0
$\begingroup$

I would like to know if it is possible to use items with different scales to compute a single variable. For example for the variable "authonomy" I would like to include 4 items for the variable computation.

Three questions are with yes/no answer, e.g., "Are you able to choose your methods of work?", and one question is "Are you able to apply your own ideas in your work?" and has a 5-point-scale (1=always, 5=never).

$\endgroup$
  • 2
    $\begingroup$ The short answer is yes of course you can. The long answer depends on just what properties you want your new variable, autonomy, to have so can you edit your question with more detail? Perhaps you already have some ideas which we could help you with? With four question to combine you may need to just use very simple methods. $\endgroup$ – mdewey May 18 '17 at 12:17
  • $\begingroup$ Thank you!You are absolutely right.My overall project is related to making a regression analysis and then verify a moderation effect. In this sense, job satisfaction will be the dependent variable (one item measure, 5-point-response scale).Autonomy (together with other variables such as: skill variety, 5-point-scale; job utility, 5-point-scale, etc.) will be included in the model as dependent variables. In the end, “technology”(7 point scale) will be included as moderator;so a multiplicative term will be computed. I hope I was a little bit more clear.Thank you so much for your precious time! $\endgroup$ – Adelaide May 18 '17 at 13:54
1
$\begingroup$

Yes, it is possible. ONe way to do it is factor analysis. However, to do factor analysis correctly with a mix of variables is a bit tricky. Some people ignore this trickiness and "just do it" but it's better to acknowledge the issue. See this thread for a good discussion of these issues.

$\endgroup$
  • 1
    $\begingroup$ Thank you so much for your super prompt reply!!! And thank you also for the useful link. Your answer lead to another doubt in my analysis. I would like to make a moderation analysis. May I do it also using factors coming from EFA? $\endgroup$ – Adelaide May 18 '17 at 12:00
  • $\begingroup$ You can look at the interaction between factors. $\endgroup$ – Peter Flom May 18 '17 at 12:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.