Survey Recall Error Truncation
RCTReview
In longitudinal labor market surveys, respondents frequently report experimental program work days as post-program outside employment during the immediate post-intervention month, contaminating outcome tracking with experimental implementation data.
Picture this
Imagine tracking someone's home-cooked meal routine right after they return from a week-long cooking workshop. During the first few days home, they might accidentally count meals cooked at the workshop as part of their home routine. Excluding the first month from data collection acts like a buffer period, ensuring that only genuine independent home cooking is measured.
What the evidence says
Excluding month 1 eliminates survey recall error where participants erroneously report recruiter-assigned experimental work as post-intervention employment, ensuring that estimated daily wage increases ($3.83 to $4.19 per day, p < 0.05) reflect independent labor market returns.
- Who was studied
- N = 227 urban male youth job seekers (aged 18+, secondary school educated) in Lilongwe, Malawi.
- How
- Methodological truncation excluding month 1 recall data from the 9-month follow-up survey reference period to isolate independent 8-month post-intervention employment outcomes.
What to do
Exclude the immediate post-program time window from longitudinal follow-up survey panels to prevent participant confusion between intervention tasks and independent employment.
From the source
"To deal with this survey recall error, we exclude the first month of recall data and rely only on the 8-month period beginning 1 month after the completion of the work related to the experiment."
b23e842f-67cd-47d7-9b15-b6b6c8e557a2-Employment Exposure- Employment and Wage Effects in Urban Malawi.pdf