220712_131018 beta - changes in engagement and dwell across trials

- model: `engage ~ b0c * b1c` (see [[220622_173908 engagement with focal headlines]]) - `c`: mean-centered - `bc1`: slope for item_order effect Changes in engagement with focal headlines (b1) **not** associated with importance/newsworthiness ratings after controlling for intercept (b0). - no significant effects for will smith study - significant effect of b0 for both climate change and monkeypox - larger b0 associated with higher importance/newsworthiness ratings - larger b1 associated with negative effects (but less consistent) - b0:b1 interaction effects ```r # study 1 will smith > summaryh(lm(importance ~ b0c * b1c, d3_b1_dwell[topic=="willsmith"])) term results 1: (Intercept) b = 2.37, SE = 0.10, t(172) = 24.12, p < .001, r = 0.88 2: b0c b = −0.04, SE = 0.09, t(172) = −0.48, p = .635, r = −0.04 3: b1c b = −3.39, SE = 5.11, t(172) = −0.66, p = .509, r = −0.05 4: b0c:b1c b = 4.19, SE = 3.24, t(172) = 1.29, p = .197, r = 0.10 > summaryh(lm(newsworthiness ~ b0c * b1c, d3_b1_dwell[topic=="willsmith"])) term results 1: (Intercept) b = 3.03, SE = 0.12, t(172) = 26.19, p < .001, r = 0.89 2: b0c b = 0.06, SE = 0.11, t(172) = 0.51, p = .612, r = 0.04 3: b1c b = −2.87, SE = 6.00, t(172) = −0.48, p = .633, r = −0.04 4: b0c:b1c b = 6.92, SE = 3.80, t(172) = 1.82, p = .071, r = 0.14 # almost # study 2 climate change > summaryh(lm(importance ~ b0c * b1c, d3_b1_engage[topic=="climate"])) term results 1: (Intercept) b = 5.10, SE = 0.08, t(240) = 62.41, p < .001, r = 0.97 2: b0c b = 0.76, SE = 0.15, t(240) = 5.11, p < .001, r = 0.31 # 3: b1c b = −61.91, SE = 30.20, t(240) = −2.05, p = .042, r = −0.13 # 4: b0c:b1c b = 63.84, SE = 31.95, t(240) = 2.00, p = .047, r = 0.13 ### > summaryh(lm(newsworthiness ~ b0c * b1c, d3_b1_engage[topic=="climate"])) term results 1: (Intercept) b = 4.85, SE = 0.09, t(240) = 56.58, p < .001, r = 0.96 2: b0c b = 0.96, SE = 0.16, t(240) = 6.19, p < .001, r = 0.37 # 3: b1c b = −49.29, SE = 31.67, t(240) = −1.56, p = .121, r = −0.10 4: b0c:b1c b = 90.25, SE = 33.50, t(240) = 2.69, p = .007, r = 0.17 # interaction!?! # study 3 monkeypox > summaryh(lm(importance ~ b0c * b1c, d3_b1_engage[topic=="monkeypox"])) term results 1: (Intercept) b = 4.08, SE = 0.07, t(439) = 59.54, p < .001, r = 0.94 2: b0c b = 0.60, SE = 0.16, t(439) = 3.69, p < .001, r = 0.17 # 3: b1c b = −36.08, SE = 31.70, t(439) = −1.14, p = .256, r = −0.05 4: b0c:b1c b = 41.11, SE = 41.29, t(439) = 1.00, p = .320, r = 0.05 > summaryh(lm(newsworthiness ~ b0c * b1c, d3_b1_engage[topic=="monkeypox"])) term results 1: (Intercept) b = 4.19, SE = 0.07, t(439) = 60.46, p < .001, r = 0.94 2: b0c b = 0.45, SE = 0.17, t(439) = 2.70, p = .007, r = 0.13 # 3: b1c b = −76.68, SE = 32.05, t(439) = −2.39, p = .017, r = −0.11 # 4: b0c:b1c b = 77.06, SE = 41.75, t(439) = 1.85, p = .066, r = 0.09 # almost ``` monkeypox study **importance** ![[1658872430.png]] monkeypox study - **newsworthiness** ![[1658872529.png]] # differences in b0 and b1 magnitudes across studies ## engagement ### b1 Engagement on focal headlines **decreased more** across trials for studies 2 and 3. ```r > summaryh(lm(b1 ~ study, s0)) term results 1: (Intercept) b = −0.002, SE = 2e−04, t(895) = −8.01, p < .001, r = −0.26 2: study2 b = −5e−04, SE = 3e−04, t(895) = −1.77, p = .077, r = −0.06 # 3: study3 b = −5e−04, SE = 3e−04, t(895) = −1.80, p = .072, r = −0.06 # # combine studies 2 & 3 > summaryh(lm(b1 ~ study23, s0)) term results 1: (Intercept) b = −0.002, SE = 2e−04, t(896) = −8.01, p < .001, r = −0.26 2: study2323 b = −5e−04, SE = 3e−04, t(896) = −1.98, p = .048, r = −0.07 # ``` ### b0 Initial engagement was much higher in climate change study than other studies. ```r > summaryh(lm(b0 ~ study, s0)) term results 1: (Intercept) b = 0.37, SE = 0.03, t(895) = 11.41, p < .001, r = 0.36 2: study2 b = 0.20, SE = 0.04, t(895) = 4.37, p < .001, r = 0.14 # 3: study3 b = 0.007, SE = 0.04, t(895) = 0.18, p = .859, r = 0.01 > summaryh(lm(b0 ~ study, s0)) term results 1: (Intercept) b = 0.57, SE = 0.03, t(895) = 18.69, p < .001, r = 0.53 2: study1 b = −0.20, SE = 0.04, t(895) = −4.37, p < .001, r = −0.14 # 3: study3 b = −0.19, SE = 0.04, t(895) = −4.97, p < .001, r = −0.16 # ``` ## dwell ### b1 Dwell on focal headlines **decreased more** across trials in study 2. ```r > summaryh(lm(b1 ~ study, s0)) term results 1: (Intercept) b = −0.02, SE = 0.001, t(785) = −13.75, p < .001, r = −0.44 2: study2 b = 0.003, SE = 0.002, t(785) = 1.96, p = .050, r = 0.07 # 3: study3 b = 0.002, SE = 0.001, t(785) = 1.48, p = .140, r = 0.05 ``` ### b0 Monkeypox study has the highest initial dwell. ```r > summaryh(lm(b0 ~ study, s0)) term results 1: (Intercept) b = 8.27, SE = 0.09, t(785) = 96.23, p < .001, r = 0.96 2: study2 b = −0.09, SE = 0.12, t(785) = −0.79, p = .433, r = −0.03 3: study3 b = −0.41, SE = 0.10, t(785) = −3.98, p < .001, r = −0.14 # > summaryh(lm(b0 ~ study, s0)) term results 1: (Intercept) b = 8.18, SE = 0.08, t(785) = 102.45, p < .001, r = 0.96 2: study1 b = 0.09, SE = 0.12, t(785) = 0.79, p = .433, r = 0.03 3: study3 b = −0.32, SE = 0.10, t(785) = −3.24, p = .001, r = −0.12 # > summaryh(lm(b0 ~ study, s0)) term results 1: (Intercept) b = 7.86, SE = 0.06, t(785) = 139.27, p < .001, r = 0.98 2: study2 b = 0.32, SE = 0.10, t(785) = 3.24, p = .001, r = 0.12 # 3: study1 b = 0.41, SE = 0.10, t(785) = 3.98, p < .001, r = 0.14 # ```