What Is Reverse Scoring — and Why It Matters
If you have ever added up the items of a questionnaire in SPSS and got a Cronbach’s alpha near zero — or even negative — there is a good chance one or more items needed reverse scoring. Reverse scoring is the step of flipping the values of negatively worded items so that, across the whole scale, a high score always means the same thing.
Take a life-satisfaction scale. One item reads “I am satisfied with my life” and another reads “I often feel my life is meaningless.” A very satisfied respondent would score 5 on the first item and 1 on the second. Added raw, the two answers cancel each other out. Reverse scoring fixes this: the 1 becomes a 5, so both answers point in the same direction.
Scale designers include negatively worded items on purpose: they guard against acquiescence bias (respondents agreeing with everything) and keep people reading carefully. But that design only works if you reverse the scores before computing totals, running reliability analysis, or testing hypotheses. Forgetting this step is one of the most common beginner mistakes in SPSS, and it can quietly wreck an entire analysis.
How to Identify Items That Need Reverse Scoring
Before you touch SPSS, you need a list of the items to reverse. There are three reliable ways to build it:
1. Check the scale manual or codebook
Every validated scale publishes a scoring key. It states plainly which items are reverse-worded, for example “Items 3, 7, and 12 are reverse-scored.” If you are using a published scale, this is your authority — follow it exactly. If you designed the questionnaire yourself, your scoring key is the list of items you worded negatively.
2. Read the wording
An item needs reversing when a high score on it means the opposite of a high score on the rest of the scale. In a self-esteem scale, “I feel I have a number of good qualities” is scored normally, but “At times I think I am no good at all” is reverse-worded. Watch for words like “not,” “never,” “rarely,” and “hardly” — the usual giveaways, though not every negative word signals a reversed item.
3. Let the data warn you
If you have no scoring key, run a reliability analysis first (Analyze > Scale > Reliability Analysis) and look at the Corrected Item-Total Correlation column. An item with a strongly negative item-total correlation is almost certainly a reverse-worded item that has not been reversed yet. After you reverse it, that correlation should turn positive. If it stays negative, the item may simply be a bad item — consider dropping it rather than forcing it.
The Reverse-Scoring Formula
The formula is simple: a reversed score equals the sum of the scale’s minimum and maximum values, minus the original score.
Reversed score = (scale minimum + scale maximum) − original score
For a 1–5 Likert scale, that is 6 − original score, so 1 becomes 5, 2 becomes 4, 3 stays 3, 4 becomes 2, and 5 becomes 1. For a 1–7 scale it is 8 − original score, and for a 0–4 scale it is 4 − original score. The middle value of an odd-numbered scale never changes — on a 1–5 scale a 3 stays a 3 — which is a useful check that you applied the formula correctly.
Method 1: Reverse-Score With the SPSS Menus (Recode into Different Variables)
The safest menu route is Transform > Recode into Different Variables: it creates new variables and leaves your original data untouched. Never recode into the same variable unless you have a backup copy of the file.
Step-by-step
- Click Transform > Recode into Different Variables.
- In the dialog, select the item you want to reverse (for example, Q3) and move it into the Numeric Variable → Output Variable box.
- Under Output Variable, type a name for the reversed variable, such as Q3_R, and click Change.
- Click Old and New Values. For a 1–5 scale, enter the pairs one at a time: Old Value 1 → New Value 5, click Add; 2 → 4, Add; 3 → 3, Add; 4 → 2, Add; 5 → 1, Add.
- Click Continue, then OK.
- SPSS creates Q3_R at the end of your data file. Check it in Data View: a respondent who scored 1 on Q3 should now score 5 on Q3_R.
To reverse several items, add all of them to the Numeric Variable list before clicking Old and New Values — SPSS applies the same old-to-new mapping to every variable in the list.
Method 2: Reverse-Score With Compute Variable
If you prefer a formula to a value-by-value mapping, Transform > Compute Variable is faster, especially for a single item.
- Click Transform > Compute Variable.
- In Target Variable, type the new name, for example Q3_R.
- In the Numeric Expression box, type the formula for your scale. For a 1–5 scale: 6 – Q3. For a 1–7 scale: 8 – Q3.
- Click OK.
Method 3: Paste-Ready SPSS Syntax
Syntax is the best option when you need to reverse many items at once, or when you want a reproducible record of exactly what you did. Paste the block below into a syntax window (File > New > Syntax), adapt the variable names, and run it.
* Reverse-score items Q3, Q7 and Q12 on a 1-5 Likert scale. * Recode into NEW variables so the originals are preserved. RECODE Q3 Q7 Q12 (1=5) (2=4) (3=3) (4=2) (5=1) INTO Q3_R Q7_R Q12_R. VARIABLE LABELS Q3_R 'Q3 reverse-scored' /Q7_R 'Q7 reverse-scored' /Q12_R 'Q12 reverse-scored'. EXECUTE. * Alternative: the same reversal with COMPUTE (1-5 scale: new = 6 - old). * Use this version if you prefer a formula to value mapping. COMPUTE Q3_R = 6 - Q3. COMPUTE Q7_R = 6 - Q7. COMPUTE Q12_R = 6 - Q12. EXECUTE. * Reverse several items at once with DO REPEAT (1-5 scale). DO REPEAT old = Q3 Q7 Q12 / new = Q3_R Q7_R Q12_R. COMPUTE new = 6 - old. END REPEAT. EXECUTE. * Now compute the scale total from the corrected items. * Suppose the scale has 12 items and Q3, Q7, Q12 were the reversed ones. COMPUTE LifeSat_Total = SUM(Q1, Q2, Q3_R, Q4, Q5, Q6, Q7_R, Q8, Q9, Q10, Q11, Q12_R). EXECUTE.
