Hi Siro Gottardi

First of all, your measurement tool is a questionnaire or checklist, so alpha coefficient or any other reliability coefficients is not suitable. Alpha is suitable for measurement tools such as scales.

I recommend you to get logical evidence (such as expert opinion), not statistical.

Good luck with your research

You are asking to measure the reliability of a scale, i.e., a measure of internal consistency, that is, the closeness of scale items when grouped. I have seen lists such as yours that were initially used as an inventory for qualitative information. The responses were subsequently recoded for MANOVA quantitative analyses, which is often done for more powerful statistical inference. But, some things to think about:

1. You want to insure that you have a very robust sample size collected prior to analyses.

2. Just a question regarding your scoring system. Over the last 30+ years, I haven't seen any inventories with mixed Likert scales...0-4, 0-2, etc. Not to say those scales aren't out there.

3. Also, in regards to your Likert scales, '0' typically means never (as in your first question)--that's fine. But, in your second example, '0' doesn't mean never (it's now less than 15 minutes)....not good. You need to be consistent with your scoring definitions for every question/item for your inventory to have any credibility and any hope for optimal statistical inference.

4. Any opportunity to convert your raw scores to t-scores is highly recommended for others to compare. Most colleagues unfamiliar with your inventory will not know if a subsequent score of 3.2 or 1.8 is above, below, or adequate when reading your findings, because they won't know what the normatives for comparison. Developing t-scores for each of your subscales, however, will require a very large sample size.

5. There are a plethora of psychometric and physiometric inventories in the literature for you to look at for comparison and improvement. That being said, maybe you want to think about using one that has already been validated for exercise with normatives, and save yourself grief and time.

In closing, to obtain optimal outcomes, inventory development is an art as well as extensive effort. If you plan on proceeding, however, you may wish to visit with a colleague that has extensive experience in psychometric or physiometric inventory development. Good luck.