Cronbach’s alpha to measure scale reliability

Cronbach’s alpha to measure scale reliability

For the Cronbach’s alpha test, it is assumed that a lower value of the alpha represents fewer questions in the questionnaires used or weak correlation among the items.Besides, the concept of reliability measured by Cronbach’s alpha assumes that unidimensionality usually exists in a sample of test items (Gerber & Finn, 2015).

Moreover, Cronbach’s Alpha as an index of reliability ought to coincide with the assumptions of the essentially tau-equivalent model.The premises were met. Incase the assumptions are not met the researcher should manipulate the data cautiously to ensure that questions are related to the subject under study. Probably the piloted questionnaire should, therefore, be edited and this would ensure reliability in data collected.If the assumption of uni-dimensionality in the sample items is violated, then the results could cause a significant underestimate of the exact reliability.

The null hypothesis, Ho: the scale is reliable according to the Cronbach’s alpha

The alternative hypothesis, H1: the scale is not reliable according to the Cronbach’s alpha

From the reliability test in the table above the Cronbach’s alpha shows an advanced case of internal consistency in the scale used. The reliability estimate is favorable, and this indicates that the measurement error is minimal. The Cronbach’s alpha value obtained usually gives an overall reliability coefficient for a set of variables.

Therefore, for all the questions asked in the questionnaire the overall Cronbach’s alpha essentially presents you with an overall measure of reliability for all of them together. Therefore the problems can be said to be reliable, and the researcher can comfortably undertake a validity test since reliability has been established.

References

Gerber, S. B. & Finn, K. V. (2015). Using SPSS for Windows: Data analysis and graphics. New York: Springer.

Britz, G., Emerling, D., Hare, L., Hoerl, R., & Shade, J. (2017). How to teach others to apply statistical thinking? Quality Progress30(6), 67.

Do you need high quality Custom Essay Writing Services?

Custom Essay writing Service