categorical indicator jointly with manifest variables?

Research topics can be discussed in this area.
Post Reply
ruchi
PLS Expert User
Posts: 75
Joined: Wed Dec 22, 2010 3:09 pm
Real name and title:

categorical indicator jointly with manifest variables?

Post by ruchi » Mon Mar 21, 2011 12:47 pm

Hi

I have some categorical variable like government policy which takes the value as
1 - for policy 1
2-policy 2
3-policy 3

I have 3 manifest variables measured on likert scale.

These 4 variables (1 categorical and 3 manifest) form a latent score.

How can i model this in smartpls to see its relation with my dependent variable (measured by 2 manifest variables on likert scale).

It will be really helpful if someone can guide me on same

Thanks
ruchi

User avatar
Diogenes
PLS Super-Expert
Posts: 905
Joined: Sat Oct 15, 2005 5:13 pm
Real name and title:
Location: São Paulo - BRAZIL
Contact:

Post by Diogenes » Sun May 08, 2011 2:06 pm

Hi,

Gov. policy is a categorical or ordinal variable?

If it is categorical:
- use the three variables as indicators of the LV, and split the data set in three. Run a multigroup analysis.

If it is ordinal:
- use the four variables as indicators of the LV.

Best regards,
Bido

ruchi
PLS Expert User
Posts: 75
Joined: Wed Dec 22, 2010 3:09 pm
Real name and title:

direct effect of gender and interaction

Post by ruchi » Sun May 08, 2011 5:32 pm

Dear Sir

I have my EXP variable with values

Respondent EXP (in years)
Respondent1 5.6
Respondent2 6
Respondent3 3
Respondent4 6
Respondent5 6
Respondent6 5.4
Respondent7 4.5
Respondent8 4.1
Respondent9 4
Respondent10 6.5
Respondent11 4
Respondent12 8.2
Respondent13 5.3

Now can I model it using smartPLS and "create moderating effect" using this data
For this first i should connect my LV ( with one indicator of EXP) to my dependent variable and then use "create moderating effect" on dependent variable with Moderator as-EXP and predictor as another IV( measured by 4 items on likert scale)
Now if i categorise my EXP variable as
1 indicates –very less experience (0-5yrs) , 2 indicates less experience(6-10 yrs) ,3 indicates medium experience (11-15 yrs) and 4 indicates high experience(more than 15).
In this case i cannot use ""create moderating effect". For this I need to do multigroup analysis. Same will be the case for gender variable (0-for male and 1 -for female).I cannot use "create moderating effect" for my gender variable. But can this be done in hierarchial regression in spss ??
Please correct me if i am wrong.
Also can the direct effect of variables like gender (0-for male and 1 -for female) and EXP variable (measured in terms of categories and as continous) can be done by creating a LV with single item and connecting it to my other dependent variables??
It will be of great help if you can please clarify the things.

Thanks
Regards
Ruchi

User avatar
Diogenes
PLS Super-Expert
Posts: 905
Joined: Sat Oct 15, 2005 5:13 pm
Real name and title:
Location: São Paulo - BRAZIL
Contact:

Post by Diogenes » Wed May 11, 2011 12:37 am

- Hi,

Now can I model it using smartPLS and "create moderating effect" using this data
For this first i should connect my LV ( with one indicator of EXP) to my dependent variable and then use "create moderating effect" on dependent variable with Moderator as-EXP and predictor as another IV( measured by 4 items on likert scale)
- It is ok.

Now if i categorise my EXP variable as
1 indicates –very less experience (0-5yrs) , 2 indicates less experience(6-10 yrs) ,3 indicates medium experience (11-15 yrs) and 4 indicates high experience(more than 15).
In this case i cannot use ""create moderating effect". For this I need to do multigroup analysis.
- In this way you are losing information (the continuous level of measurement is better than ordinal). I think that the first option is better.

Same will be the case for gender variable (0-for male and 1 -for female).I cannot use "create moderating effect" for my gender variable. But can this be done in hierarchial regression in spss ??
- As a dummy variable (0/1) the gender could be used in the same way that we use it in the multiple regression. Yes, you could create the moderating effect.

Also can the direct effect of variables like gender (0-for male and 1 -for female) and EXP variable (measured as continuous) can be done by creating a LV with single item and connecting it to my other dependent variables??
- Yes.

- Best regards,

- Bido

ruchi
PLS Expert User
Posts: 75
Joined: Wed Dec 22, 2010 3:09 pm
Real name and title:

Post by ruchi » Wed May 11, 2011 5:35 am

Dear Sir

Thanks for clarifying my doubt

It means , That do study the direct effect of gender variable (coded a s0-male and 1 for female) , I can use it as a single variable with single Item and connect to my dependent variables in smartPLS. then no need to do group comparison (as my samples in group are less).

The result/path co-effcient will tell about female. ??
Please correct me if i am wrong

But If i want to see Moderating effect of gender , It is must to do group analysis in smartPLS. I have to create 2 groups (datasets) and do same modelling and then campare.

I cannot do it in smartPLS using "create moderating effect" with my gender variable ( as single LV with one item haning values 0-male and 1)-female).

But then how to go if my sample size for the1 group (like male) is less -30 people

Please let me know .
I hope after this my queries will largely be clarified on moderating effect.


Thanks
Warm Regards
Ruchi

User avatar
Diogenes
PLS Super-Expert
Posts: 905
Joined: Sat Oct 15, 2005 5:13 pm
Real name and title:
Location: São Paulo - BRAZIL
Contact:

Post by Diogenes » Wed May 11, 2011 12:51 pm

You could use the "create moderating effect" even with gender (dummy coded), see details in

AGUINIS, H. (2004). Regression analysis for categorical moderators. New York, the guilford Press.

One example in p.24-35.

Best regards,

Bido

ruchi
PLS Expert User
Posts: 75
Joined: Wed Dec 22, 2010 3:09 pm
Real name and title:

Post by ruchi » Mon Aug 01, 2011 4:24 am

Dear Sir

For direct effect of age , do i need to connect it to all my dependent variables or to only main dependent variable or to all the the variables where i want to see the effect ??

Does this direct effect same as main effect. How to interpret it?

Thanks
Ruchi

Post Reply