Theoretical examples

Example 1 Suppose we want to examine the association between gender (x) and alcohol abuse (y). Gender has the values 0=Man and 1=Woman, whereas alcohol abuse has the values 0=No and 1=Yes. Now, we get an OR of 0.66. This would mean that women are less likely compared to men to abuse alcohol.

Example 2 Here, we want to examine the association between having young children (x) and owning a pet (y). Having young children is measured as either 0=No young children and 1=Young children. Owning a pet has the values 0=No and 1=Yes. Let us say that we get an OR that is 1.49. We can hereby conclude that it is more common to own a pet in families with young children compared to families without young children.

Practical example

Dataset

StataData1.dta

Variable name	earlyret
Variable label	Early retirement (Age 50, Year 2020)
Value labels	0=No 1=Yes

Variable name	sex
Variable label	Sex
Value labels	0=Man 1=Woman

sum earlyret sex if pop_logistic==1

The variable sex is binary: 0=Man, 1=Woman. When we add it to the model, the category with the lowest value will be the reference category (i.e. Man). 

logistic earlyret sex if pop_logistic==1

When we look at the results for sex, we see that the odds ratio (OR) is 1.70. Now, it is important to remember the coding of sex: 0=Man, 1=Woman. Thus, a unit increase in sex is the same as being a woman compared to a man. Since men are the reference category, they automatically get the OR 1.00. The specific odds ratio of 1.70 can be interpreted as women having higher odds of earlyret compared to men. 

There is a statistically significant association between sex and earlyret, as reflected in the p-value (0.000) and the 95% confidence intervals (1.48-1.96). 

Summary Women are more likely to have experienced early retirement at the age of 50, as compared to men (OR=1.70, 95% CI=1.48 to 1.96).