How do I calculate the adjusted odds ratio using jamovi (Java Statistical Analysis Module)?

How to Calculate Adjusted Odds Ratio Using Jamovi

To calculate adjusted odds ratios in Jamovi, use the logistic regression module under the "Analyses" tab, add your outcome variable, predictor variables, and covariates, then interpret the exponentiated coefficients as odds ratios.

Understanding Odds Ratios

An odds ratio (OR) represents the odds that an outcome will occur given a particular exposure, compared to the odds of the outcome occurring in the absence of that exposure. When we adjust for covariates, we're controlling for potential confounding variables that might influence this relationship.

Step-by-Step Guide to Calculate Adjusted Odds Ratios in Jamovi

1. Open your dataset in Jamovi

   • Import your data file (.csv, .sav, etc.)
   • Ensure variables are correctly defined (nominal, ordinal, continuous)

2. Access the Logistic Regression module

   • Click on "Analyses" tab
   • Select "Regression"
   • Choose "Logistic Regression"

3. Set up your model

   • Add your dependent variable (outcome) to the "Dependent Variable" box
   • Add your independent variable(s) of interest to the "Covariates" box
   • Add your adjustment variables (confounders) to the "Covariates" box

4. Configure additional options

   • Under "Model Builder," ensure all variables are included in the model
   • In "Model Coefficients," check "Odds ratio" to display ORs instead of log odds
   • Select "Confidence interval" and set your desired level (typically 95%)

5. Interpret the results

   • The "Odds ratio" column shows your adjusted odds ratios
   • The confidence intervals show the precision of these estimates
   • The p-values indicate statistical significance

Advanced Options

Interaction Terms

To include interaction effects:

• Go to "Model Builder"
• Select the variables you want to interact
• Click "Add Interaction" button

Categorical Variables

For categorical predictors with multiple levels:

• Jamovi automatically creates dummy variables
• The reference category can be changed in the variable settings

Common Pitfalls to Avoid

1. Multicollinearity: Check for high correlations between predictors, as this can inflate standard errors and make odds ratios unstable.

2. Complete separation: If a predictor perfectly predicts the outcome, the model won't converge properly.

3. Sample size: Ensure you have sufficient cases per predictor variable (typically at least 10 events per predictor).

4. Misinterpreting odds ratios: Remember that odds ratios are not the same as relative risks, especially when outcomes are common (>10% prevalence) 1.

5. Confusing adjusted and unadjusted odds ratios: The adjusted OR from logistic regression represents the effect of each variable while controlling for all other variables in the model 2.

Statistical Background

The logistic regression model uses the log odds link function to relate predictors to the binary outcome 2:

log(odds) = β₀ + β₁X₁ + β₂X₂ + ... + βₚXₚ

Where:

• log(odds) is the natural logarithm of the odds of the outcome
• β₀ is the intercept
• β₁, β₂, ..., βₚ are the regression coefficients
• X₁, X₂, ..., Xₚ are the predictor variables

The exponentiated coefficients (e^β) represent the odds ratios, which Jamovi can display directly in the output 2.

By using this approach, you'll obtain more precise estimates of odds ratios compared to dichotomizing count data before analysis 3.