Educational Migration Prospects and Human Capital Investments

The intervention aims to inform students in Uganda about the Malengo program. The merit-based program offers mentoring and financial support to young adults in Uganda who want to complete a university degree in Germany and cannot otherwise afford to study at university (see https://malengo.org/ for more information). The intervention targets O- and A-level students in their final year who are about to sit their final exams. Ordinary-level (O-level) education consists of four years of lower secondary schooling at the end of which students undertake UCE (Uganda Certificate of Education) exams. Advanced-level (A-level) education consists of two years of upper secondary schooling at the end of which students undertake UACE (Uganda Advanced Certificate of Education/university entrance) exams.

The intervention has four bundled components. First, immediately after the baseline interviews, candidate students watch a 16-minute video that summarizes important aspects of the Malengo program and portrays current Malengo scholars. Field staff brings projectors and speakers to each school to ensure good video and sound quality. The video screening ends with a short Q&A session, which allows students to ask questions about the program.

Second, immediately after the video screening, field staff facilitates discussions among students. The discussions take place in small groups and last for about 30 minutes. Students are asked to reflect on the video, set goals for themselves, and discuss whether and how they might become Malengo scholars.

Third, before leaving the school, field staff hang a Malengo poster in a prominent spot on the school premises. The poster aims to remind students of the intervention.

Fourth, shortly after the school visit, field staff send a link to a short video summarizing the Malengo program to all contacts provided by surveyed students. The video aims to inform parents or other primary caregivers about Malengo and give students a chance to discuss this opportunity with them.

2022-07-11

2022-08-31

We will analyze whether the intervention changes students’ expectations and intentions concerning international educational migration and whether these changes translate into increased human capital investments and higher aspirations.

We list our primary outcomes below, marked with a number sign (#). For primary outcomes based on an index, we will also report treatment effects on the individual index components (but our main focus remains on the respective index).

1) Test scores
- Lower/upper secondary school (UCE/UACE) exam score #

2) Index of migration intentions #
- Would you like to study at a university abroad?
- Are you planning to apply for programs that help you study abroad?
- In the next five/seven years, would you like to live here in Uganda or elsewhere?
- Where do you WANT to live when you are 35 years old? (dummy for abroad)

3) Index of expected migration status #
- On a scale from 0-100, what do you think is the percent chance that you will get a degree from a university abroad in the near future?
- On a scale from 0-100, what do you think is the percent chance that you will live abroad in the next five/seven years?
- From zero to hundred, what is the percent chance that you will be living in Europe at the age of 35 with university education?

4) Expected employment outcomes in Europe
a) Suppose, hypothetically, you had completed university education in Europe. From zero to hundred, what is the percent chance that you will be working at the age of 35 if you lived in Europe? # (for interpretation, we will also compare the treatment effect on this outcome to the treatment effect on two auxiliary outcomes: (i) What is the percent chance that you will be working at the age of 35 if you lived in Uganda and had completed secondary school with a UACE certificate in Uganda? (ii) What is the percent chance that you will be working at the age of 35 if you lived in Uganda and had completed university education in Europe?)
b) Suppose, hypothetically, you had completed university education in Europe. How much do you think you will earn in a typical month at the age of 35 if you lived in Europe? # (for interpretation, we will also compare the treatment effect on this outcome to the treatment effect on two auxiliary outcomes: (i) How much do you think you will earn in a typical month at the age of 35 if you lived in Uganda and had completed secondary school with a UACE certificate in Uganda? (ii) How much do you think you will earn in a typical month at the age of 35 if you lived in Uganda and had completed university education in Europe?)

5) Index of aspirations #
- Until what educational degree would you like to study?
- How much money would you like to earn when you are 35 years old? Please give me a number for the monthly income/salary you would like to earn.
- At which age would you like to have your first child?

Our secondary outcomes focus on mechanisms and include knowledge about educational migration programs and the returns to international educational migration, perceived benefits and costs of international educational migration, study effort, effort spent on planning the future, agency, self-efficacy, and subjective wellbeing.

We list our secondary outcomes below, marked with a plus sign (+). Outcomes that measure mechanisms are marked with an asterisk (*).

