How do I... TRANSFORM VARIABLES in Jamovi? (2022)

Summary

This tutorial demonstrates how to reverse-code a Likert-scale variable in Jamovi 2.2.5 by creating a transformed copy while preserving the original data.

The example uses a five-point agreement scale to measure preferences for dogs and cats. Because “Cats are better” is worded in the opposite direction from “Dogs are better,” its scores must be reversed before combining the items into an overall preference measure. The demonstration treats the two items as opposite ends of the same preference scale.

Reverse-Coding Procedure in Jamovi

1. Open the Data tab and choose Transform for the variable you want to change.
2. Create a new transformation. Jamovi creates a transformed variable based on the original.
3. Give the transformation a descriptive name, such as “cats reverse coded.” This is the transformation’s name, which may differ from the resulting variable’s name.
4. Optionally, add a description or a suffix to append to the transformed variable’s name.
5. Add recoding conditions using an if/else-style format. Jamovi uses R-like logic for these conditions and supplies the original source variable automatically.
6. For a 1–5 scale, enter these reverse-coding mappings:
   • 1 → 5
   • 2 → 4
   • 3 → 3
   • 4 → 2
   • 5 → 1
7. Set a fallback value for responses that do not match any condition, such as a missing value or, as in the demonstration, zero. Choose a fallback that is appropriate for the dataset.
8. Check the resulting column to confirm that the values were recoded as intended.

Transformations can be reused: once configured, the same transformation can be applied to multiple variables with the same response scale. Jamovi indicates how many variables use a transformation and marks transformed variables in the interface. The transformed variable can also be renamed, or the transformation can be duplicated or deleted.

The presenter briefly notes that reverse-coded items can later be used to calculate a combined score, such as an average. Calculating that score is left for a subsequent tutorial.

Speakers and Sources Featured

• Presenter/instructor: Alexander Swan, Ph.D. (identified by the video’s channel metadata; the subtitles do not state his name).
• Software/interface demonstrated: Jamovi.
• Technical reference mentioned: R-style logical if/else expressions used in Jamovi’s transformation conditions.