Validity of scales without Cronbach's alpha

[Maria Neto]

Hi everyone,
I want to build a scale using the Comparative Manifesto Project data, but I read (Prosser 2014) that Cronbach's alpha is not the right tool to assess its validity for this type of variables. ( In fact, the alpha for the right and left components of the rile scale is very low).
Can someone explain how I can use a different method and write the syntax, please?
Thank you in advance

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of helpful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data. Remember, few of us are from your specific area, so few of us have the least idea what Comparative Manifest Project data means.

You need to pose your question in general terms. What kinds of variables are you working with? What is Prosser's objection to alpha? What are left and right components and rile scale?

I assume you are talking reliability not validity? Obviously, some important forms of validity depend on the tie of the questions to the construct - something on which you must be the expert,not us.

Beyond alpha, many use forms of factor analysis or SEM to assess reliability.

[Rich Goldstein]

in addition to Phil's comment, note that Chuck Huber of StataCorp has given at least one user-group talk related to this: 2012 at the US meeting; there may be others also; all these are available at the Stata web site