Provision of Religious Information about Female Labor Force Participation to College Students in Pakistan

Lab Experiments: We will use lab experiments to conduct the proposed research. The controlled environment of the lab would enable us to identify the precise impact of Islam on Female Labor Force Participation (FLFP) attitudes, which can be difficult in a field where numerous other factors can possibly dilute this relationship.

We will use two types of Islamic teachings to establish the causal impact of Islam on FLFP attitudes. These Islamic teachings are already checked and confirmed by an Islamic scholar. One type encourages FLFP and the other potentially discourages it. While most scholars agree that Islam permits FLFP while remaining within the Islamic framework, most of the public does not know because these teachings do not get highlighted in society.

There will be one control and two treatment arms(T1 & T2). The control arm will not be exposed to any religious information. T1 will provide Islamic teachings emphasizing women’s right to work (positive frame). T2 will provide Islamic teachings about the rules and regulations women should follow if they work outside home (can be perceived negative). While T1 and T2 both contain religious information about FLFP, there is an important difference: T1 emphasizes women’s right to participate in labor force, while T2 emphasizes religious rules that must be followed by women participating in labor force. Therefore, T1 is inherently positive towards FLFP, while T2 can be perceived negative towards FLFP due to strict conditions that must be observed by women. These two conditions accurately capture the divided perceptions about FLFP in Pakistan, and therefore, are essential elements of our study. The control arm will not be exposed to any information. The exact experimental Instructions are provided in below. These instructions in Experimental Design will be provided to the participants in either T1 or T2 treatment arms.

2026-02-01

2026-02-21

1. FLFP attitudes measured by Dictator Game (DG) donations.
2. FLFP attitudes measured by List Experiments (LE).
3. Subjective FLFP attitudes, i.e., their attitudes toward FLFP in general and toward their future wife or own sister to work outside the home.
4. Willingness to hire/work with women, i.e., future willingness to hire or work with men and women, respectively.

Right after the intervention, following the DG framework, we will elicit monetary donations to a foundation working to promote FLFP in Pakistan. The amount donated will be our first incentive-compatible outcome variable. The DG procedure is explained in the detailed experiments below (4. DG).

Following DG, we conduct LE to elicit FLFP attitudes. The LE procedure is explained in the detailed experiments below (5. List experiments).

Outcomes 3 and 4, attitudes towards FLFP, are measured by the questions F1-F8 in the questionnaire (Appendix). Specifically, F1 measures their attitude toward general women working outside the home. F3 measures their attitude toward their own (future) wife/sister (if any) working outside the home. F5-F6 measure the gap of their willingness to hire men and women. F7-F8 measure the gap of their willingness to work with men and women.

5. Gender attitudes, such as responsibility for household chores, decision-making, autonomy, mobility, women's rights (vote, education), FLFP, marriage.
6. Religiosity, such as frequency of prayers, visit to mosques, level of Islamic knowledge and understanding.

Outcomes 5 and 6 are not, in principle, considered outcomes of the intervention. However, they will serve as controls if they are balanced between the treatment and control groups.

Outcomes 5, gender attitudes, are measured by the questions GA1-GA15 in the questionnaire (Appendix). The respondents answer each of the 15 statements using a 5-point Likert scale (1=strongly disagree to 5=strongly agree). GA2 and GA3 will be added to measure their attitudes toward women's mobility. GA5 and GA6 will be added to measure their attitudes toward patriarchy. GA7-GA9 will be added to measure their attitudes toward women's rights to education. GA10 and GA11 will be added to measure their attitudes toward women working outside the home. GA14 and GA15 will be added to measure their marriage market evaluation toward working brides. We will also use the gender attitudes index, which is the sum of GA1-GA15 measures; higher = more progressive.

Outcomes 6, religiosity, are measured by the questions R1-R15 in the questionnaire (Appendix). R1 (frequency of prayers) is the main variable to measure the respondents' ritual-based religiosity. We check the robustness of ritual-based measures using R2 (frequency of mosque visits) and R12 (frequency of Quran recitation). R5 (belief-based) and R6 (knowledge-based) will be used as other supplementary religiosity measures. R5 and R6 are 5-point Likert scale measures: self-declared religiosity (1 =Very religious; 5 =Not at all religious, reversed so higher = more religious) and knowledge-based: overall Islamic knowledge (1 =Expert; 5 =Never studied, reversed). Note that R8 and R11, i.e., the correct knowledge about FLFP in Islam, may be used as the outcomes of the intervention if they have sufficient variations. We will also use the religiosity index, which is the sum of all three measures; higher = more religious.

The lab experiments will be implemented with students at the National University of Sciences and Technology (NUST) located in Islamabad, Pakistan. Within NUST, we will implement the study with students of NUST Business School (NBS). A full-time faculty member of NBS has already made organizational arrangements for implementing our experiments. As our study intends to examine the impact of religious information on FLFP attitudes of young male members of the society, our participants in the experiments will be male students, both undergraduate and graduate. We will exclude female students as they fall outside the aims and objectives of the current study.

We will obtain schedule of those classes at NBS after which the classroom is exclusively available to us for at least one hour to conduct experiments. At the end of the selected classes, Co-PI (Hamza Umer) will enter the classroom and announce that he would like to conduct an economic experiment, clarify that participation is completely voluntary, and the decision to participate or leave will have no impact on student’s academic grades. Interested male students will stay in the classroom and participate in experiment.

On average, there are 50 students in every class at NBS, and approximately half are male. Our previous experience of implementing experiments at NBS shows that on average, three to five students from every class do not participate in the experiment and leave. Therefore, to obtain the required sample of 300 male participants, we will randomly select thirteen classes from the list of classes after which the classroom is available for at least one hour, and hope to obtain 300 participants. However, if our required sample is not achieved, we will again randomly select classes from the list and repeat this procedure until we obtain our required sample. In case we are unable to obtain the required sample from NBS, we will target male students in other departments of NUST by obtaining necessary permissions from those departments. As a backup, we have discussed the possibility of implementing experiments with faculty members of the NUST School of Social Sciences and Humanities (S3H), and they are happy to accommodate our request.

In the reserved classroom for experiments, Co-PI (Hamza Umer) will proceed by announcing the experiment and instructing that the interested students sit in an exam formation to ensure participants make decisions privately. Afterwards, an informed consent form (Appendix) will be distributed among the students. Once students read and sign it, the experimental instructions sheet containing 1 to 3 below (detailed) will be distributed. Once everyone reads the instructions, the decision sheet (4 - 6) will be distributed. Afterwards, the post-experiment survey (7 - 8) will be distributed. At the end, participants will be paid in private the fixed participation fee and the money based on their decision in DG and a signed payment slip (9) will be obtained. The experimental materials 1 – 2, and 4, 6– 9 are the same for treatment and control groups. The intervention component 3 is different for T1 and T2 arms, and absent for the control arm. The list experiments (LE) component 4 is different between LE treated and LE controlled.

Not available

Public lottery across three arms.

Individual

No

300 male university students

300 male university students

100 students control, 100 students positive frame treatment, 100 students negative frame treatment

We calculate the minimum detectable effect size using the STATA command 'power'. Although World Value Survey asks the questions related to the attitudes toward women working outside the home, there are no reported statistics for Pakistan. Therefore, we use the FLFP rate as a proxy for these attitudes. According to the latest ILOSTAT and most recent Pakistan Labour Force Survey (2024-25), the rate of college degree women who work outside the home is 25%. With 0.25 standard deviations for the control and treatment groups, type I error at 5% level, the statistical power at 80%, and a given sample size (N=100 per arm), the minimum detectable effect size is 10 ppt.