Smallest Path coefficient to report?

[Mattias_O]

Hi - I am working on a complex model, and I wish to reduce the number of paths. Some of the paths are strong and some are weak, - it would make some kind of sense to not include some paths.

The question is, - how strong or weak can a path be for inclusion? 0,1? 0,2? Is there some kind of best practice, or reference to cite on this issue?

I realize there may be differences in between fields, - but in marketing - what is a too weak path to include? What is a path strong enough to include?

Very Best Regards!

:-)

[nayabeyes]

Hi - I am working on a complex model, and I wish to reduce the number of paths. Some of the paths are strong and some are weak, - it would make some kind of sense to not include some paths.

The question is, - how strong or weak can a path be for inclusion? 0,1? 0,2? Is there some kind of best practice, or reference to cite on this issue?

I realize there may be differences in between fields, - but in marketing - what is a too weak path to include? What is a path strong enough to include?

Very Best Regards!

:-)

Hi,
I think if the path is significant by checking its t-value then no problem. For more details check Hair et al., 2013

[Hengkov]

Hi,

Path > 0.3 is strong, but < 0.3 is less.
If your beta small, the T-statistic not significant because beta/standard error = T-statistic.

Regards,