Experimental Design
I will implement a randomized controlled trial of a leadership training workshop with incentives among youth interested in running for political office. The experiment will be conducted in Sorsogon Province, Philippines, where the experimental design has been piloted this last year (Aug. 2013--Sep. 2014). It will be implemented in partnership with the Angara Centre for Law and Economics (ACLE) and Innovations for Poverty Action (IPA). The intervention will occur a year before the October 2016 Village Level Elections. Standing in and winning election are the key outcomes of interest.
Sampling. The sample will consist of N=900 Filipino residents of Sorsogon Province, qualified and interested in running for an elective post in the village council as first-time candidates. (See statistical power calculations in a separate section below.)
Treatments. Study participants will be randomly assigned to a control condition (no workshop) or one of two workshop treatments---a workshop with unconditional incentives or a workshop with conditional incentives. Assignment will take account of baseline measures of public service motivation (PSM) to maximize statistical power.
Workshop Interventions. The leadership training workshop has three goals: (1) provide basic leadership skills; (2) give participants a chance to run a simulated village council meeting in which they will brainstorm a flagship project; and (3) serve as a screening mechanism by which study participants can reveal their 'type' (i.e. competence and honesty, as will be explained below), as they perform in the various workshop tasks and in the simulation of the council meeting.
To prevent differential take-up across treatment arms, participants invited to the workshop will not be informed of the incentives until after the workshop is over. Moreover, unbeknownst to participants, performance in the workshop will be monitored and evaluated, and will be assigned scores based on a scoring rubric. This is to prevent differential behavior among participants across treatment arms, due to awareness of being monitored.
Workshop staff members will be assigning scores. If a participant is assigned to the group with unconditional incentive, then he shall receive the incentive regardless of his performance score. On the other hand, if a participant is assigned to a group with conditional incentive, then he shall receive the incentive only if his performance score is above a pre-determined cutoff, which is known only to the Principal Investigator (PI).
The incentive is a combination of two things: (1) a plaque of merit awarded at the end of the workshop, (2) and a few (less than 10 pcs.) of standard-sized campaign posters, should they decide to file an official certificate of candidacy.
Data Collection and Measurement of Political Selection. The field experiment begins with a call for applications by ACLE to the leadership training workshop. The study team will hand out posters and invitation letters to schools, offices of incumbent village officials, as well as to different civic organizations, to capture as many applicants as possible. Posters and letters will provide a general description of the workshop, application guidelines, as well as information on a pre-screening session. Applicants will be informed that selection to the workshop is random.
The pre-screening session is designed to measure personal characteristics of applicants before any random assignment to treatment groups occurred. The session will involve a series of tests designed to measure two broad categories of personal characteristics: competence and honesty. Competence is subcategorized into public service motivation (using Perry's PSM Index (Perry, 1996)), aptitude (based on Wechsler's Test of Memory for Digit Span (Wechsler, 1987), and on a Matrices Test adapted from (Ariely et al., 2009)), personality (using the Big 5 Personality Index (John, 1990; John, Naumann and Soto, 2008)), and aspiration (Kasser and Ryan's Aspiration Index (Kasser and Ryan, 1996)). Honesty will be measured in terms of participants' propensity to cheat in a simple roll-a-die game a la Fischbacher and Follmi-Heusi (2013); Hanna and Wang (2013)).
Applicants who successfully completed the pre-screening session will be enrolled as study participants and randomly assigned into the treatment groups described above. Political selection will be measured in several stages after randomization, as follows:
Stage 1: Selection into the workshop.
Stage 2: Filing a certificate of candidacy. (After the workshop, but before the elections.)
Stage 3: Indicator for winning an elective post and vote-share margin of victory. (After elections.)