Reliability analysis with XLStat

Reliability analysis is used to assess the quality of measurement scales and, in particular, the accuracy of the scales adopted in a questionnaire. The Cronbach's alpha is a reliability score often used to assess measurement scales' reliability.

What is Cronbach's alpha?

Cronbach's alpha measures internal consistency, which translates how closely related a set of items is. It is considered to be a measure of scale reliability. Indeed, Cronbach's alpha assesses the degree to which a scale consistently reflects the construct it is measuring.

The alpha coefficient lies between 0 and 1. The closer Cronbach's alpha coefficient is to 1, the greater the internal consistency of the items in the scale.

George and Mallery (2019) provided the following rules of thumb:

• > 0.9 Excellent
• 0.8 – Good
• 0.7 – Acceptable
• 0.6 – Questionable
• 0.5 – Poor
• < 0.5 – Unacceptable

Description of the dataset

In this fictitious database, the researcher measured three variables using 7-points Likert scales in a questionnaire:

• Perceived ad relevance with three items
• Perceived ad credibility with four items
• Click intention with two items

We want to know the internal consistency of these three measurement scales using a reliability analysis and, more precisely, by obtaining Cronbach's alpha scores.

Setting up a reliability analysis

Open XLStat and Describing data > Reliability analysis

Once you have clicked on the button, the dialog box "Reliability analysis" appears.

In Observation/Item table, select the three items of the perceived ad relevance measurement scale (columns A, B, and C).

In Type of reliability, select "Internal model".

Since the column title for the variables is already selected, leave the Variable labels option activated.

In Missing data, select "Do not accept missing data". In Outputs, only select "Cronbach's alpha statistics". Finally, make sure that no chats are selected in the Chats tab.

Interpreting the results of reliability analysis

In the column Cronbach's alpha, the Cronbach's alpha is 0.95 for the perceived ad relevance measurement scale, which indicates an excellent internal consistency of the items.

To find the internal consistency for the other scales, you need to reproduce the same procedure in XLStat. To do that, you can select the four items of the perceived ad credibility scale and then the two items of the click intention scale.

The table shows the Cronbach alpha scores obtained from the data. The variables of perceived ad relevance (α = 0.95), perceived ad credibility (α = 0.87), and click intention (α = 0.88) show good to excellent measurement scales reliability.

Table: Cronbach's alphas for reliability analysis