 1 if the answers to the questions F1 and F2 are 1, respectively, and (c) binary variables taking the value of 1 if the answers to the questions F1 and F2 are 1 or 2, respectively. The sum of these binary variables are also consructed as ordinal variables taking the values 0-2 in both cases ((d) for the sum of (b) and (e) for the sum of (c)).

(from (ii) endline survey)
DW1. Would you allow your daughter to work outside the home?
1= Yes, 2= Yes, only if it is white collar job, 3= No
DW3. How far would you allow your daughter to travel to get a job?
1= Home-based work only, 2= Within the union, 3= Within the tehsil, 4= Within the district, 5= Within the division, 6= Another division, 7= Another province, 8= Abroad

-> The outcome variable will be constructed (a) as ordinal variables, i.e., the ordered answers to the questions DW1 and DW3, respectively. We will construct (b) a binary variable taking the value of 1 if the answer to the question DW1 is 1, (c) a binary variable taking the value of 1 if the answer to the question DW3 is 2 and above, and (d) a binary variable taking the value of 1 if the answer to the question DW3 is 3 and above. The sum of these binary variables are also consructed as ordinal variables taking the values 0-2 in either cases ((e) for the sum of (b) and (c), (f) for the sum of (b) and (d)).

GA10.Women should not work outside but should rather take care of the family.
GA11.Women should not work outside even if other household members (e.g., sisters) can take care of the family.
GA12.Income earned by a women should be remitted to her husband.
GA13.A woman should ask her husband for permission if she wants to work outside.
GA14.If a girl works outside home, that increases her chances of finding a better groom.
GA15.If a girl works outside home, that decreases the amount of dowry that the parents need to pay.
To above questions, the respondent answers 1=strongly disagree, 2=disagree, 3=neutral, 4=agree, 5=strongly agree.

-> The outcome variable will be constructed (a) as binary variables taking the value of 1 if the answers to the questions above are two less conservative values, (b) the sum of these binary variables taking the values 0-6, and (c) using factor analysis.

2. Girls' empowerment will be measured based on the answers to the following questions.
The answer 5=I'm usually at home (do not go out) to the question D1. To which level do you observe purdah? (Though purdah itself is not a measure of empowerment, the answer 5 shows the level of girl's mobility)
D2. Please tell us your educational aspiration.
1=I do not intend to pursue higher education, 2=I plan to complete master’s or higher degree.

-> The outcome variable will be constructed (a) as a binary variable taking the value of 1 if the answer to the question D1 is not 5, (b) a binary varialbe taking the value of 1 if the answer to the question D2 is 2, and (c) the sum of these binary variables taking the values 0-2.

3. Girls' attitude toward FLFP will be measured based on the answers to the following questions.
D3. Please tell us your job aspiration in the future. Do you want to work for pay?
1=Yes, 2=No, 3=Not sure.
D3.1 If Yes above, which occupation do you have in your mind?

-> The outcome variable will be constructed (a) as a binary variable taking the value of 1 if the answer to the question D3 is 1. We will construct some (b) categorical (by occupations) variables based on the answers to the question D3.1.

1. Conduct the baseline survey to the eligible household. Eligibility requires that the household has at least one girl who should be unmarried, completed 14 years of education, and is not currently working for pay outside the home. The sampling is based on convenience sampling and details are provided below.

Baseline questionnaire consists of 6 sections: Section 1 (household roster), Section 2 (income), Section 3 (assets), Section 4 (attitudes toward FLFP and gender), Section 5 (religiosity), Section 6 (questions to daughters). Sections 1 to 5 should be responded by male household members (fathers or any male decision makers such as elder brother, uncle, or grandfather) who has a decision-making power. Section 6 is the questionnaire to eligible daughters.

Conduct the baseline questions up to the section 5 to the male household member, followed by the section 6 to the daughter.

2. Randomization of treatment assignment
33% probability for each treatment arm.

3. Detach the intervention and follow-up questions from the baseline survey, and conduct the intervention and follow-up questions independently from the baseline survey.

4. Proceed the donation game (the instruction will be given in the protocol).

5. Right after the donation game, conduct the follow-up survey to the male respondent.

6. (Around 6 months after the intervention) Conduct the endline survey which includes the same questions in the baseline survey Section 2 (income), Section 4 (attitudes toward FLFP and gender), Section 5 (religiosity), and Section 6 (questions to daughters)

Not available

Computer-based randomization across three arms.

Household

No

63 union councils in 4 tehsils: Lahore Cantt, Lahore City, Model Town, Shalimar Town

500 households. Eligible households are those with at least one unmarried daughter who completed 14 years of education and were not currently work full time.

167 households (1/3 of 500 households)

In terms of the rate of FLFP, the expected control mean is 0.25 following the latest ILOSTAT 2019 (the rate of college degree women who work outside the home is 25%). The minimum detectable effect size is 0.09 ppt with 0.34 standard deviation.