Reconsidering the impact of behavior prediction questions on illegal drug use: The importance of using proper analytic methods

Jon Krosnick is University Fellow at Resources for the Future. We thank Patti Williams, Lauren Block, and Gavan Fitzsimons for providing us with their data sets for the analyses reported here.

Figures & data

Table 1. Distribution of illegal drug use among participants in Williams et al. (2006a)

Uses of illegal drugs	All participants		Treatment		Control
	Frequency	Percent	Frequency	Percent	Frequency	Percent
0	122	73.05	62	72.94	60	73.17
1	13	7.78	7	8.24	6	7.32
2	7	4.19	3	3.53	4	4.88
3	3	1.80			3	3.66
4	7	4.19	4	4.71	3	3.66
5	2	1.20	1	1.18	1	1.22
6	1	0.60			1	1.22
7	1	0.60			1	1.22
10	3	1.80	1	1.18	2	2.44
12	2	1.20	2	2.35
15	1	0.60			1	1.22
20	1	0.60	1	1.18
25	2	1.20	2	2.35
50	2	1.20	2	2.35
N	167	100	85	100	82	100

Table 2. Means, SD, and tests of the significance of the effect of behavior prediction questions on drug use

	All participants	Excluding one outlier	Excluding two outliers
Mean of the control group (SD)	1.07 (2.58)	1.07 (2.58)	1.07 (2.58)
Mean of the treatment group (SD)	2.80 (8.78)	2.24 (7.13)	1.66 (4.83)
p‐value: Ordinary least squares (t‐test)	.084	.162	.329
p‐value: OLS on log‐transformed DV	.409	.590	.828
p‐value: OLS on square‐root‐transformed DV	.268	.438	.704
p‐value: OLS on fourth‐root‐transformed DV	.582	.762	.972
p‐value: Zhou Z‐Score test	.028	.110	.369
p‐value: Negative binomial regression	.026	.092	.290
p‐value: Mann‐Whitney‐Wilcoxon	.849	.982	.881
p‐value: Kolmogorov‐Smirnov	.979	.998	1.000
p‐value: Permutation test	.087	.185	.342
p‐value: Bootstrap test	.082	.154	.319
p‐value: Lachenbruch 2‐part model with a t‐test	.140	.236	.374
p‐value: Lachenbruch 2‐part model with a Mann‐Whitney‐Wilcoxon test	.673	.802	.897
p‐value: Lachenbruch 2‐part model with a t‐test and log transformation	.448	.620	.799
N	167	166	165

Williams , P. , Block , L. G. and Fitzsimons , G. J. 2006a . Simply asking questions about health behaviors increases both healthy and unhealthy behaviors. . Social Influence , 1 (2) : 117 – 127 .

 Google Scholar