Russia is an oppressive dictatorship. In such states, as UCLA and Yale scholars have explained, people falsify their preferences on political issues. Fear of bad consequences for “wrong” answers (harassment, imprisonment, or worse) is omnipresent 3/7 https
1,809 followers
16,709 followers
@AjayVerghese @AlexanderMkLee @VEDmechanic On the flip side the q I (try to) ask at the design stage for survey experiments is what is the randomization buying us here for measurement and inference. I love this paper by @graemedblair @aecoppock and Margare
50 followers
Como de praxe, seguem artigos pra dar mais corpo ao q foi dito: (1) https://t.co/qR2zuuRJrv (2) https://t.co/t0riQ848gm (3) https://t.co/LSBYMUWdNK
1,488 followers
The item count technique (or list experiment) can reduce social desirability bias. However, it requires large samples if the expected bias is rather small. E.g. if I expect a 5 percentage point bias and have N=500, a direct question is preferable (see http