n = 145
# lean
- condition: dummy (proT 0, proN 1)
- strategy
- positive (reference group; highlight ingroup positives)
- anti (highlight outgroup negatives)
- both (contrast positive + anti)
```r
r
gt; summ(feols(lean2 ~ conditionN * lean1Z * strategyF, data = d0, se = "HC1"))
term result sig
<char> <char> <char>
1: (Intercept) b = 43.31 [37.45, 49.17], p < .001 ***
2: conditionN b = -0.58 [-10.82, 9.66], p = .912
3: lean1Z b = 44.43 [40.98, 47.89], p < .001 ***
4: strategyFanti b = 1.71 [-5.08, 8.51], p = .619
5: strategyFboth b = -7.19 [-16.80, 2.41], p = .141
6: conditionN × lean1Z b = -7.78 [-17.10, 1.55], p = .101
7: conditionN × strategyFanti b = -1.17 [-12.64, 10.29], p = .840
8: conditionN × strategyFboth b = 14.90 [-0.28, 30.08], p = .054 . # treatment effect much bigger when using both positive/negative
9: lean1Z × strategyFanti b = -2.68 [-6.76, 1.39], p = .195
10: lean1Z × strategyFboth b = -7.89 [-18.27, 2.49], p = .135
11: conditionN × lean1Z × strategyFanti b = 6.40 [-3.62, 16.42], p = .208
12: conditionN × lean1Z × strategyFboth b = 9.05 [-6.62, 24.71], p = .255
# anova
Contrasts set to contr.sum for the following variables: condition, strategy
Anova Table (Type 3 tests)
Response: lean2
Effect df MSE F ges p.value
1 condition 1, 133 299.54 1.88 .014 .173
2 lean1Z 1, 133 299.54 729.37 *** .846 <.001
3 strategy 2, 133 299.54 0.05 <.001 .947
4 condition:lean1Z 1, 133 299.54 0.80 .006 .372
5 condition:strategy 2, 133 299.54 3.25 * .047 .042 # interaction (same as above)
6 lean1Z:strategy 2, 133 299.54 0.71 .011 .492
7 condition:lean1Z:strategy 2, 133 299.54 0.83 .012 .438
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘+’ 0.1 ‘ ’ 1
```
# strategy manipulation
![[strategy_use 4.png]]