17 summary(toxicity_calibrated)
18 qplot(labelr$toxic,type='hist')
26 hist(log(fbyg$weight))
29 fbyg$weight[fbyg$weight != 1]
30 fbyg[fbyg$weight != 1]
31 fbyg[,fbyg$weight != 1]
32 fbyg[[fbyg$weight != 1]]
33 fbyg[fbyg$weight != 1,]
35 summary(toxicity_calibrated)
45 labelr[labelr$data=='yg']
46 labelr[,labelr$data=='yg']
47 labelr[labelr$data=='yg',]
49 summary(toxicity_calibrated)
51 yg3[,['toxic','toxic_pred']]
52 yg3 %>% select('toxic','toxic_pred')
53 yg3 |> select('toxic','toxic_pred')
55 yg3[,c('toxic_pred','toxic')]
56 corr(yg3[,c('toxic_pred','toxic')])
57 cor(yg3[,c('toxic_pred','toxic')])
58 cor(yg3[,c('toxic_pred','toxic')],na.rm=T)
59 cor(yg3[,c('toxic_pred','toxic')],rm.na=T)
60 ?cor(yg3[,c('toxic_pred','toxic')],use=
62 cor(yg3[,c('toxic_pred','toxic')],use='all.obs')
64 cor(yg3[,c('toxic_pred','toxic')],use='complete.obs')
65 cor(yg3[,c('toxic_pred','toxic')],use='complete.obs',method='spearman')
94 preds1$diff - preds$diff
96 preds1$diff - preds1$diff
97 preds1$diff - preds$diff
98 preds1$diff - preds$diff
99 preds1$diff - preds$diff
100 preds1$diff - preds$diff
119 yg3[,.(toxic_pred,toxic_var)]
120 yg3[,.(toxic_pred,toxicity_2_pred_sigma,toxicity_1_pred_sigma)]
121 yg3[,.(toxic_pred,toxicity_2_pred_sigma,toxicity_1_pred_sigma,cov(toxicity_2_pred,toxicity_1_pred))]
132 yg3[,.(toxic_pred,toxicity_2_pred_sigma,toxicity_1_pred_sigma,cov(toxicity_2_pred,toxicity_1_pred))]
133 yg3[,.(toxic_pred,toxicity_2_pred_sigma,toxicity_1_pred_sigma,toxic_var)]
134 yg3[,.(toxic_pred,toxicity_2_pred_sigma,toxicity_1_pred_sigma,toxic_var,toxic_sd)]
142 stargazer(w1,w2,w3,w4,w5,t1,t2,t3,t4,t5, type="text",
143 keep = c("cond1","meantox","cond1:meantox","Constant"),
144 keep.stat=c("n","adj.rsq"),
146 dep.var.labels = c("DV = Willingness to comment","DV = Toxicity of YG respondent comments"),
147 covariate.labels = c("Treatment (top comments shown)",
148 "Average toxicity of top comments",
149 "Treatment $\times$ top comments toxicity",
151 add.lines = list(c("Article fixed effects","No","No","No","Yes","Yes","No","No","No","Yes","Yes")),
152 star.cutoffs = c(0.05,0.01,0.005),
153 notes = "Standard errors are clustered at the respondent level.",
154 column.labels = c("(1)","(2)","(3)","(4)","(5)","(6)","(7)","(8)","(9)","(10)"),
168 labelr[,'comment_id']
180 yg3[,.(.N),by=.(toxic,fb)]
190 is.data.table(labelr)
197 setwd("presentations/ica_hackathon_2022/")
202 example_2_B.plot.df[(variable=='x') && (m < 1000)]
203 example_2_B.plot.df[(variable=='x') && (m < 1000)]
204 theme_set(theme_default())
205 theme_set(theme_minimal())
206 theme_set(theme_classic())
207 example_2_B.plot.df[(variable=='x') && (m < 1000)]
208 example_2_B.plot.df[(variable=='x') && (m < 1000),unique(method)]
218 example_2_B.plot.df$method %>% unique
219 example_2_B.plot.df$method |> unique
220 example_2_B.plot.df$method |> uniq
221 unique(example_2_B.plot.df$method)
222 example_2_B.plot.df$method
223 example_2_B.plot.df$method
224 example_2_B.plot.df$method
225 example_2_B.plot.df$method
226 example_2_B.plot.df <- r$example_2_B.plot.df
229 setwd("presentations/ica_hackathon_2022/')
230 setwd("presentations/ica_hackathon_2022/")
231 example_2_B.plot.df$method
232 example_2_B.plot.df$method
235 example_2_B.plot.df$method
236 example_2_B.plot.df$method
239 example_2_B.plot.df$method
240 example_2_B.plot.df$method
247 plot.df[,.N,by=.(N,m)]
248 plot.df[,.N,by=.(N,m,method)]
249 plot.df[variable=='x',.N,by=.(N,m,method)]
251 plot.df[(variable=='x') & (m < 1000) & (!is.na(p.true.in.ci))]
252 plot.df[(variable=='x') & (m != 1000) & (!is.na(p.true.in.ci))]
255 install.packages("ggplot2")
256 devtools::install_github("tidyverse/ggplot2")
269 plot.df[method="2SLS+gmm"]
270 plot.df[method=="2SLS+gmm"]
271 df <- example_2_B.plot.df
278 plot.df.example.2[m==50][method=2SLS+gmm]
279 plot.df.example.2[m==50][method==2SLS+gmm]
280 plot.df.example.2[(m==50) & (method==2SLS+gmm)]
281 plot.df.example.2[(m==50) & (method=="2SLS+gmm")]
285 plot.df.example.3[N=25000]
286 plot.df.example.3[N==25000]