I’m working with panel data comprising several years of observations of schools. My DV is a proportion of exam passers but is not normally distributed, and many observations of the DV are > 0.8. A panel linear model using plm() is therefore inappropriate, so I am trying to treat the DV as a binary response […]

Categories

## Proportion vs. binary response with pglm

- Post author By Full Stack
- Post date October 23, 2020
- No Comments on Proportion vs. binary response with pglm

- Tags -38, -48, .022) X2 <- c(145, .062, .44, .74) X1 <- c(.05, "0:15, "021", "034", "041", "2018"], [ 55, [ 75, [[ 87, [148, 012, 031, 037, 055, 081, 1, 101, 112, 119, 145, 146, 150, 151) takers <- c(50, 152], 154, 155], 156, 2, 2019 ] }, 3, 35, 36, 4, 4) proportion <- c(.67, 4) year <- rep(c(2017, 43, 44) fails <- takers - passers data <- as.data.frame(cbind(id, 45], 52, 57, 60) passers <- c(34, 62, 66, 69, 77, 79, 81, 85, 88, 89, 90, 99], and many observations of the DV are > 0.8. A panel linear model using plm() is therefore inappropriate, any help on figuring out what the error is would be much appreciated. Comments on methodology are also welcome, but a random effects approach does not make theoretical sense given my research question and a Hausman test indicates I need fixed effects an, but I do not see a why to use fixed effects with betareg(). I can also run this code using glmer() and setting a random intercept (1|id), data = data) #> Error in `.rowNamesDF<-`(x, data = data) And I am also familiar with an alternative to the treat-DV-as-binary approach, fails) ~ X1 + X2, fails)) pglm::pglm(cbind(passers, family = binomial(link = "logit"), I'm working with panel data comprising several years of observations of schools. My DV is a proportion of exam passers but is not normally di, index = c("ID", model = "within", namely the betareg() package which uses beta regression]2, of course, passers, proportion, so I am trying to treat the DV as a binary response and use logistic regression. I have counts of the numbers of test takers and passers. I, takers, value = value): duplicate 'row.names' are not allowed Created on 2020-10-21 by the reprex package (v0.3.0) I do not encounter an issue runn, x1, x2, Year