- media-bias-fact-check (MBFC) results - 3216 rated domains - seems like anything below 60 is actually bad (relative to fact-checker ratings) ![[s20220310_174640.png]] ![[s20220310_175010.png]] Missing values ```r # missing values condition variable n_na N n_na_prop domain_type 3: c mbfc_mean_t1_retweet 1939 16473 0.1177 mbfc 4: t mbfc_mean_t1_retweet 1986 16415 0.1210 mbfc ``` # Basic model ```r > m1 <- feols(mean_t1 ~ conditionC * mean_t0C | block, dt1[domain_type == "overall"], vcov = "HC1") > m1 OLS estimation, Dep. Var.: mean_t1 Observations: 32,888 Fixed-effects: block: 5,424 Standard-errors: Heteroskedasticity-robust Estimate Std. Error t value Pr(>|t|) conditionC -0.130612 0.095103 -1.373373 0.16965 mean_t0C 0.522977 0.007389 70.782558 < 2.2e-16 *** conditionC:mean_t0C 0.008265 0.010761 0.768050 0.44246 ``` # 3-bin model ```r # bin means and sizes > dt1[, .(mean_t0 = mean(mean_t0), n = .N), keyby = .(bin = mean_t0_bin)] bin mean_t0 n 1: _1 52.73154 10963 2: _2 64.72663 10962 3: _3 77.46787 10963 > m3.1 <- feols(mean_t1 ~ conditionC * mean_t0_bin | block, dt1[domain_type == "overall"], vcov = "HC1") > m3.1 OLS estimation, Dep. Var.: mean_t1 Observations: 32,888 Fixed-effects: block: 5,424 Standard-errors: Heteroskedasticity-robust Estimate Std. Error t value Pr(>|t|) conditionC -0.340247 0.193581 -1.757644 0.078819 . mean_t0_bin_2 6.290744 0.143933 43.705957 < 2.2e-16 *** mean_t0_bin_3 13.662622 0.161858 84.411212 < 2.2e-16 *** conditionC:mean_t0_bin_2 0.522196 0.265948 1.963523 0.049596 * conditionC:mean_t0_bin_3 0.188127 0.248824 0.756066 0.449616 ``` ![[dv_mbfc_retweets_oct17-oct24_model_interact_3bins 1.png|700]] ## model comparison ```r m101 <- feols(mean_t1 ~ conditionC * mean_t0_bin | block, dt1[domain_type == "overall"]) m102 <- feols(mean_t1 ~ mean_t0_bin | block, dt1[domain_type == "overall"]) test_wald(m102, m101) > test_wald(m102, m101) Name | Model | df | df_diff | F | p ---------------------------------------------- m102 | fixest | 32886 | | | m101 | fixest | 32883 | 3.00 | 2.19 | 0.087 # model with condition # condition effect for each bin > m201 <- feols(mean_t1 ~ condition * mean_t0_bin | block, dt1[domain_type == "overall"]) > summary(m201, vcov = "HC1") OLS estimation, Dep. Var.: mean_t1 Observations: 32,888 Fixed-effects: block: 5,424 Standard-errors: Heteroskedasticity-robust Estimate Std. Error t value Pr(>|t|) conditiont -0.340247 0.193581 -1.757644 0.078819 . # bin 1 conditiont 0.181949 0.171614 1.06022 0.289054 # bin 2 conditiont -0.152120 0.153334 -0.992078 0.32117 # bin 3 ``` # Threshold analysis - see [[220305_142915 fact-checker threshold analysis]] for explanation - negative condition effects! ![[dv_mbfc_thresholds_interact 1.png|800]]