- [prereg](https://doi.org/10.17605/OSF.IO/K9YNS) - n: 490 - [[20211206_093907 results - first 250 with counterbalancing - study 1|lucid study with counterbalancing]] - three-way interaction `veracity:demrep_c:bfi_c` in model 3 is our preregistered Bayes-factor test (BF must be either < 0.1 or > 10) # Basic models ```r > sum2(m0c) term res 1: (Intercept) b = 0.42 [0.38, 0.46], p < .001 2: demrep_c b = -0.03 [-0.07, 0.01], p = .110 3: veracity b = 0.03 [0.00, 0.06], p = .028 4: demrep_c:veracity b = 0.01 [-0.02, 0.05], p = .452 ``` # Model 1 (false headlines) ```r m1_1 <- glm(share ~ demrep_c * bfi_c, family = binomial, data = dt1[veracity == 0]) m1_1c <- coeftest(m1_1, vcovCL(m1_1, ~ responseid + headline_id, NULL, fix = FALSE)) m1_1c z test of coefficients: Estimate Std. Error z value Pr(>|z|) (Intercept) -0.334158 0.085955 -3.8876 0.0001013 *** demrep_c -0.135811 0.084723 -1.6030 0.1089337 bfi_c -0.344286 0.070232 -4.9021 9.48e-07 *** BF10 (alternative against null) = 2187 (alternative 2187x more likely) demrep_c:bfi_c 0.082608 0.073235 1.1280 0.2593266 BF01 (null against alternative) = 40 (null 40x more likely) ``` ![[bf_model1 5.png]] # Model 2 (false headlines) ```r m2_1 <- glm(share ~ demrep_c * (bfi_c + bfi_e + bfi_a + bfi_n + bfi_o + age + gender + edu + attention_score + ctsq_aot), family = binomial, data = dt1[veracity == 0]) m2_1c <- coeftest(m2_1, vcovCL(m2_1, ~ responseid + headline_id, NULL)) m2_1c z test of coefficients: Estimate Std. Error z value Pr(>|z|) (Intercept) -0.4246003 0.1013577 -4.1891 2.800e-05 *** demrep_c -0.0016335 0.0974291 -0.0168 0.986623 bfi_c -0.0916724 0.1051332 -0.8720 0.383228 bfi_e 0.3215759 0.0792559 4.0574 4.961e-05 *** bfi_a -0.1565917 0.1008670 -1.5525 0.120553 bfi_n 0.1306789 0.0918976 1.4220 0.155025 bfi_o -0.0082770 0.0984212 -0.0841 0.932979 age -0.3639037 0.0795577 -4.5741 4.783e-06 *** gender 0.1048974 0.0689965 1.5203 0.128428 edu 0.0329299 0.0702801 0.4686 0.639390 attention_score -0.1119604 0.0687920 -1.6275 0.103627 ctsq_aot -0.5920495 0.0919042 -6.4420 1.179e-10 *** BF10 = 13882962 demrep_c:bfi_c 0.0875519 0.1062531 0.8240 0.409943 BF01 = 52.67897 demrep_c:bfi_e -0.0859712 0.0846848 -1.0152 0.310015 demrep_c:bfi_a 0.0345281 0.0882050 0.3915 0.695463 demrep_c:bfi_n 0.1210913 0.0803796 1.5065 0.131941 demrep_c:bfi_o 0.1623594 0.0856716 1.8951 0.058074 . demrep_c:age -0.0182371 0.0792841 -0.2300 0.818074 demrep_c:gender 0.0453677 0.0751471 0.6037 0.546030 demrep_c:edu -0.1214864 0.0691395 -1.7571 0.078897 . demrep_c:attention_score -0.0612781 0.0757995 -0.8084 0.418847 demrep_c:ctsq_aot 0.2161046 0.0714941 3.0227 0.002505 ** BF10 = 1.303 ``` ![[bf_model2 5.png]] # Model 3 (false and true headlines) - all variables are centered (except `veracity`) ```r m3_1 <- glm(share ~ veracity * demrep_c * (bfi_c + bfi_e + bfi_a + bfi_n + bfi_o + age + gender + edu + attention_score + ctsq_aot), family = binomial, data = dt1) m3_1c <- coeftest(m3_1, vcovCL(m3_1, ~ responseid + headline_id, NULL, fix = TRUE)) m3_1c z test of coefficients: Estimate Std. Error z value Pr(>|z|) (Intercept) -0.4246004 0.1001107 -4.2413 2.222e-05 *** veracity 0.1764828 0.0867829 2.0336 0.041991 * demrep_c -0.0016335 0.0962766 -0.0170 0.986463 bfi_c -0.0916724 0.1048715 -0.8741 0.382042 bfi_e 0.3215759 0.0790141 4.0699 4.704e-05 *** BF10 = 37.773 bfi_a -0.1565917 0.1004060 -1.5596 0.118858 bfi_n 0.1306789 0.0916277 1.4262 0.153812 bfi_o -0.0082770 0.0980813 -0.0844 0.932747 age -0.3639038 0.0792391 -4.5925 4.380e-06 *** gender 0.1048974 0.0689680 1.5210 0.128270 edu 0.0329299 0.0703018 0.4684 0.639493 attention_score -0.1119604 0.0685794 -1.6326 0.102560 ctsq_aot -0.5920495 0.0912372 -6.4891 8.633e-11 *** BF10 = 13311215 veracity:demrep_c 0.0209703 0.0997464 0.2102 0.833484 veracity:bfi_c -0.0417054 0.0504126 -0.8273 0.408078 veracity:bfi_e -0.0600631 0.0314675 -1.9087 0.056297 . veracity:bfi_a -0.0319695 0.0423644 -0.7546 0.450471 veracity:bfi_n -0.0168106 0.0384176 -0.4376 0.661694 veracity:bfi_o 0.1474760 0.0507561 2.9056 0.003666 ** BF10 = 0.6511165 veracity:age 0.1986642 0.0382275 5.1969 2.027e-07 *** veracity:gender 0.0779460 0.0270181 2.8850 0.003915 ** veracity:edu 0.0009371 0.0349414 0.0268 0.978604 veracity:attention_score -0.0041516 0.0375730 -0.1105 0.912018 veracity:ctsq_aot 0.0271017 0.0622993 0.4350 0.663544 demrep_c:bfi_c 0.0875519 0.1058060 0.8275 0.407968 demrep_c:bfi_e -0.0859712 0.0841998 -1.0210 0.307236 demrep_c:bfi_a 0.0345281 0.0879294 0.3927 0.694556 demrep_c:bfi_n 0.1210913 0.0801457 1.5109 0.130817 demrep_c:bfi_o 0.1623594 0.0855308 1.8983 0.057662 . demrep_c:age -0.0182371 0.0790524 -0.2307 0.817551 demrep_c:gender 0.0453677 0.0748590 0.6060 0.544486 demrep_c:edu -0.1214864 0.0688994 -1.7632 0.077859 . demrep_c:attention_score -0.0612781 0.0755362 -0.8112 0.417227 demrep_c:ctsq_aot 0.2161046 0.0713421 3.0291 0.002453 ** BF10 = 0.9394558 veracity:demrep_c:bfi_c -0.0344228 0.0593720 -0.5798 0.562062 BF01 = 88.42902 veracity:demrep_c:bfi_e -0.0257107 0.0483012 -0.5323 0.594519 veracity:demrep_c:bfi_a 0.0360511 0.0454585 0.7931 0.427746 veracity:demrep_c:bfi_n 0.0217580 0.0470228 0.4627 0.643572 veracity:demrep_c:bfi_o -0.0883531 0.0287278 -3.0755 0.002101 ** BF10 = 1.082373 veracity:demrep_c:age 0.0260097 0.0416118 0.6251 0.531934 veracity:demrep_c:gender -0.0540934 0.0415936 -1.3005 0.193421 veracity:demrep_c:edu 0.0157688 0.0313753 0.5026 0.615255 veracity:demrep_c:attention_score 0.0664275 0.0322343 2.0608 0.039325 * BF10 = 0.07990714 veracity:demrep_c:ctsq_aot -0.1058099 0.0403929 -2.6195 0.008805 ** BF10 = 0.2954283 ``` ![[bf_model3 6.png]] Opposite effect?!? Higher conscientious conservatives **more likely** to share fake news?? ![[Pasted image 20211208233813.png]]