Skip to main content

Table 4 Ever smoked WP: Logistic regression estimates (aORs [95% CI]), SES index used as an associated factor

From: Water pipe smoking among public versus private university students in Ankara, Turkey: an online survey

 

Private University Students

Public University Students

All Students

Private university

  

1.54***

   

[1.35,1.75]

Gender

 Female

1

1

1

 Male

2.31***

2.77***

2.53***

[1.91,2.80]

[2.31,3.32]

[2.22,2.88]

Age group

 Ages 18-19

1

1

1

 Ages 20-21

1.40***

1.78***

1.58***

[1.10,1.79]

[1.37,2.30]

[1.32,1.88]

 Ages 22-23

2.45***

2.13***

2.24***

[1.87,3.22]

[1.62,2.79]

[1.85,2.70]

 Ages 24 or older

2.81***

3.29***

3.07***

[1.88,4.22]

[2.34,4.62]

[2.37,3.97]

SES index

1.13***

1.12***

1.13***

[1.06,1.20]

[1.06,1.19]

[1.08,1.18]

Observations

2184

2352

4536

  1. Notes: The SES index was calculated using the first principal component of eight binary variables (taking values of zero or one): Has a car, Lives with family, Lives in dormitory, Lives alone, Has roommate(s), Income source: Family support, Income source: Scholarship, Income source: Work. The first principal component had positive loadings on four variables (has a car, lives with family, lives alone, and receives family support). These four variables can be thought to be associated with higher socioeconomic status. The other four variables had negative factor loadings. The mean (and standard error) values of the index for private and public university students were 0.0966 (0.0302) and -0.0893 (0.0265), respectively. A two-sample t-test for difference in means yielded a t-statistic of -4.62; therefore, the null hypothesis of equality of means was rejected
  2. The first two columns in the table show estimates of adjusted odds ratios (adjusted for all associated factors variables listed in the table) from binary logistic regressions, separately for private and public university students. The last column shows the estimates for the entire sample, adding the “Private university” dummy (binary) variable to the regression. 95% CI are shown in square brackets. ***p<0.01