Reliability Analysis in SPSS

How legitimate or justifiable a research is, that is based on an inconsistent instrument? Consistency of data is central to the concept of reliability. Reliability simply refers to the confidence a researcher places on the questionnaire to provide the same numeric value when the measurement is repeated on the same subject. How to do it is described below. If you wish to follow along with this example, you should start SPSS and open the Training Finalized.sav file.

Reliability Coefficient

Commonly used technique for assessing reliability is Cronbach’s alpha for internal reliability of a set of questions (Interval and Ratio Scale). Ideally, the Cronbach’s alpha coefficient of a scale should be above .7. Following guideline developed Gliem & Gliem (2003) by presented in the table can be a guide to evaluate the reliability coefficient.

Cronbach’s Alpha

Cronbach’s alpha values are quite sensitive to the number of items in the scale. With short scales (e.g. scales with fewer than ten items) it is common to find quite low Cronbach’s values (e.g. .5). Reliability is normally reported under the head of instrumentation in the methodology section. If your scale contains some items that are negatively worded (common in psychological measures), these need to be ‘reversed’ before checking reliability. For instance we have a Scale named “Work Morale” having the following questions answered on likert scale (1 – Strongly Disagree to 5 – Strongly Agree)

1. The atmosphere at work is pretty good.
2. Everyone around here looks forward to come to work.
3. The Company is going places.
4. There is no future for this company (R).
5. We all pull together this company.

We can see that item 4 is in reverse order, it is negative while all other questions are positive. Now when you enter the response for item 4 into SPSS, you need to reverse the entry, for instance if the respondent has said 5, you will enter 1 into SPSS, similarly 4 will be exchange with 2 and vice versa.

Procedure for checking the reliability of a scale

1. 1. Choose Analyze → Scale → Reliability Analysis
2. 2. You will see Reliability Analysis dialog box.
3. 3. Select the items whose reliability is to be assessed from the variable list box. Select only the variables for One Construct (Scale) at a time. For this example we will select TNA1 to TNA4 related to scale Training Needs Analysis (Once Added your Dialog box should resemble the one in figure 8.1).
4. 4. Add the selected list of variables to items list box.
5. 5. Click on the Statistics button which will open a dialog box. Check Item, Scale, and Scale if item deleted from Descriptive Statistics group box. Click on Continue to return to the main dialog box then click on OK to run the analysis.

The output from analysis is shown below:

Interpretation of Results

The output shows a number of tables. The first table shows the Case Processing Summary, showing the total number of valid cases and if any data was excluded from the analysis. The second table of Reliability Statisticsis the table of interest, it gives the value of the Cronbach’s alpha and the number of items selected for the scale. For our scale of Training Needs AnalysisCronbach’s alpha value reported to be 0.900. This recommends that the scale is consistent and highly reliable.

SPSS also provides us with descriptive statistics. The table titled Item Statisticsgives item-wise mean and standard deviation values. Item-Total Statisticstableis important. The fourth column in this table, titled Corrected Item-Total Correlationgives an indication of the degree to which each item correlates with the composite score for the scale. The last column labeled Cronbach’s Alpha if Item Deletedcan help improve the reliability of the scale. It shows if removing a certain item will improve the overall reliability of the scale, however in this particular case the Cronbach’s alpha won’t improve by removing any of the items.

Please Check!

• Check that the number of cases is correct (in the Case Processing Summary table) and that the number of items is correct (in the Reliability Statistics table). In this particular case both are correct, none of the cases are excluded and the number is items are 4 which are correct.
• Check the Inter-Item Correlation Matrix for negative values. All values should be positive, indicating that the items are measuring the same underlying characteristic and account for the same construct. The presence of negative values could indicate that some of the items have not been correctly reverse scored.

Improving Reliability

The overall reliability of the scale can be improved by following a few simple guidelines:

• The Corrected Item-Total Correlation column in the Item-Total Statistics table provides an indication of the degree to which each item correlates with thetotal score. Low values (less than .3) here indicate that the item is measuring somethingdifferent from the scale as a whole. If scale’s overall Cronbach’s alpha istoo low (e.g. less than .7) and you have checked for reverse items that might not have been entered properly, It would be a good idea to consider removing items with low item-total correlations.
• In the column headed Alpha if Item Deleted, the impact of removing each item from the scale is given. Compare these values in the column headed Alpha if Item Deleted with the alpha value obtained. If any of the values in this column are higher than the final alpha value, you may want to consider removing this item from the scale.

Reporting Cronbach’s Alpha

It is normally reported in the methodology section where instrument are discussed. After discussing the scale i-e their Number of items in the scale, scale for response you can describe the reliability of the instrument. An example of how to report is shared below.

“For recruitment there were ten items asking how the organizations recruit new employees. The reliability coefficient for recruitment was 0.787.”

(Source: Hashim, J. (2010). Human resource management practices on organisational commitment: The islamic perspective. Personnel Review, 39(6), 785-799.)