Decision regarding Formative Construct Items.

[Aliza]

Respected Sir/ Ma'am,

In SMARTPLS-4 bootstrapping result, the few of the formative construct items p-values are not coming significant in outer wights, however in outer loadings all the p-values are significant but orignal sample (O) value is showing less than 0.05 and negative. 

So, is it advisable to retain the formative item as p-value is showing significant or the item should be deleted as original sample value is 0.05 and negative?

Thanks & Regards,

Aliza.

[Dr. Vimla]

Dear Aliza,

Considering that:

You are assessing the significance and relevance of the formative indicators

...if the values of outer weights are less than 0.5, it is necessary to check the significance of outer loadings.

...if outer loading value<0.5 and not significant,  then the indicator is considered for the possibility of removal (refer Hair, et al, 2016)

....if indicators have significant outer loadings, having value>0.5, may be retained as being important for the model.

Hope it helps,

Vimla

--
The members of this group are expected to follow the following Protocols:
1. Please search previous posts in the group before posting the question.
2. Don't write the query in someone's post. Always use the option of New topic for the new question. You can do this by writing to dataanaly...@googlegroups.com
3. It’s better to give a proper subject to your post/query. It'll help others while searching.
4. Never write Open-ended queries. This group intends to help research scholars, NOT TO WORK FOR THEM.
5. Never write words like URGENT in your posts. People will help when they are free.
6. Never upload any information about National Seminars/Conferences. Send such information
in personal emails and feel free to share any RESEARCH-related information.
7. No Happy New Year, Happy Diwali, Happy Holi, Happy Birthday, Happy Anniversary, etc. allowed in this group.
8. Asking or sharing Research Papers is NOT ALLOWED.
9. You can share your questionnaire only once.
---
You received this message because you are subscribed to the Google Groups "DataAnalysis" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataanalysistrai...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataanalysistraining/CADeObbKSXYU38QAi_RD04v7yv%2BAWhu1NBQci5dKzqpt7d7W-%3Dg%40mail.gmail.com.

[Aliza]

Thankyou for the reply ma'am. In my study, outer loadings are less than 0.5, but still it's showing significant p value.

In this case, what should be done? 

In Hair, et al, 2022 it's written consider removal of the indicator.

But as per theory, the indicator seems important so could we retain it?

To view this discussion on the web visit https://groups.google.com/d/msgid/dataanalysistraining/CALvRphJoO3DOyBmCMNSWJDvywpXANVWSZDK-VO%2Bvb%2B82yqwgoQ%40mail.gmail.com.

[Neeraj Kaushik]

Dear Aliza

Dr Vimla has given a perfect explanation. Now let me add a few points here:

1. Most important- while talking about reflective construct we talk about the term factor loading and while talking about the formative construct, we talk about factor weight.

2. The various items in a formative construct indicate the various dimensions of that construct, hence all shd be considered. Technically, there shd not be any correlation among these items as all are different dimensions. 

3. But unknowingly, there are patterns in responses which generate correlation among these items.

So please note that most of the time, the items are out of a formative construct because of multicollinearity.

So drop such items having p>0.05

4. However, if the factor weight is large enough say >0.5 then its significance does not matter and you can etain that even it its p>0.05

Hope it helps

Best wishes

Neeraj 

To view this discussion on the web visit https://groups.google.com/d/msgid/dataanalysistraining/CADeObbJs9KVw9X4%3Dq8jqqqKoM4XD5hSm-vazGvH0hS0-c9DCVA%40mail.gmail.com.