1
$\begingroup$

A questionnaire survey was conducted to explore underlying factors affecting pedestrian road crossing behavior. The respondents were asked to answer Likert type questions along with demographic questions (such age, gender etc.).

As the questionnaire contained new questions as compared to some existing validated scales, therefore, exploratory factor analysis (EFA) was carried out after collecting the data.

I got four underlying factors after conducting EFA.

Can I use the factor scores (obtained through EFA) for further analysis?

For example, can I use two-way ANOVA to determine the impact of age and gender on a factor?

$\endgroup$

1 Answer 1

3
$\begingroup$

First, supposing you asked them the following three questions (among others) and the scale is "never", "seldom", "often", "always":

  1. How often do you cross intersections out of turn?
  2. How often do you run across large roads instead of using pedestrian bridges?
  3. How often do you cross intersections doing cartwheels?

Then, supposing your EFA yielded a factor that loads highly on these three items and you decided to call that factor "recklessness".

And last, supposing you only included the questionnaire items in the EFA (not the background questions).

Then yes! It is a great idea to regress recklessness on age and see, for example, if younger people are more reckless.

$\endgroup$
10
  • $\begingroup$ Could you explain why you think background questions need to be excluded from such an EFA? $\endgroup$
    – rolando2
    Commented Jul 28, 2018 at 11:23
  • 1
    $\begingroup$ Suppose we include age in the EFA, and our factors have some (positive or negative, large or small) loading for age. What would it then mean to regress the factors on age? We already know age can explain them, since we put age information in the factors. $\endgroup$
    – suckrates
    Commented Jul 28, 2018 at 13:01
  • $\begingroup$ Thank you very much for your reply. What I actually wanted to ask is that whether it is necessary to validate the scale by conducting CFA before regressing the factor on age? $\endgroup$ Commented Jul 28, 2018 at 14:49
  • 1
    $\begingroup$ Maybe the courteous thing would be to recognize (upvote and possibly accept) Alvaro Fuentes' answer before moving on to this new question about CFA. $\endgroup$
    – rolando2
    Commented Jul 28, 2018 at 15:01
  • 2
    $\begingroup$ You can find ideas here under "Validity": courses.lumenlearning.com/… $\endgroup$
    – suckrates
    Commented Jul 28, 2018 at 16:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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