Perceptions of plain packaging and health warnings among university students in Turkey: a survey-based experiment

BMC Public Health

Table 2 Regression models explaining the three outcome variables

	(1)	(2)	(3)	(4)	(5)	(6)
Outcome Variable:	Negative affect		Avoidant responses		Intentions to quit
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
Condition 1	0	0	0	0	0	0
Condition 2	0.21*	0.21*	0.17	0.17	0.049	0.039
Condition 3	0.54***	0.53***	0.56***	0.59***	0.047	0.047
Condition 4	0.74***	0.70***	0.68***	0.62***	0.079**	0.065**
Condition 5	0.79***	0.75***	0.75***	0.70***	0.089**	0.065*
Age		0.033*		-0.015		-0.0072
Gender: Male		0		0		0
Female		0.32***		0.50***		-0.027
Income (in ‘000)		-0.042		0.032		-0.032
Age at initiation		0.041**		0.033**		-0.0012
Quit attempt		0.069		0.083		0.019
Nicotine dependence: 0: > 60 min		0		0		0
1: in 31–60 min		-0.082		-0.17		-0.044
2: in 6–30 min		-0.38***		-0.18		-0.068**
3: within 5 min		-0.38**		-0.057		-0.080**
Past 30-day consumption: None		0		0		0
10/day or less		-0.38***		-0.39***		-0.15***
More than 10/day		-0.32**		-0.56***		-0.11***
Smoking is addictive
Disagree/Definitely disagree		-0.081		-0.19		0.017
Neither agree, nor disagree		0		0		0
Agree/Definitely agree		0.11		-0.013		0.015
Cigarettes absolutely harmful		0.24**		0.068		0.057**
Smoking but not harming self:
Disagree/Definitely disagree		-0.032		0.048		0.019
Neither agree, nor disagree		0		0		0
Agree/Definitely agree		-0.098		0.073		0.030
Constant	1.71***	0.33	1.79***	1.66***	0.043**	0.29**
N	623	623	623	623	623	623

Columns in the table show estimates from different regressions. Coefficient estimates for reference categories are zero. Model 1 included only the dummy variables for conditions. Model 2 added other control variables. Ordinary least squares was used to estimate regressions in columns (1)-(4). Linear probability model was used to estimate regressions in columns (5)-(6). Condition 1: Brand logos + text, Condition 2: Brand logos + old text + old pictures, Condition 3: Brand logos + new text + new pictures, Condition 4: PP + old text + old pictures, Condition 5: PP + new text + new pictures. Negative affect was the average of scores in eight questions (1 to 5). Avoidant responses was the average of scores in five questions (1 to 5). Intentions to quit was a dummy variable that was 1 if the respondent intended to quit within six months or sooner, and 0 otherwise. The categories of ‘Past 30-day consumption’, ‘Smoking is addictive’, and ‘Smoking but not harming self’ were condensed to present the results more concisely. Since a large majority of the responses to harm perception question was ‘absolutely harmful’, a new variable ‘Cigarettes absolutely harmful’ was created as a dummy variable equal to 1 if ‘absolutely harmful’ was chosen, 0 if one of the other three responses was chosen. *** p < 0.01, ** p < 0.05, * p < 0.10

ISSN: 1471-2458