1) Manipulation check based on immediate follow-up and endline survey +
a) Individual items to be analyzed as before-after comparison for treated students and comparison between treated and non-treated students:
- On a scale from 0 to 10, where 0 means not important at all and 10 extremely important, how important is your performance in the lower/upper secondary school (UCE/UACE) exams for your future quality of life?
- Would you like to study at a university abroad?
- On a scale from 0-100, what do you think is the percent chance that you will get a degree from a university abroad in the near future?
b) Which aspect of the Malengo program have students understood? Share of correct responses among treated students for each individual item:
- What is the Malengo program?
- Which of the following is NOT a requirement to apply for the Malengo program?
- What subjects are the Malengo scholars allowed to study?
- Which of the following is NOT an option that Malengo scholars use to support themselves?
- Out of 100 university graduates in Germany, how many of them find jobs after graduation?
c) What do students take away from the intervention? Tabulate responses for each item:
- I can relate to one of the Malengo scholars in the video.
- I can imagine being a Malengo scholar.
- Studying abroad would increase my chances of getting a well-paid job.
- Studying abroad would be an opportunity to acquire new skills and broaden my horizons.
- Studying abroad would allow me to meet new people and experience new things.
- Studying abroad would be fun.
- Studying abroad would allow me to become independent from my family.
- Studying abroad would be challenging.
- What is the *one* thing that you take away from the video? (grouped by main themes)
d) Share of treated students who have visited the Malengo website

2) Knowledge (first stage)
- Have you heard of any programs that help you to study abroad? [If yes]: Which programs? (dummy for mentioning Malengo) +
- What are the average monthly earnings of a worker with university education in Europe? +
- What are the average monthly earnings of a worker with university education in Uganda? +
- What are the average monthly earnings of a worker with only an upper secondary school certificate (UACE) in Uganda? +
- Ratio of average monthly earnings of a worker with university education in Europe to average monthly earnings of a worker with only an upper secondary school certificate (UACE) in Uganda +

3) Index of perceived benefits/costs of international educational migration +
a) Index of perceived benefits from educational migration *
- Studying abroad would increase my chances of getting a well-paid job.
- Studying abroad would be an opportunity to acquire new skills and broaden my horizons.
- Studying abroad would allow me to meet new people and experience new things.
- Studying abroad would be fun.
- Studying abroad would allow me to become independent from my family.
b) Index of perceived costs of international migration (reverse coded) *
- Studying abroad would be challenging.
- Studying abroad would be expensive and hard to finance.
- Number of reasons mentioned which might discourage you to live, study, or work abroad

4) Index of study effort +
- On a scale from 0 to 10, where 0 means not important at all and 10 extremely important, how important is your performance in the lower/upper secondary school (UCE/UACE) exams for your future quality of life? *
- How many days have you arrived late for school or skipped some classes?
- How many hours of private tutoring have you received in total?
- How many practice tests, including mock exams offered by the school, have you completed?
- How many hours did you spend on preparing for the lower/upper secondary school (UCE/UACE) exams outside of class in total? *

5) Other outcomes related to test scores
- Difference between expected and realized lower/upper secondary school (UCE/UACE) exam score +
- Dummy for two principal passes in upper secondary school (UACE) exam +

6) Index of parental inputs +
- My parents/caregivers provide emotional support for my educational efforts.
- My parents/caregivers support my educational efforts by involving me less in other tasks such as housework, childcare, or helping out in the family business/farm.

7) Index of teacher inputs +
a) Index of teacher inputs as reported by students *
- My teachers care a lot for me and are genuinely pleased to see me succeed in my studies.
- My teachers support me in planning my future after school.
- My teachers make an extra effort to prepare me for the lower/upper secondary school (UCE/UACE) exams.
- My teachers show an interest in every student’s learning.
b) Index of teacher inputs as reported by teachers *
- The learning outcomes of my students are more important than their social connections for their future.
- Good teaching has an impact on students beyond the school, leading to a better life, good jobs and higher earnings for students.
- I care for my students and want them to succeed in their studies.
- I help my students in planning their future after school.

8) Efforts spent on planning the future
a) Index of effort spent on future abroad +
- I looked for information about studying abroad.
- I looked for information about careers abroad.
- I looked for information about programs that help me study abroad.
- I looked for information about migrating to another country.
- I talked to my family or friends about studying abroad.
- I talked to my family or friends about working abroad.
- I talked to my teachers about studying abroad.
- I talked to my teachers about working abroad.
b) Index of effort spent on future in Uganda +
- I looked for information about studying in Uganda.
- I looked for information about careers in Uganda.
- I looked for information about programs that help me study in Uganda.
- I talked to my family or friends about studying in Uganda.
- I talked to my family or friends about working in Uganda.
- I talked to my teachers about studying in Uganda.
- I talked to my teachers about working in Uganda.

9) Agency and self-efficacy
a) Index of agency +
- Some people believe that individuals can decide their own destiny, while others think that it is impossible to escape a predetermined fate. Please tell me which comes closest to your view on this scale on which 1 means “everything in life is determined by fate” and 10 means “people shape their fate themselves.” *
- Trying hard at school will help me get a good job.
- Trying hard at school will help me get into a good university.
b) Index of self-efficacy +
- If someone opposes me, I can find the means and ways to get what I want.
- It is easy for me to stick to my aims and accomplish my goals.
- I am confident that I could deal efficiently with unexpected events.
- Thanks to my resourcefulness, I know how to handle unforeseen situations.
- I can remain calm when facing difficulties because I can rely on my coping abilities.
- I can usually handle whatever comes my way.

