Hi,
I have a general question regarding weather or not one can use cronach's alfa for measuring scale reliability? As far as I understand alfa is only used for metric data, right ?
Thanks
Davit
The alpha coefficient gives you an insight into the reliability of the used construct. If the alpha should be to low, you should take a look at the loadings of the single dimensions (factor analysis). Yes, this indicator is used for metric scaled data.
Dear Samuel,
Thank you for your quick feedback.
Looking forward to hearing from you
Best regards,
Davit
I am not too sure whether I got all the points. However, if you composed a scale by "random" items of your choice, you will need to show that the items do fit together and the scale is valid. Besides, the question of reliability (cronbachs alpha) , you will need to show also validity of your scale. There are several possibilities you should look at, e.g. exploratory factor analysis.
Concerning the cronbachs alpha, I guess that you can apply it to your scale to prove reliability.
Dear Samuel,
Let me clarify some of the points in detail.
We collected data to conduct Psychographic Market Segmentation. For this purpose VALS technique (Values and Lifestyles) was employed. In total, 36 items had been designed (all of the measure the same construct; they are not random items) each measured on 4 point scale. Firstly, we intend to conduct factor analysis and then move to cluster analysis. For factor analysis, data has to be measured on at least metric scale, however, in my case can I treat 4 point scale as metric? Personally, it seems 4 point scale (1 strongly disagree, 2 disagree, 3 agree, 4 strongly agree) is by nature "ordinal" but can I treat it as a scale?
Best regards,
Davit
Hi Davit,
from my point of view, you can treat it as a metric scale because it is an interval scale in particular. Such a scale is not ordinal, but metric.
Best
samuel
Cronbach Alpha is used routinely in many application with this type of data (Lickert Scale or variations thereof), even though its theory was originally derived for multivariate normal data. In most software packages (such as SPSS) you may ask for the difference it would make if a variable is eliminated. A potential problem could be caused by questions which are inverted, in which case they may have to be recoded from 1 2 3 4 to 4 3 2 1. Another thing to take into account is that some of the 36 items may be measuring different factors (aspects). I would recommend doing a factor analysis first, and then calculating the Cronbach Alpha separately for the items loading on each factor.
Francois E Steffens sir, i have a question regarding the reliability testing,
in my questionnaire, i have interval scale questions that means closed ended questions, so can i use Cronbach alpha test for checking the reliability?
if possible please attach some article so that i can read that article.
thank you
- Badges
- Science topic