Yusuf, I had a look at the paper and I am not sure what exactly the index is and how it was measured. The authors performed several analyses and found a CFA with three factors to fit the data best for all investigated populations. As the factors were correlated, they specified a second-order factor which is statistically equivalent to the correlated factors model. The assumption seemed to be that the common variance was large enough to be a measure of overall brand equity, called "higher-order brand equity" - whatever that means.

The authors used composite scores and correlated this variable with their newly developed MBE index, using the same data. I don't think that they used a formative measurement model, as far as I could find out they just summed the items.

You do not have to create an index after having performed a factor analysis. A first-order or second-order factor analysis would be just fine and better to be justified. in order to estimate a higher-order model, you just have to let all three first-order factors load onto the second-order factor. You need to fix one factor loading on the second-order factor to one in order to set a scale for this factor, just as you would do for the first-order factors.

Does this help?

Regards, Karin

Thank you dear Ms. Karin,

Sometimes it is really hard finding your way especially studying with an unestablished theory. Your answer really helped, thank you.

Regards,

Yusuf