10) Direction of interest
- Which subject would you like to study in a university for a degree program? +
- In which occupation do you want to work when you are 35 years old? +

11) Subjective wellbeing
- All things considered, how satisfied are you with your life as a whole these days on a scale from 1 to 10? +

Additional longer-run outcomes (based on data from additional follow-up surveys conditional on obtaining more funding)
- Mental health
- Applied to Malengo or other scholarships for university studies (if so, was application successful)
- Enrolled in A-level (for O-level students) or university (for A-level students) (if so, which subject)
- Labor market outcomes such as employment status, income, and type of job
- Place of residence (moved, Greater Kampala, abroad)
- Marital status and children
- Effects on siblings

Our randomization is based on a complete list of schools in Uganda that are also used as exam centers (obtained from the Uganda National Examinations Board). We exclude (i) schools with fewer than 25 students in the 2020 university entrance (UACE) examinations to ensure a sufficiently large number of students per school, and (ii) schools in districts that are not eligible for the Malengo program (about half of the districts outside of Greater Kampala, see https://malengo.org/eligible-districts). We then randomly draw schools and use stratified randomization at the level of schools to assign the intervention. We create strata by grouping schools by terciles of (i) number of students in the 2020 university entrance (UACE) examinations, (ii) share of female students, and (iii) average exam scores. The intervention is the same for all treated schools. There is one treatment and one control group.

Our main data source is a survey that we will conduct in schools. We will interview twelve randomly drawn O-level students plus one of their teachers, as well as twelve randomly drawn A-level students plus one of their teachers, and follow them over time. The baseline survey is scheduled to be conducted in July and August 2022. The intervention will be implemented immediately after the baseline interviews. In addition, treated students and teachers will complete a short follow-up survey immediately after the intervention. The endline survey is scheduled to be conducted in September and October 2022, just before students sit their final lower/upper secondary school (UCE/UACE) exams. In addition, we are planning to collect grades from administrative sources based on individual registration numbers that students will share with us in interviews.

We will use the following equation to estimate the impact of the intervention:

Y_i = a + b T_i + X’_i c + u_i

where Y_i is the outcome variable of interest (based on data collected in endline interviews), T_i the treatment dummy indicating whether student i has been exposed to the intervention and X_i a vector of control variables (based on data collected at baseline). We will control for the baseline value of the respective outcome variable wherever possible. We will also include strata fixed effects. We will use the double post-lasso estimation proposed by Belloni et al. (2014) to select additional control variables. The set of potential control variables that we will use for this estimation will include all baseline variables, the dates of the baseline and endline interview as well as enumerator and district fixed effects. We will use dummies to indicate missing baseline data and replace missing values with zero, including both variables in the set of potential control variables for the double post-lasso estimation. This procedure will ensure maximum sample size and power. We will use OLS to estimate the equation above and cluster standard errors at the school level. In case there are large imbalances in the number of students interviewed per school, we will consider reweighing observations so schools have equal weight. In case of differential attrition, we will consider using inverse probability weights and Lee/Manski bounds.

We will test for effect heterogeneity along five dimensions: (i) student type (O-level vs. A-level), (ii) gender (male vs. female), (iii) ability (based on baseline grades), (iv) location (Greater Kampala defined as the districts of Kampala, Mukono, Wakiso vs. rest of country), (v) education of main caregiver (A-level and higher vs. below A-level). We will do so by interacting the treatment dummy with a variable that captures the respective dimension of heterogeneity.

We will rely on outcome indices, as defined by Kling et al. (2007), to reduce the number of hypotheses. These indices are averages of standardized z-scores, where all outcomes are recoded so that higher values correspond to “better” or pro-migration outcomes. In addition, we will adjust for multiple testing across indices of primary outcomes controlling for the false discovery rate (Anderson 2008). We will not adjust for multiple testing across indices of secondary outcomes or outcomes within domains as we put less emphasis on these outcomes.

We will make the following adjustment to variables if needed. First, some variables might have minimal variation and thus reduce the power to detect an impact. We will therefore exclude all variables for which 90 percent of subjects have the same value. Second, we will winsorize continuous variables (e.g., incomes) at the 95th percentile and carry out the inverse hyperbolic sine transformation to reduce the influence of outliers.

We will consider replacing any methods mentioned above with superior methods if they become available by the time of conducting the analysis.

Note that we follow the guidance provided by Duflo et al. (2020) on pre-analysis plans and only use these fields in the AEA RCT Registry rather than a separate document.

Not available

Stratified randomization by computer

School

Yes

400 schools

9,600 students (12 O-level students and 12 A-level students per school * 400 schools) 800 teachers (1 O-level teacher and 1 A-level teacher per school * 400 schools)

200 treatment schools, 200 control schools