Cronbach Alpha is a reliability test conducted within SPSS in order to measure the internal consistency i.e. reliability of the measuring instrument (Questionnaire). It is most commonly used when the questionnaire is developed using multiple likert scale statements and therefore to determine if the scale is reliable or not.
Example of Cronbach Alpha
A researcher developed a 9 question questionnaire to determine how safe people feel at their work. In addition the questions were on the 5-point Likert Scale with responses ranging from “Strongly agree” to “Strongly disagree”. In order to determine if the questionnaire could “reliably” measure the latent variable i.e. feeling of safety, Cronbach alpha test was conducted. The acceptable reliability value is .6. Therefore if your questionnaire’s reliability result is more than .6 then your questionnaire is considered “reliable”.
Cronbach Alpha in SPSS
- Step 1: On the Menu, Click Analyze -> Scale -> Reliability Analysis (Figure 1).
- Step 2: Next the Reliability Dialog box would open which is shown below. All the questions which are to be checked for Reliability are transferred to the “Items box” (Figure 2)
- Step 3: Next Click on “Statistics” which lead to another dialog box. Here under “Descriptives for” check Item, Scale, and Scale if Item Deleted. In Inter-Item Check Correlations. Click “Continue” and then “OK” to Run the test (Figure 3).
Output and results
Table 1 given below is the Reliability Statistics Table which provides the value for Cronbach alpha which in this case is .805 and reflects high reliability of the measuring instrument. Furthermore, it indicates high level of internal consistency with respect to the specific sample.
Table 2 “Item Total Statistics” presents the results for Cronbach Alpha if Item Deleted. It is the measure of Reliability to determine the “Item” which when deleted would enhance the overall reliability of the measuring instrument.
As Table 2 shows above, that other than Question 8, if one delete any other question then the reliability will result lower Cronbach Alpha. However the Corrected Item-Total Correlation value (which denotes the correlation between each item or question within the questionnaire and total reliability score of the questionnaire)for Question 8 is very low i.e. .128 and therefore low correlation indicates that we should consider removing the item from the overall questionnaire.
Latest posts by Priya Chetty (see all)
- What is the relevance of significant results in regression analysis? - February 28, 2020
- How to improve the correlation between the variables? - February 27, 2020
- Manufacturing procurement management strategies - February 25, 2020