In SPSS you go to Analyze > Scale > Reliability Analysis, move your scale items into the Items box, set Model to Alpha, and click OK. The Cronbach's Alpha value in the Reliability Statistics table is your answer. The one step people forget is to reverse-code any negatively worded item first, or the number comes out wrong.
Cronbach's alpha is a measure of internal consistency: it tells you how closely a set of items on a scale, such as a five-item anxiety questionnaire, behave as a single measure. This guide is part of our statistics homework help and our wider SPSS help. It gives the exact menu path, the settings to tick, the one number that answers the question, and the two extra columns that tell you whether a weak item is dragging your scale down. It also shows the step most students miss, which quietly produces a wrong or negative alpha.
The rule in one line. Alpha ranges from just below zero up to one. Higher means the items hang together more tightly. A value of 0.70 or above is generally acceptable, 0.80 is good, and 0.90 is excellent, though very high values can mean some items are redundant.
Cronbach's alpha is only meaningful when these hold. Sort them out first and the analysis is a few clicks.
| Requirement | Why it matters |
|---|---|
| Two or more items measuring one construct | Alpha describes how well a set of items measures a single underlying thing. It is not meaningful on a single item. |
| Items on the same response scale | Likert-type items, for example one to five agreement scales, are the usual case. Mixing very different scales without standardising distorts alpha. |
| Negatively worded items reverse-coded | An item worded in the opposite direction must be recoded first, or its negative correlations pull alpha down, often below zero. |
| One subscale at a time | Run alpha separately for each subscale. Pooling items from different subscales understates reliability. |
Not sure which items belong to which subscale, or whether an item is reverse-worded? Every step below shows how to check, and a statistician can confirm it on your own file.
This step comes before you touch the reliability menu, and skipping it is the single most common reason a Cronbach's alpha comes out wrong. On most scales every item points the same way, so a high score always means more of the trait. When one item is worded the opposite way, for example "I feel calm in crowds" sitting inside an anxiety scale, a high raw score on that item means less of the trait, not more. Left as is, that item correlates negatively with the rest and drags alpha down, sometimes below zero.
The fix is to reverse the scoring so the item runs in the same direction as the others. In SPSS use Transform > Recode into Different Variables. On a one-to-five scale you map 1 to 5, 2 to 4, 3 stays 3, 4 to 2 and 5 to 1, and save the result as a new variable such as item3_r. Then feed the reverse-coded versions, not the originals, into the reliability analysis. Do this for every negatively worded item before you continue.
The menu path, the settings to tick, the exact numbers to read, and how to report them. Example values follow a five-item Likert scale.
The whole procedure lives under one menu once your items are ready and any reverse items are recoded.
What each setting does. Model set to Alpha is what produces Cronbach's alpha. "Scale if item deleted" adds the two diagnostic columns you need, the Corrected Item-Total Correlation and Cronbach's Alpha if Item Deleted. Inter-Item Correlations prints the matrix so you can spot an item that correlates negatively with the rest, which is the fingerprint of a reverse item you forgot to recode.
Working with a large questionnaire, several subscales, or unsure which items to reverse? A statistician will set it up correctly on your file and interpret every value. Get a quote →
This is the headline result, the answer to the question.
What to read. The Reliability Statistics table has two cells that matter: Cronbach's Alpha, which is your reliability coefficient, and N of Items, the number of items in the scale. If SPSS also prints "Cronbach's Alpha Based on Standardized Items", report the plain Cronbach's Alpha unless your items are on different scales, in which case the standardized value is the fair one.
| Cronbach's Alpha | Cronbach's Alpha Based on Standardized Items | N of Items |
|---|---|---|
| .847 | .851 | 5 |
Here alpha is .847 across five items, which counts as good internal consistency.
This table tells you whether a single weak item is holding the scale back, and what alpha would become without it.
Two columns decide it. Corrected Item-Total Correlation is how well each item correlates with the sum of the others; flag any item below .30 as a poor fit. Cronbach's Alpha if Item Deleted shows what overall alpha would be if that single item were removed; if a value here sits noticeably above your overall alpha, that item is weakening the scale.
| Item | Scale Mean if Item Deleted | Scale Variance if Item Deleted | Corrected Item-Total Correlation | Cronbach's Alpha if Item Deleted |
|---|---|---|---|---|
| Item 1 | 15.32 | 9.41 | .688 | .808 |
| Item 2 | 15.18 | 9.03 | .712 | .801 |
| Item 3 | 15.44 | 9.77 | .641 | .821 |
| Item 4 | 15.51 | 9.28 | .699 | .805 |
| Item 5 | 15.09 | 11.62 | .214 | .883 |
Item 5 is the problem: its Corrected Item-Total Correlation of .214 is below .30, and Cronbach's Alpha if Item Deleted (.883) is higher than the overall .847. Removing or revising Item 5 would raise the scale's reliability.
