How do concessions in national parks impact development? 🌱
In our new article, we use matching and diff-in-diff to evaluate socioeconomic effects of visitor in Brazilian parks.
Results challenge assumptions.
Read: doi.org/10.1080/1368...
#Tourism #NationalParks #diffindiff
Abstract: This study used a difference-in-differences approach to examine the effects of The Kalamazoo Promise (KPromise) on college stopout during students’ first year of postsecondary enrollment, a stopout at any time, reenrollment after a stopout, and credential completion after a stopout. Results indicated that for Promise-eligible high school graduates who started college within 12-months, KPromise decreased first-year stopout for all students by -16% - which included strong effects for female students and students of color. Additionally, KPromise produced a considerable increase in first year stopout for white students (+30%). Overall stopout declined post-Promise (-70%) and the effect was consistently negative across all subgroups. KPromise also led to an overall increase in student return after a stopout (+14%); however, there was a negative effect for students of color (-74%). For students that experienced a stopout and later returned to college, KPromise produced an increase in credential attainment (+38%) by 10 years from high school graduation. Attainment effects were relatively consistent across the studied groups. Although generally non-significant, our findings suggest that KPromise is overall helping students persist, encouraging some students to reengage, and revealing that some gaps in persistence and completion remain between racial, gender, and socioeconomic groups.
After a few years of letting this sit on the shelf - I am finally looping around to getting this #DiffinDiff #KalamazooPromise paper focused on stop out, reenrollmet, and degree attainment within 10 years cleaned up with some updates, and back in cycle.
@isabelmcmullen.bsky.social - updates soon.
R2: ‘I’m no expert, but here’s a shallow take.’
Also R2: ‘Report 5 staggered Diff-in-Diff estimators or perish!’
Stat nerds, rejoice—we’re all just trying to 2x2 our way out of chaos.
#DiffinDiff #StatsHumor #EconometricsForTheWin #TrustTheProcessButVerify
In Feb 25 I'll teach a course on Longitudinal Data Analysis at @komex.bsky.social @unikonstanz.bsky.social
We'll start with basics:
➡️what is it good for #Design
➡️how to handle & visualize the data #HandsOn
But we also cover analyses: fixed effects, random effects, & (a little) #DiffInDiff
⬇️⬇️⬇️
6. We also conducted *several* sensitivity analyses (also checking robustness of #DiffInDiff assumptions): we fit comparative #ITS regression models w/ #GEE, #PropensityScore-weighted models, & augmented #SyntheticControl models w/ state-level summaries (proportions or means)...
3. Our primary analysis used #DiffInDiff methods {Callaway & Sant'Anna, @pedrosantanna.bsky.social} that accommodate staggered exposures; we used 2-group, pre-vs-post models for each outcome at each year, with #Medicaid *non*-expansion states as comparators and time-varying expansion status...
