Batas Angka Cronbach Alpha
Berapa batas angka Cronbach Alpha yang dapat diterima?
Kesimpulan: Tak ada angka keramat, tak ada titik potong ajaib
Diskusi Cronbach Alpha
Sing Yee Ong
i ran cronbach alpha for an established instrument for more than 10 years, 23 items, with sample size of 30. this is social science research
cronbach alpha showed .67.
i read many articles saying that .70 is acceptable.
my question: .
1. .67, is it can’t be accepted? if it is, how can i find some articles to support .67?
2. is sample size of 30, is a factor of it? or should i increase the sample size?
or any other suggestions with it?
Acceptable for what? There is no magic cut-off point that has a cliff to drop off. However, there are many problems with alpha. Ask prof Google
I think it’s still acceptable. Sekaran and Bougie (2010) suggested below 0.6 is considered low or weak.
Sing Yee Ong
hi despite, an established instrument (we cant aspect the same result to the circumstances we test it. I believe you would have tested it in differing for example study sites, or groups, for them such importance on the items would differ. Hence, it depends on how you have mold the established measure based on the respondents type and place. And yes increasing sample size has chances to increase alpha.Since you collected the data may want consider deleting some items based on items deleted to increase cronbach. Or can consider Elan VK suggestion.
Larger sample size may improve the Cronbach’s alpha.
Increase sample size
(Schmitt, 1996) at summary and conclusions 3, “there’s no sacred level…”
Thanks Alex. I should have said that :”there’s no sacred level…”; I said “no magic cut-off point “. But scholars in Malaysia seem to think that above .70 is perfect but .67 or even .69 is rubbish! Little understand of correlation or probability.