CRONBACH'S ALPHA SCALE: The scale had been interpreted as follow:

> .9 – Excellent,

> .8 – Good,

> .7 – Acceptable,

> .6 – Questionable,

> .5 – Poor, and

< .5 – Unacceptable

See attached file for reference; p. 87 (quoting Mallery).

NOTE:  Although researchers still rely on Cronbach's alpha, it is worthing noting that Cronbach himself had recently admitted that the Cronbach’s alpha may not be an appropriate test for reliability. Cronbach wrote that:

 “I no longer regard the alpha formula as the most appropriate way to examine most data. Over the years, my associate and I developed the complex generalizability (G) theory. (Cronbach et al. (1963); Cronbach et. al. (1973); see also Brennan (2001); Shavelson and Webb (1991), which can be simplified to deal specifically with a simple two way matrix and produce coefficient alpha (Cronbach (2004), p. 403). Cited in N.M. Webb, R.J. Shavelson and E.H. Haertel (2006). “Reliability Coefficients and Generalizibility Theory.”  See Handbook of Statistics, Vol. 26, p. 2. ISSN 0169-7161.

Cronbach's original article may be here: Coefficient alpha and the internal structure of tests. Psychometrika, 16,297-334. You need account login to access at https://www.psychometricsociety.org/content/psychometrika

See attached for G Theory.

Hi, 

I won't discuss much because there is already a very good discussion about this topic. I will just post some more literature that could add up to what both professors have already posted.

http://pioneer.netserv.chula.ac.th/~ppongsa/2013605/Cronbach.pdf

This is a conference paper regarding internal consistency.

Hope it helps

This traditional method (Cronbach alpha) is still implemented even the previous article found that the inconsistency of this method. Thats why i prefer use Composite reliability of joreskog rho for reliability test.

Okay, thanks everybody! This is very useful to me, I already read that there were doubts about the Cronbach alpha but I didn't know they had already proven that it kind of fails as a measure for the reliability ! 

See also Cronbach's reflection on his paper:

And his paper itself is avialable from:

http://garfield.library.upenn.edu/classics1978/A1978EQ39200002.pdf

http://kttm.hoasen.edu.vn/sites/default/files/2011/12/22/cronbach_1951_coefficient_alpha.pdf

I do agree with what Dr. Paul present above, and simply speaking of Cronbach alpha it is the result of all possilbe intercorrelation among all items in the scale. of course the statistical program is capable to handle this.

For all the issues with alpha, it doe remain useful and incredibly conventient.  However, remember that interpretation guidelines like those above are purely arbitrary.

The best article on the topic is Cortina (1993) http://www.psycholosphere.com/what%20is%20coefficient%20alpha%20by%20Cortina.pdf.