# Is a logistic regression relevant to my problem?

Hi so for my proposal I plan on conducting a logistic regression but I'm not sure if it is the right method. My independent variable is a composite variable made up of 8 items, 4 reverse coded but all summed together to create a new variable representing need fulfillment. These items are 5 point Likert scale (strongly disagree... neutral... Strongly agree) half the items represent satisfaction and the other frustration hence why I plan to reverse code the latter.

My outcome variable is a dichotomized item representing substance misuse or no misuse. I would like to see if lower scores or frustration are related to misuse and vice versa. Since the independent variable is Likert but a composite I'm not sure if a logistic regression would still be appropriate if not what other analysis could I use?

edit: I didn't mean sum I meant average of the 8 items will be used as the new variable sorry for the confusion.

The appropriateness of logistic regression does not depend on how the independent variable is coded. As long as the outcome is binary, it can be used.

If your goal is simply to test the association between a binary and a continuous variable, you could also just use a t-test or Mann-Whitney U-test. If you want to build a more complicated model, logistic regression is a good choice.

Your Likert scale is computed by addition. Addition works only for metric data so you already decided to consider your Likert data to be quasi-metric. Thus you can use it for the regression .

• Sorry I meant the average not sum. Ill be reverse coding than taking the average of all 8 items to represent need fulfillment
– J M
Nov 23, 2021 at 17:25

Yes, logistic regression is a good choice for the structural model (i.e., the regression of your binary outcome variable onto your predictor variable). However, you may also consider using a measurement model such as factor analysis (FA) or item response theory (IRT) instead of using summed scores. Nowadays, there are plenty of software tools for both types of measurement models (e.g., in R the lavaan package in for (FA) and the mirt package for IRT). The two links below are good discussions of the differences between FA and IRT.

Item Response Theory vs Confirmatory Factor Analysis

Difference between IRT and EFA to find factors