A note on the last block: SUM in SPSS ignores missing values, so a respondent who skipped one item still gets a total based on the items they did answer.
How to Check Your Reverse Scoring Worked
1. Eyeball a few cases
Open Data View and compare the original and reversed columns for five or six respondents. A 1 should have become a 5, a 2 should have become a 4, and a 3 should still be a 3. If any row looks wrong, you applied the mapping backwards — the single most common error.
2. Run Frequencies on both versions
Go to Analyze > Descriptive Statistics > Frequencies and run it on Q3 and Q3_R. The frequency counts should be mirror images: the number of 1s on Q3 should equal the number of 5s on Q3_R. The mean of Q3_R should equal 6 minus the mean of Q3 (for a 1–5 scale). If the means are identical, nothing was reversed.
3. Re-run reliability analysis
Run Analyze > Scale > Reliability Analysis on the full set of items, using the reversed versions where they apply. Cronbach’s alpha should rise, often dramatically, and every item-total correlation should be positive. If alpha barely moves, double-check that you used the reversed variables in the analysis and not the originals.
4. Watch for these common mistakes
- Reversing twice: running the recode a second time flips the item back. If you re-run syntax, delete the old reversed variables first or use fresh names.
- Reversing the wrong items: reversing a normally worded item damages the scale exactly as much as forgetting to reverse a reversed one.
- Mixing scales: applying the 6 − x formula to a 1–7 item, or to a 0–4 item, produces out-of-range values. Always match the formula to the item’s own response scale.
- Analysing the originals: after reversing, every later step — totals, reliability, correlations, group comparisons — must use the _R variables.
Reporting Reverse Scoring in APA Style (7th ed.)
Describe your scoring procedure in one or two sentences in your Method section — enough for a reader to reproduce it. Name the items and the rule you applied, then report the scale’s reliability after scoring. For example:
“The scale consists of five items rated on a 7-point scale (1 = strongly disagree, 7 = strongly agree). One item was reverse-scored so that higher scores indicated greater life satisfaction. Total scores were computed as the mean of the five items (Cronbach’s α = .XX — insert your own value here).”
Points to cover in your own write-up:
- Which items were reverse-scored, by number or wording.
- The response scale and the reversal rule (for example, “scores were reversed by subtracting each response from 8”).
- That higher scores indicate more of the construct after reversal, and the Cronbach’s alpha for the scale after reverse scoring, not before.
- How total or mean scores were computed, including how you handled missing items.
In a results section, you never need to mention reverse scoring again — by that point the reader assumes all scores point in the same direction. But if a reviewer asks, your syntax file is the proof that the step was done correctly.
Key Takeaways
- Reverse scoring flips negatively worded items so that a high score means the same thing on every item of a scale.
- Identify the items to reverse from the scale’s scoring key first; a negative item-total correlation is your data-driven backup check.
- The formula is (scale minimum + scale maximum) − original score — 6 − x for a 1–5 scale, 8 − x for a 1–7 scale.
- Always recode into a new variable (for example, Q3_R) and keep the originals untouched.
- Use syntax (RECODE, COMPUTE, or DO REPEAT) when reversing several items — it is faster and leaves an auditable record.
- Verify with Frequencies and a re-run of reliability analysis: mirrored frequency counts, positive item-total correlations, and a higher Cronbach’s alpha mean the reversal worked.
Frequently Asked Questions
What does reverse scoring mean in SPSS?
Reverse scoring means flipping the values of a negatively worded questionnaire item so it points in the same direction as the rest of the scale. On a 1–5 scale, a 1 becomes a 5 and a 2 becomes a 4. In SPSS use Transform > Recode into Different Variables, Transform > Compute Variable, or a short syntax command.
What is the formula for reverse scoring a 1–5 Likert scale?
Subtract the original score from 6: reversed score = 6 − original score. The general formula is (scale minimum + scale maximum) − original score.
Should I recode into the same variable or a different variable?
Always recode into a different variable. Create a new variable such as Q3_R and leave the original Q3 untouched — recoding into the same variable overwrites the raw responses, so a mistake cannot be undone without re-importing the data.
How do I know which items to reverse score?
Follow the scoring key in the scale’s manual — it lists the reverse-worded items by number. If you have no key, read the wording (items where agreement means the opposite of the construct) and confirm with a reliability analysis: an item with a negative item-total correlation is a strong candidate for reversal.
What happens if I forget to reverse score?
Your scale total becomes meaningless because opposite-worded items cancel each other out, Cronbach’s alpha drops sharply (sometimes below zero), and every correlation or group comparison based on the total will be wrong. If your alpha is unexpectedly low, forgotten reverse scoring is the first thing to check.
Need this done on your data? Get a free quote → /services/