Table 4. Logistic regression models to assess the impact of highlighting mechanism on acceptability for the five interventions in the USA sample.
| Size B (SE) | Shape B (SE) | Location B (SE) | Taxation B (SE) | Education B (SE) | |
|---|---|---|---|---|---|
| (Intercept) | -0.41(0.29) | 0.51(0.30) | 0.92(0.32)*** | -0.13(0.29) | 1.55(0.38)*** |
| Conscious condition | 0.13(0.15) | -0.16(0.15) | -0.25(0.16) | -0.04(0.15) | 0.16(0.27) |
| Non-conscious condition | 0.21(0.15) | 0.27(0.16) | 0.14(0.17) | -0.02(0.15) | -0.35(0.25) |
| Gender (F) | 1.22(0.50) | 0.98(0.57) | 0.21(0.53) | -0.60(0.49) | 0.67(0.71) |
| Age [25;30[ | 0.43(0.45) | 0.25(0.48) | 0.22(0.51) | 0.27(0.45) | 1.24(0.81) |
| Age [30;40[ | -0.68(0.50) | -1.07(0.48) | -1.12(0.48) | -0.82(0.50) | 0.40(0.64) |
| Age [40;Inf[ | 0.15(0.48) | -1.03(0.49) | -0.90(0.49) | 0.61(0.49) | -0.25(0.59) |
| Income [25K,50K[ | 0.66(0.44) | 0.08(0.46) | -0.51(0.46) | -0.21(0.44) | 2.01(1.07) |
| Income [50K,Inf[ | 0.99(0.41) | 0.33(0.43) | 0.35(0.46) | 0.19(0.40) | 0.77(0.59) |
N.B.: *** = p < 0.0027358
The control condition taken as reference. To take into account the non-representativeness of the MTurk sample, we control for variables explaining the difference between the MTurk and USA populations. We chose age (4 levels, with [18;25 [y.o. as base), gender (2 levels), and income (3 levels, with [0;25K [as base) and considered all possible interactions between our control variables (i.e., 4x2x3 groups) but only report here their main effects.