- see [[230719_164713 descriptives|descriptives]] comparing pre-treatment distributions across waves model: t0 ~ t0 * condition, fixed effect block, day, cluster SEs on block ```r # users per wave (before any exclusion) campaign N 1: exp1-pilot 32888 2: exp2-big 81290 3: exp3-trucker 23833 4: exp4-deepstate 19042 ``` # t0 active proportion days: >= 0.9 - all four waves included ```r # users per wave > x[t0_prop_days_w_tweets >= 0.9, .N, campaign] campaign N 1: exp1-pilot 23655 2: exp2-big 11772 3: exp3-trucker 16208 4: exp4-deepstate 11721 ``` ![[_mainfig_count_conditionC 7.png]] ```r threshold estimate CI CI_low CI_high evid_ratio post_prob type 1: 40 0.00032 0.95 -0.01814 0.01966 0.94595 0.48611 Intent-to-treat effect 2: 45 -0.00173 0.95 -0.01994 0.01673 1.33611 0.57194 Intent-to-treat effect 3: 50 -0.00056 0.95 -0.01809 0.01773 1.10266 0.52441 Intent-to-treat effect 4: 55 0.00009 0.95 -0.01676 0.01677 0.98837 0.49707 Intent-to-treat effect 5: 60 0.00228 0.95 -0.01373 0.01874 0.64518 0.39216 Intent-to-treat effect 6: 65 0.00069 0.95 -0.01515 0.01592 0.85714 0.46154 Intent-to-treat effect 7: 70 -0.00356 0.95 -0.01860 0.01130 2.10078 0.67750 Intent-to-treat effect 8: 75 -0.00290 0.95 -0.01802 0.01171 1.85643 0.64991 Intent-to-treat effect 9: 80 -0.00171 0.95 -0.01606 0.01248 1.44993 0.59182 Intent-to-treat effect 10: 40 0.00221 0.95 -0.02806 0.03209 0.79324 0.44235 Treatment effect on the treated 11: 45 -0.00104 0.95 -0.03077 0.02849 1.12789 0.53005 Treatment effect on the treated 12: 50 0.00072 0.95 -0.02804 0.02923 0.92530 0.48060 Treatment effect on the treated 13: 55 0.00161 0.95 -0.02597 0.02869 0.82707 0.45267 Treatment effect on the treated 14: 60 0.00469 0.95 -0.02064 0.03136 0.56436 0.36076 Treatment effect on the treated 15: 65 0.00267 0.95 -0.02220 0.02783 0.71799 0.41792 Treatment effect on the treated 16: 70 -0.00404 0.95 -0.02859 0.02004 1.68222 0.62717 Treatment effect on the treated 17: 75 -0.00325 0.95 -0.02692 0.02118 1.51209 0.60192 Treatment effect on the treated 18: 80 -0.00128 0.95 -0.02446 0.02192 1.18537 0.54241 Treatment effect on the treated ``` ## interaction effect ![[_mainfig_count_conditionC-t0SC 7.png]] # t0 active proportion days: >= 0.5 - all four waves included ```r # users per campaign > x[t0_prop_days_w_tweets >= 0.5, .N, campaign] campaign N 1: exp1-pilot 29094 2: exp2-big 16347 3: exp3-trucker 21515 4: exp4-deepstate 15632 ``` ```r threshold estimate CI CI_low CI_high evid_ratio post_prob type 1: 40 -0.00380 0.95 -0.02174 0.01385 1.95094 0.66112 Intent-to-treat effect 2: 45 -0.00468 0.95 -0.02270 0.01273 2.31112 0.69799 Intent-to-treat effect 3: 50 -0.00408 0.95 -0.02104 0.01265 2.12769 0.68027 Intent-to-treat effect 4: 55 -0.00215 0.95 -0.01850 0.01412 1.52406 0.60381 Intent-to-treat effect 5: 60 -0.00184 0.95 -0.01755 0.01371 1.44245 0.59057 Intent-to-treat effect 6: 65 -0.00356 0.95 -0.01825 0.01142 2.11442 0.67891 Intent-to-treat effect 7: 70 -0.00747 0.95 -0.02182 0.00721 5.34065 0.84229 Intent-to-treat effect 8: 75 -0.00633 0.95 -0.02067 0.00811 4.16096 0.80624 Intent-to-treat effect 9: 80 -0.00584 0.95 -0.01953 0.00803 3.94315 0.79770 Intent-to-treat effect 10: 40 -0.00468 0.95 -0.03424 0.02366 1.66516 0.62479 Treatment effect on the treated 11: 45 -0.00651 0.95 -0.03550 0.02087 2.08261 0.67560 Treatment effect on the treated 12: 50 -0.00543 0.95 -0.03290 0.02158 1.86697 0.65120 Treatment effect on the treated 13: 55 -0.00248 0.95 -0.02876 0.02387 1.33693 0.57209 Treatment effect on the treated 14: 60 -0.00217 0.95 -0.02713 0.02321 1.31007 0.56711 Treatment effect on the treated 15: 65 -0.00450 0.95 -0.02828 0.02021 1.79916 0.64275 Treatment effect on the treated 16: 70 -0.01096 0.95 -0.03442 0.01344 4.46299 0.81695 Treatment effect on the treated 17: 75 -0.00924 0.95 -0.03294 0.01351 3.58216 0.78176 Treatment effect on the treated 18: 80 -0.00857 0.95 -0.03124 0.01418 3.39416 0.77242 Treatment effect on the treated ``` ![[_mainfig_count_conditionC 6.png]] ## interaction effect ![[_mainfig_count_conditionC-t0SC 6.png]]