- see [[230719_184036 bayesian fixed-effects meta for studies 1 3 4]]
# main effect (and moderation by t0)
- quasipoisson models predicting winsorized 95% # bad link RTs
```r
# meta-analytic treatment effect for different thresholds
threshold estimate se low high pval
1: 40 -0.0194 0.0125 -0.0438 0.0051 0.1210
2: 45 -0.0231 0.0122 -0.0471 0.0009 0.0594
3: 50 -0.0262 0.0120 -0.0498 -0.0026 0.0296
4: 55 -0.0222 0.0116 -0.0448 0.0005 0.0553
5: 60 -0.0217 0.0112 -0.0437 0.0003 0.0529
6: 65 -0.0218 0.0108 -0.0429 -0.0006 0.0435
7: 70 -0.0241 0.0106 -0.0448 -0.0033 0.0230
8: 75 -0.0228 0.0105 -0.0434 -0.0023 0.0296
9: 80 -0.0208 0.0102 -0.0408 -0.0007 0.0425
```
![[1689904854.png]]
![[1689904910.png]]
![[1689904941.png]]
## ATT
```r
# meta-analytic treatment effect for different thresholds
threshold estimate se low high pval
1: 40 -0.0318 0.0204 -0.0718 0.0081 0.1184
2: 45 -0.0379 0.0200 -0.0772 0.0013 0.0583
3: 50 -0.0425 0.0197 -0.0811 -0.0039 0.0310
4: 55 -0.0355 0.0189 -0.0725 0.0015 0.0603
5: 60 -0.0343 0.0183 -0.0702 0.0016 0.0608
6: 65 -0.0341 0.0176 -0.0686 0.0004 0.0525
7: 70 -0.0383 0.0173 -0.0723 -0.0044 0.0268
8: 75 -0.0361 0.0171 -0.0697 -0.0025 0.0350
9: 80 -0.0327 0.0167 -0.0654 0.0001 0.0504
```
![[1689904856.png]]
![[1689904912.png]]
![[1689904942.png]]