The number only means something once you place it on the accepted scale.
| Cronbach's alpha | Internal consistency |
|---|---|
| 0.90 and above | Excellent, but check for redundant items above about 0.95 |
| 0.80 to 0.89 | Good |
| 0.70 to 0.79 | Acceptable |
| 0.60 to 0.69 | Questionable, sometimes accepted for short or exploratory scales |
| Below 0.60 | Poor, and below 0.50 unacceptable |
Two cautions sit at the ends of the scale. A very high alpha, above roughly 0.95, is not automatically better; it often means several items are near-duplicates and you could shorten the scale without losing information. A negative alpha is not a tiny reliability, it is a signal that an item was not reverse-coded, so go back to Step 0.
Marks come from the sentence that states the coefficient, the number of items and what it means, not from pasting the table.
APA style reports alpha to two decimals with no leading zero, because it cannot exceed one. State the number of items and, where relevant, that reverse items were recoded first.
Each of these turns a correct procedure into a wrong result or a lost mark.
Alpha is an index of internal consistency, the degree to which the items respond together. It rises with the number of items and with the average correlation between them, which is why a long scale of loosely related items can still post a respectable alpha. Keep three limits in mind when you write it up.
Procedure on this page was checked against the IBM SPSS Statistics documentation, the UCLA statistical computing SPSS guides and university statistics tutorials from Laerd and Kent State.
When items need reverse-coding, subscales must be split out, or the deadline is unforgiving, a specialist runs it in SPSS on your file and explains the output so you can defend it.
Upload your data file, the list of scale items and your deadline, then get a free quote from support.
Approve the price and a statistician who works in SPSS every day starts right away.
Get the output, interpretation and write-up in your account, with free revisions if you need them.
Your work goes to a specialist who passed a subject-specific statistics test and a background check, and who works in SPSS every day. What we deliver is a model answer and study aid for reference, and many students use it to learn the method and check their own analysis.
Everything stays private. Your data and details are confidential, we never contact your school, payments are secure, and the work is original and plagiarism-checked.
Get SPSS help you can trust →"I was a little skeptical at first but this is definitely the best decision I could have made when it comes to getting help with my homework. This is a legit business and the staff are very helpful and professional."
"Without these guys there is no way I would have graduated. Am particularly happy with Moses for all the help he offered. Thanks man! Be blessed."
A common rule of thumb is that alpha of 0.70 or above is acceptable, 0.80 or above is good, and 0.90 or above is excellent. Values above 0.95 can signal that some items are redundant. Below 0.70 the scale is usually treated as having questionable internal consistency, though 0.60 is sometimes accepted for short scales or early-stage research. Report the threshold your field or instructor uses.
A negative alpha almost always means at least one item was not reverse-coded, so its correlations turn negative and drag alpha below zero. Recode those items with Transform then Recode into Different Variables before you run the analysis. A low but positive alpha usually means the items do not measure a single construct, the scale is too short, or a poor item is included. Check the Corrected Item-Total Correlation column and remove or revise items below 0.30.
Yes, if any item is negatively worded relative to the rest of the scale. Reverse-code it first with Transform then Recode into Different Variables, so a high raw score becomes a low score in the same direction as the other items, then feed the reverse-coded versions into the reliability analysis. Skipping this produces a wrong, often negative, alpha.
It is the correlation between each item and the sum of the other items in the scale. Values below 0.30 flag an item that does not fit the rest of the scale and is a candidate for removal or rewording. It is corrected because the item itself is left out of the total it is compared against.
Read the Cronbach's Alpha if Item Deleted column. If removing an item raises alpha meaningfully above the overall value, that item is weakening the scale and may be dropped, but only if doing so still makes theoretical sense and you report the change. Do not delete items just to chase a higher number.
No. Alpha measures internal consistency reliability, meaning how closely the items move together. It does not tell you whether the scale measures the concept it claims to measure, which is validity. A scale can be highly reliable yet measure the wrong thing.
Hand it to a statistician and get accurate output and a clear interpretation before your deadline.
Get my free quote →