Exploring Behavioral Biases in Contraceptive Decision-Making: Evidence from a Field Experiment in Urban Malawi

How to cite

Karra, Mahesh; Zhang, Kexin; Ngwira, Bagrey; Longwe, Abiba. 2019. "Exploring Behavioral Biases in Contraceptive Decision-Making: Evidence from a Field Experiment in Urban Malawi." AEA RCT Registry. AEARCTR-0004194. https://doi.org/10.1257/rct.4194-2.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Mahesh Karra* , Kexin Zhang , Bagrey Ngwira , Abiba Longwe

* Primary investigator

Registry ID
AEARCTR-0004194
DOI
https://doi.org/10.1257/rct.4194-2.0
First registered
Last updated
Status
On going
Trial dates
2019-07-23 – 2022-12-31
Countries
Private
JEL codes
I15, I18, I31, I32, J13, J21
Registry record
http://www.socialscienceregistry.org/trials/4194

Abstract

This study seeks to investigate some of the behavioral mechanisms that play a role in contraceptive decision-making through a field experiment, in which we manipulate the specific architectural conditions under which women are counseled on family planning. The study population is married women of reproductive age who have a need for family planning in Lilongwe, Malawi. Our approach is motivated by evidence suggesting that women’s stated and realized preferences for contraception are likely to be both malleable and highly sensitive to a range of biases. To this end, we explore how key biases are characterized through a woman’s revealed preferences and decisions about a method. We test the following two hypotheses:
1. Women’s stated preferences for and choices of contraception are sensitive to the number of method choices that are presented to them, consistent with prior evidence on choice overload. Specifically, decision-making under the presence of too many choices may either be determined heuristically, in which choices over methods are driven by key method attributes (e.g. effectiveness in preventing pregnancy, incidence of side effects, etc.) or may be avoided altogether (decision deferral).
2. Giving women the choice whether or not to involve their husbands or partners as part of the counseling process is likely to affect 1) women’s decision-making around and choice of a contraceptive method; 2) the extent to which women are able to realize their preferences for family planning.

Intervention

All participants will receive a counseling session on contraceptive methods and family planning. Following receipt of counseling, women in all treatment arms (including the control arm) will be asked to reveal their preferred choice of method. Following this elicitation, women in all arms of the study will be offered the following package of services:
1. A free taxi ride to our partner family planning clinic in Lilongwe, the Good Health Kauma Clinic, which has been previously identified as a high quality family planning clinic that provides the full range of contraceptive methods.
2. A free consultation with a family planning service provider at the Good Health Kauma Clinic.
3. Free family planning services at the Good Health Kauma Clinic.

These services have already been tested and implemented in the field as part of a recently completed family planning study in Lilongwe (NHSRC Protocol No. 16/7/1628). To increase the salience of the offer, the package of services will be time-limited for one month following the counseling session. By offering the package of three services above, which together aim to reduce the barriers to accessing family planning services at our partner clinic, we will be able to elicit a measure of demand for family planning and will also be able to observe how a woman’s demand (her final choice of method) aligns with her initially stated preferences for family planning immediately following counseling.

A follow-up survey will be conducted one month following a woman’s counseling session. Outcome data that captures a woman’s preferences and decision-making around contraception will be collected at baseline survey, pre-counseling session, post-counseling session, and at every clinic visit. In situations where women do not show up at the clinic, our enumerators will make phone calls or home visits to conduct end-line surveys with the women and ask about their preferences and decision-making around contraception.

Experimental design

Participants taking part in this study will receive a survey will ask about their marriage, pregnancy status, fertility preferences, and views on family planning, maternal and child health, and reproductive health. Participants will then receive a second survey after two months at a local clinic, by phone, or at the participant’s home.

Design details

Randomization method
Randomization will take place after the baseline interview. We will randomize women to intervention and control arms using prepared codes in Stata such that intervention assignment is balanced according to the following baseline characteristics: neighborhood/household cluster, distance to the nearest family planning clinic, number of living children, months since last live birth, current use of family planning, age of marriage, educational attainment, and household wealth.
Randomization unit
Units of randomization will be implemented at an individual level.
Clusters
7-9 areas in Lilongwe, Malawi
Observations
The key study population is 700 married women of reproductive age in Lilongwe, Malawi. For the initial part of our study, we will recruit women who: 1. Are married at baseline 2. Age between 18 and 35 at screening 3. Live in the city of Lilongwe (permanent residents) 4. Are currently non-pregnant and did not give birth in the 6 months prior to the initial screening 5. Have neither been sterilized nor have had a hysterectomy 6. Have given birth to at least one child (one live birth) in their lifetime 7. Live with their husbands at the time of the screening. Women who successfully meet these inclusion criteria and who consent to participate in the main part of our study will be recruited. In addition, no two eligible women will be enrolled from the same household. If multiple women from the same household are potentially eligible to be recruited into this part of our study based on the four criteria inclusion above, we shall choose the youngest eligible woman from the household to participate. In addition, we will interview husbands of women who are invited to participate in the counseling intervention.
Arms
100 women in T0 (Control group), 200 in T1, 200 in T2, and 200 in T3. The descriptions of treatment groups T1-T3 can be seen from "Intervention" above.
Primary outcomes
Our main outcomes of interest (stated preferred method of contraception immediately following counseling, intention to use or not use contraception immediately following counseling) aim to capture women's immediate stated preferences for contraceptive methods. In addition, by observing how women may seek family planning following their counseling visit, we will examine the short-run stability of their stated preferences and the extent to which these stated preferences are subsequently realized in the face of other barriers to use. Data examining the realization of women’s eventual contraceptive preferences (e.g. contraceptive uptake, eventual method choice) will be collected during clinic visits and at follow-up across the three arms.

Cited by

No citation data available for this registration yet.

Keywords