Section author: Jonas Rafi

Logistic Regression

How to perform a logistic regression in jamovi:
  1. You need one nominal ordinal categorical dependent variable (nominal or ordinal), and at least one continuous explanatory variable.
    A correct setup should look similar to this:

    data_format_regression_logistic


  2. Logistic regression can be found by selecting Analyses → Regression. If the outcome variable is nominal (as in the above image), select 2 Outcomes if it has 2 steps / different values, or N outcomes if it has more than 2 steps. If the outcome variable is ordinal (e.g., low, medium, high), select Ordinal Outcomes.

    select_regression_logistic


  3. Drag and drop your dependent variable to Dependent Variable and your predictor to Covariates.

    add_var_regression_logistic


  4. Scroll down to the Model Coefficients drop-down-menu and check the options Odds ratio and Confidence interval.

    options_regression_logistic

  5. The result is shown in the right panel:

    output_regression_logistic


Tip

Those who have previous experience with SPSS may want to have a look at the side-by-side-comparison of how a logistic regression is conducted in SPSS and jamovi.

Further help from the community resources

A little more comprehensive introduction into this statistical method is provided by this two videos, explaining logistic regression with two levels (to predict, e.g., gender or clinical vs. control group) and with more than two levels (to predict, e.g., food preferences: fast food, healthy food, high protein food, vegan food, etc.).