Hi fellows,

I have some issues interpreting the outcome of my CFA.

Basically I'm checking the impact of various second order construct on performance.

Somehow like this:

Items <-- factor 1 (>.60)
Items <-- factor 2 (>.60)
Items <-- factor 3 (>.60)

factor 1 <-- second oder factor 1 (.54)
factor 2 <-- second oder factor 1 (.82)
factor 3 <-- second oder factor 1 (.30)

second oder factor 1 --> Performace (.40)


All item show standardised loadings > 0.6 on their destinct factor which is inline with the suggestion of Bagozzi and Yi (1988)

However, factor loadings on the proposed 2nd order constructs vary from 0.3 to 0.8 and I wonder if there are also recommended tresholds for factor / second oder facor loadings, or if this just means that there are stronger and weaker factors within the higher construct.

How'd you interpret the results above?

Would be great if I can get your thoughts on that!

Happy researching!

Thanks to you two!

I quickly checked Ying and Stake and it looks interesting...

@CR1980: Sure access played a role, usually you don't get executives for an hour +...
I guess the study cost minimum 10000€+ (traveling, time etc...)

That's a good hint!

@DROT: I'm a critical realist, so that somehow should fit :)

Concerning the Case Study approach: Labour Union gave me the restriction that neither the name of the company nor names of the employees will show up in any publication. Is there a good argument (besides the given restriction) for this anonymity?

Cheers,

Dear fellow PhDs,

I'm doing a PhD (Marketing/Sales) in a company together with a Uni in the UK.

I already completed my first study (interviews) within my company and I'm very happy about the results.
A lot of new insights were given, what for the first time made me feel very comfortable with my topic. ;)

However, I'm working on my data collection chapter right now and do not really find literature that supports the approach of performing interviews in only one company. One might argue that this study is biased and cannot be generalised...

I already thought about the following argumentation, however it would be great if you could give me some input as well...

- Purposive sampling
- World leading 200k employee company
- Management System that is in the focus of my research is successfully applied for 20 years
- Data is fully comparable - no cross sectional influence
- Hofstede (1989) argues out that the use of a single multinational employer eliminates the effect of the corporate policy and management practices from different companies...

- ....

Thanks for your support, have a great Weekend!