In an educational setting, assessments will play a vital role in differentiating between the better performers and the poor performers. In order to achieve this, each test instrument should produce a high degree of reliability as well as validity. Many statistical tools as well as theories are applied in making an assessment reliable and valid, and among them, the Cronbach’s alpha will also play a major part.
When we consider assessment instruments (e.g Set of MCQ questions), the correlation between each test item within the tool would determine the reliability on how we differentiate between ‘good’ and the ‘bad’ students to a certain extent. Such relationship between each item of a test instrument is considered as the ‘internal consistency’ and one of the widely used methods to assess the same would be the ‘Cronbach’s alpha’.
How to calculate?
In order to calculate Cronbach’s alpha, we can make use of a statistical package such as the SPSS. Once the data sheet is prepared by having students representing each row and each test item as the columns, the marks obtained by the students for each test item will fill the table. Thereafter, the SPSS will provide the user with a score known as the Cronbach’s alpha and depending on the number, the user can determine the internal consistency of its test items.
How to interpret?
In general, a Cronbach’s alpha value of 0.8 and above is considered ideal whereas a score of more the 0.7 is desirable. But, the reason for a high Cronbach’s alpha value to be interpreted as a score that indicates a ‘good’ question or a test item needs to be understood. The fact is, the Cronbach’s alpha measures the variance between the test scores and higher the variation that is obtained from the test instrument, higher the discriminatory value it will have in separating the ‘good’ students from the ‘bad’ ones. Therefore, higher Cronbach’s alpha value is desirable whereas a value lower than this would lessen the reliability of that particular test instrument.
How to improve Cronbach’s alpha?
The same SPSS package would give use a tool to see how the Cronbach’s alpha would change when one particular test item is deleted. As such, if the deletion would lead to an overall reduction in the alpha value, it means that the test item is a good one and that it should be kept. Similarly, if the Cronbach’s alpha rises with the deletion, it means that the test item is less reliable and would improve the reliability if it is removed from the assessment instrument or modified to suite the need.