How to Reverse code negative items/questions of Likert scale in SPSS

TL;DR

Reverse-coded items are negatively worded questionnaire items that must be flipped so higher scores align with the construct direction.

Briefing Cornell Notes

Briefing

Reverse-coded Likert items need to be flipped so that higher scores consistently reflect the same direction of a construct. In practice, negatively worded questionnaire items—often inserted to check response consistency—must be transformed before computing scales or running analyses like correlations. For a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree), the standard reversal rule is to replace each response x with (8 − x). So a response of 6 becomes 2, 5 becomes 3, 1 becomes 7, and the midpoint 4 stays 4.

The tutorial uses affective commitment as an example, where three items are negatively worded (e.g., “I do not feel like a part of family at my organization”). Those items are identified in the dataset (named ac3, ac4, and ac6). Because manually recoding hundreds of cases is impractical, the workflow in SPSS relies on built-in transformation tools to reverse the items in bulk.

One approach is to create a new reverse-coded variable for each negatively worded item using Transform → Compute Variable. The new variable names are created by appending an “r” (for example, ac3 becomes ac3r, ac4 becomes ac4r, and ac6 becomes ac6r). The compute step applies the reversal formula for a 7-point scale: 8 minus the original item value. The tutorial also notes that SPSS can handle different Likert ranges by adjusting the constant (e.g., using 6 for a 5-point scale), but the example stays with the 7-point version.

A second SPSS option is Transform → Recode into Different Variables, which is especially convenient when many items are reverse-coded. This method preserves the original variables and creates new ones, reducing the risk of losing the raw data. In the recode dialog, the user selects all negatively worded items at once (using multi-select with Ctrl), renames the outputs with an “r” suffix, and defines the value mapping explicitly: 1→7, 2→6, 3→5, 4→4, 5→3, 6→2, 7→1.

After reverse coding, scale construction and downstream analyses must use the new variables (the “r” versions), not the original negatively worded items. For example, when computing an affective commitment score, the tutorial demonstrates using Transform → Compute Variable to average the relevant items—ac3r, ac4r, and ac6r—so the resulting composite score reflects the intended direction of the construct. The same principle applies to reliability checks and correlation modeling: all calculations should rely on the consistently oriented, reverse-coded items.

Cornell Notes

Reverse-coded Likert items must be flipped so that all items point in the same direction for a construct. For a 7-point scale, the reversal rule is to replace each response x with (8 − x), meaning 1↔7, 2↔6, 3↔5, and 4 stays 4. In SPSS, the recommended workflow creates new variables (e.g., ac3r, ac4r, ac6r) rather than overwriting the originals. Using Transform → Recode into Different Variables is efficient when many items are negatively worded, while Transform → Compute Variable works well for formula-based recoding. After recoding, any scale score (like an affective commitment average) and analyses such as correlations should use the new reverse-coded variables.

What makes an item “reverse coded,” and why does it matter for analysis?

An item is reverse coded when its wording is negative relative to the construct direction. In the affective commitment example, negatively worded statements (ac3, ac4, ac6) require flipping so that higher values consistently represent stronger affective commitment. If the flip isn’t done, composite scores and correlations can be distorted because some items would move in the opposite direction.

What is the exact reversal rule for a 7-point Likert scale in SPSS?

For a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree), each response x is replaced with (8 − x). Examples from the tutorial: 6 becomes 2, 5 becomes 3, 1 becomes 7, and the midpoint 4 remains 4.

How can SPSS reverse-code items without manually changing hundreds of responses?

Use transformation tools. One method is Transform → Compute Variable to create new variables (e.g., ac3r) and apply the formula 8 − ac3. Another method is Transform → Recode into Different Variables, which lets users map values directly (1→7, 2→6, 3→5, 4→4, 5→3, 6→2, 7→1) for selected negatively worded items.

Why does the tutorial prefer “recode into different variables” over overwriting?

It preserves the original dataset. By creating new variables with an “r” suffix (ac3r, ac4r, ac6r), the original negatively worded items remain available for future checks or reprocessing, reducing the risk of losing raw responses.

After reverse coding, which variables should be used to compute a construct score?

Use only the reverse-coded versions (the “r” variables). For affective commitment, the composite should be computed from ac3r, ac4r, and ac6r (not ac3, ac4, ac6). This ensures the scale reflects the intended direction across all items.

Review Questions

In a 7-point Likert scale, what value does SPSS produce when reversing an original response of 6? What about 4?
What are the two SPSS transformation paths mentioned for reverse coding, and how do they differ in workflow?
When building a composite score for affective commitment, which variables should be included: the original negatively worded items or the reverse-coded “r” variables?

Key Points

1
Reverse-coded items are negatively worded questionnaire items that must be flipped so higher scores align with the construct direction.
2
For a 7-point Likert scale, reverse coding uses the rule x → (8 − x), mapping 1↔7, 2↔6, 3↔5, and keeping 4 unchanged.
3
SPSS can reverse-code multiple items efficiently using Transform → Compute Variable or Transform → Recode into Different Variables.
4
Creating new variables with an “r” suffix (e.g., ac3r) preserves the original data and reduces the chance of accidental overwriting.
5
When computing scale scores (like affective commitment), use the reverse-coded variables (ac3r, ac4r, ac6r), not the original items.
6
Downstream analyses such as correlations and reliability checks should rely on consistently oriented items to avoid sign errors.

Highlights

Reverse coding on a 7-point Likert scale follows a simple constant: replace each response x with (8 − x).

The tutorial’s example flips negatively worded affective commitment items (ac3, ac4, ac6) into new variables (ac3r, ac4r, ac6r).

SPSS value mapping for reversal is explicit: 1→7, 2→6, 3→5, 4→4, 5→3, 6→2, 7→1.

Composite scores must be computed using the reverse-coded variables; otherwise, the construct direction can be wrong.

Topics

Reverse Coding
Likert Scale
SPSS Transformations
Affective Commitment
Data Recoding

Mentioned

Cameron