Bottom-up Anti-corruption: Applications of Market Segmentation for Demarketing Corruption
Description
This data is part of 'Bottom-up Anti-corruption: Applications of Market Segmentation for Demarketing Corruption' published in the Southeast Asian Journal of Economics. It includes 2 datasets collected from Thai population and randomized survey experiment participants at Chulalongkorn University between July 2020 and September 2020. The datasets are intended to support the segmentation analysis and randomized survey experiment presented in the accompanying paper, helping to identify behavioral and attitudinal clusters relevant to designing targeted anti-corruption interventions. The data collection cites approval from the Research Ethics Review Committee for Research Involving Human Research Participants, Group 2 (Social Sciences, Humanities, and Fine Arts), Chulalongkorn University, COA No. 052/2562. 01_baseline_survey consists of the face-to-face baseline survey (Phase 1, March 2020) dataset used for the Latent Profile Analysis (persona segmentation) reported in accompanying paper. 02_survey_experiment consists of the online randomized vignette/role-play experiment (Phase 2, July-September 2020, two rounds) dataset used to test H1-H4 reported in accompanying paper.
Files
Steps to reproduce
A. Baseline Survey - Table 1 is reproduced by tabulating Generation, Gender, Education, Occupation, and Personal Income. - Table 4 is reproduced by running a factor analysis on 41 items, dropping five items with loadings below 0.5, and computing the AVE, Construct Reliability, and Cronbach's alpha for the eight latent constructs using the remaining 36 items. - A separate factor analysis is run on the anti-corruption items to produce scores for the perception, prevention, persistence, and enforcement dimensions. - A Latent Profile Analysis is performed on the 719 samples using Mplus 8.3 to segment respondents based on the eight latent class variables. - Table 6 is reproduced by estimating models ranging from three to six groups, comparing their AIC, BIC, SABIC, and LRT statistics, and selecting the four-group model. - Table 7 is reproduced by computing the mean of each of the eight latent variables within each of the four persona groups and across the total sample. - Table 8 is reproduced by computing the mean and standard deviation of the four anti-corruption sub-dimensions within each of the four persona groups. B. Survey Experiment - Table 2 is reproduced by tabulating role from both the full Round 1 response file and the paired Round 1/Round 2 response file. - Table 9 is reproduced by running a factor analysis on the 19 anti-corruption items, dropping three of them to leave 16, and running a second factor analysis on the 14 masculinity, uncertainty-avoidance, and personal-norms items, dropping four of them to leave 10. - The manipulation check is reproduced by comparing masculinity scores and uncertainty-avoidance scores between the high and low role conditions across both rounds. - Hypotheses I and II are reproduced by comparing the four anti-corruption dimension scores across role conditions using a between-subject design. - Hypothesis III is reproduced by splitting the sample into high and low groups at the mean of personal norms and comparing the four anti-corruption dimension scores between these two groups. - Hypothesis IV is reproduced by comparing each respondent's Round 1 and Round 2 anti-corruption scores using the paired response file in a within-subject design.
Institutions
- Chulalongkorn UniversityBangkok, Bangkok