US study ```r # candidate preference # original term result sig <char> <char> <char> 1: (Intercept) b = 42.67 [41.99, 43.34], p < .001 *** 2: conditionproTrump b = 2.88 [2.02, 3.74], p < .001 *** 3: lean_bidentrump_1Z b = 40.94 [40.35, 41.53], p < .001 *** 4: topicZ b = -0.71 [-1.39, -0.04], p = .038 * 5: conditionproTrump × lean_bidentrump_1Z b = 0.77 [0.04, 1.50], p = .038 * 6: conditionproTrump × topicZ b = 0.93 [0.07, 1.79], p = .035 * 7: lean_bidentrump_1Z × topicZ b = -0.12 [-0.71, 0.47], p = .689 8: conditionproTrump × lean_bidentrump_1Z × topicZ b = -0.11 [-0.84, 0.62], p = .768 # missing post values filled with pre values term result sig <char> <char> <char> 1: (Intercept) b = 42.78 [42.15, 43.41], p < .001 *** 2: conditionproTrump b = 2.70 [1.89, 3.51], p < .001 *** 3: lean_bidentrump_1Z b = 41.04 [40.49, 41.59], p < .001 *** 4: topicZ b = -0.66 [-1.30, -0.03], p = .040 * 5: conditionproTrump × lean_bidentrump_1Z b = 0.72 [0.03, 1.41], p = .040 * 6: conditionproTrump × topicZ b = 0.86 [0.05, 1.67], p = .037 * 7: lean_bidentrump_1Z × topicZ b = -0.11 [-0.67, 0.44], p = .688 8: conditionproTrump × lean_bidentrump_1Z × topicZ b = -0.10 [-0.79, 0.59], p = .773 ``` ```r variable term estimate std.error statistic p.value <fctr> <char> <num> <num> <num> <num> 1: income attrit -0.236 0.086 -2.754 0.006 # attriters have lower income 2: vote_chance_1 attrit -0.207 0.096 -2.159 0.031 # attriters are less likely to vote 3: per_imp_1 attrit 0.163 0.081 2.025 0.043 # attriters think personality of politician matters more 4: collegeD attrit -0.160 0.088 -1.828 0.068 5: politics_involve attrit 0.149 0.084 1.775 0.076 6: politics_interest attrit 0.133 0.078 1.705 0.088 7: economic_conserv attrit -0.142 0.086 -1.654 0.098 8: policy_imp_1 attrit -0.103 0.081 -1.273 0.203 9: orient attrit -0.085 0.087 -0.969 0.332 10: social_conservE attrit 0.046 0.087 0.531 0.596 11: therm_repub_1 attrit -0.042 0.087 -0.488 0.626 12: genderD attrit -0.040 0.084 -0.473 0.637 13: vote_1_trumpZ attrit -0.039 0.083 -0.463 0.643 14: age attrit -0.037 0.089 -0.420 0.674 15: god attrit 0.030 0.086 0.354 0.724 16: economic_conservE attrit 0.028 0.089 0.316 0.752 17: therm_dem_1 attrit 0.026 0.087 0.300 0.764 18: vote_1_harrisZ attrit -0.022 0.084 -0.260 0.795 19: ethnicityD attrit -0.017 0.085 -0.206 0.837 20: pol attrit 0.013 0.087 0.150 0.881 21: ai_general_1 attrit -0.011 0.082 -0.136 0.892 22: social_conserv attrit 0.007 0.087 0.081 0.936 23: lean_bidentrump_1E attrit 0.001 0.083 0.013 0.990 24: lean_bidentrump_1 attrit 0.001 0.084 0.010 0.992 variable term estimate std.error statistic p.value ``` canada study ```r # candidate preference # original term result sig <char> <char> <char> 1: (Intercept) b = 34.21 [32.40, 36.02], p < .001 *** 2: conditionproPoilievre b = 10.05 [7.52, 12.58], p < .001 *** 3: lean1Z b = 32.59 [30.57, 34.62], p < .001 *** 4: strategyD b = 2.83 [0.30, 5.36], p = .029 * 5: conditionproPoilievre × lean1Z b = 1.11 [-1.51, 3.73], p = .406 6: conditionproPoilievre × strategyD b = -5.88 [-9.30, -2.45], p = .001 *** 7: lean1Z × strategyD b = -0.08 [-3.00, 2.84], p = .957 8: conditionproPoilievre × lean1Z × strategyD b = 1.72 [-1.92, 5.36], p = .354 # missing post values filled with pre values term result sig <char> <char> <char> 1: (Intercept) b = 34.47 [32.79, 36.16], p < .001 *** 2: conditionproPoilievre b = 9.34 [6.99, 11.68], p < .001 *** 3: lean1Z b = 32.69 [30.72, 34.65], p < .001 *** 4: strategyD b = 2.87 [0.55, 5.19], p = .015 * 5: conditionproPoilievre × lean1Z b = 1.38 [-1.09, 3.85], p = .275 6: conditionproPoilievre × strategyD b = -5.68 [-8.82, -2.53], p < .001 *** 7: lean1Z × strategyD b = 0.40 [-2.33, 3.13], p = .776 8: conditionproPoilievre × lean1Z × strategyD b = 1.07 [-2.30, 4.44], p = .534 ``` poland study ```r # candidate preference # original term result sig <char> <char> <char> 1: (Intercept) b = 37.78 [35.31, 40.25], p < .001 *** 2: conditionN b = 9.94 [6.67, 13.20], p < .001 *** 3: lean1Z b = 31.46 [28.62, 34.31], p < .001 *** 4: strategygeneric b = 1.06 [-1.98, 4.09], p = .495 5: strategynofacts b = 3.50 [0.46, 6.55], p = .024 * 6: strategynoinstr b = 0.75 [-2.48, 3.97], p = .649 7: conditionN × lean1Z b = 2.12 [-1.31, 5.55], p = .226 8: conditionN × strategygeneric b = -1.57 [-6.04, 2.89], p = .489 9: conditionN × strategynofacts b = -7.73 [-11.76, -3.71], p < .001 *** 10: conditionN × strategynoinstr b = -1.81 [-6.31, 2.68], p = .429 11: lean1Z × strategygeneric b = 3.47 [0.22, 6.72], p = .036 * 12: lean1Z × strategynofacts b = 3.16 [-0.25, 6.57], p = .069 . 13: lean1Z × strategynoinstr b = 1.31 [-2.20, 4.82], p = .464 14: conditionN × lean1Z × strategygeneric b = -4.61 [-9.14, -0.07], p = .046 * 15: conditionN × lean1Z × strategynofacts b = -2.22 [-6.44, 1.99], p = .300 16: conditionN × lean1Z × strategynoinstr b = -1.38 [-5.93, 3.17], p = .553 # missing post values filled with pre values term result sig <char> <char> <char> 1: (Intercept) b = 38.21 [35.93, 40.49], p < .001 *** 2: conditionN b = 9.16 [6.21, 12.11], p < .001 *** 3: lean1Z b = 31.92 [29.28, 34.57], p < .001 *** 4: strategygeneric b = 1.01 [-1.78, 3.79], p = .478 5: strategynofacts b = 3.23 [0.39, 6.06], p = .026 * 6: strategynoinstr b = 0.72 [-2.30, 3.74], p = .641 7: conditionN × lean1Z b = 2.03 [-1.11, 5.17], p = .206 8: conditionN × strategygeneric b = -1.57 [-5.56, 2.42], p = .441 9: conditionN × strategynofacts b = -6.97 [-10.64, -3.29], p < .001 *** 10: conditionN × strategynoinstr b = -1.88 [-6.01, 2.25], p = .372 11: lean1Z × strategygeneric b = 3.25 [0.24, 6.27], p = .035 * 12: lean1Z × strategynofacts b = 2.82 [-0.37, 6.02], p = .084 . 13: lean1Z × strategynoinstr b = 0.77 [-2.54, 4.08], p = .649 14: conditionN × lean1Z × strategygeneric b = -4.10 [-8.21, 0.02], p = .051 . 15: conditionN × lean1Z × strategynofacts b = -2.02 [-5.91, 1.86], p = .307 16: conditionN × lean1Z × strategynoinstr b = -0.99 [-5.21, 3.24], p = .646 ```