Hi there,
I have ran a logistic regression on a number of Socio-economic predictor variables. The DV is joining a political party due to its internet use.
I have coded gender (0= female, 1= male).
My coefficient is sizable, negatively signed and significant. However, the odd ratio is below 1 (.704).
How can I interpret this? Does the negatively signed coefficient mean that I should interpret this differently? Or do I still need the OR to be above 1?
Any help would be greatly appreciated.
Neil
If you prefer an OR more than one, reverse the coding of gender.
Interpreting the coefficient for DV is depending of variable levels. (if yes=reference level): odds ratio compares chances of the level "no" for DV to record a success, relative to the reference level (yes). If the odd ratio for the EV=gender (F=0-reff, M=1) is below 1 (.704), that means that M has lower chance than F to join a political party due to its internet use.
It is easy to interpret the negative odds of a categorical variable. For example the odds interpretation for your example is as follows. Males using the Internet are 29.6% (1-0.704) less likely to join a political party than females (Reference = female)
- Badges
- Science topic
- Similar topics
- Medicine
- Disease
- Infectious Diseases
- Immunization
- Vaccination
- Vaccines