Interaction effects
-
- PLS Expert User
- Posts: 31
- Joined: Mon Jun 14, 2010 7:35 pm
- Real name and title:
Hey Christian,
it still doesn´t work and I really don´t know why :(
I have now 2 data files:
1) male dataset with male variable and 1 standing in the male column -> 224 cases left
2) female dataset with female variable and 0 standing in the female column -> 278 cases left
But still the calculation errors as mentioned abouve occur :( But I have enough cases, so I really don´t understand what I do wrong.
When I split up the data file into two and leave all the 520 cases within the set, and in one data file the code for male is 1 and the rest are 0 and in the other data file the code for female is 1 and the rest is 0, then the calculation works, but the path coefficient then is e.g for male -0,154 and for female 0,154...I think that this is wrong.
Best regards,
Larissa
it still doesn´t work and I really don´t know why :(
I have now 2 data files:
1) male dataset with male variable and 1 standing in the male column -> 224 cases left
2) female dataset with female variable and 0 standing in the female column -> 278 cases left
But still the calculation errors as mentioned abouve occur :( But I have enough cases, so I really don´t understand what I do wrong.
When I split up the data file into two and leave all the 520 cases within the set, and in one data file the code for male is 1 and the rest are 0 and in the other data file the code for female is 1 and the rest is 0, then the calculation works, but the path coefficient then is e.g for male -0,154 and for female 0,154...I think that this is wrong.
Best regards,
Larissa
-
- PLS Senior User
- Posts: 28
- Joined: Thu Jun 24, 2010 5:38 am
- Real name and title:
-
- PLS Expert User
- Posts: 248
- Joined: Sat Jul 25, 2009 1:34 pm
- Real name and title:
-
- PLS Junior User
- Posts: 3
- Joined: Fri Jan 07, 2011 2:08 pm
- Real name and title:
Hello,
I have a similar problem I guess, but I don't speak German, so I cannot use the paper. I would greatly appreciate someone's help!
I have 4 excel files with data of 4 different groups. These 4 groups were 4 different manipulations in my survey. In order to include these manipulations in my model (independent variable), I created a dummy variable. Since I have 4 different excel files, I just listed all respondents in the colomn with the number 1, because I want to include all the data. If I link this variable to my model, and try to run the analysis, I get a message saying:
A singular matrix occurred during the estimation of the path coefficients using the path weighing scheme. Setting another weighting scheme could solve the problem.
I have no idea how to solve this. Could anyone please help me with this?
Thanks a million in advance.
I have a similar problem I guess, but I don't speak German, so I cannot use the paper. I would greatly appreciate someone's help!
I have 4 excel files with data of 4 different groups. These 4 groups were 4 different manipulations in my survey. In order to include these manipulations in my model (independent variable), I created a dummy variable. Since I have 4 different excel files, I just listed all respondents in the colomn with the number 1, because I want to include all the data. If I link this variable to my model, and try to run the analysis, I get a message saying:
A singular matrix occurred during the estimation of the path coefficients using the path weighing scheme. Setting another weighting scheme could solve the problem.
I have no idea how to solve this. Could anyone please help me with this?
Thanks a million in advance.
-
- PLS Expert User
- Posts: 248
- Joined: Sat Jul 25, 2009 1:34 pm
- Real name and title:
Hey,
You should not use the variable/column with your dummy code in your model. After you splitting your file in the different groups there are only e. g. 1 in the column. Therefore you have got an error message. (A second reason could be, that your model is not recursive.)
Here a very good article, this time in English, who also perform a group comparison with four groups in PLS:
Eberl, Markus (2010): An Application of PLS in Multi-Group Analysis, in: Esposito Vinzi, V./Chin, W./Henseler. J./Wang, H. (Eds.): Handbook of Partial Least Squares: Concepts, Methods and Applications, Berlin, pp. 487-514.
Best regards,
Christian
You should not use the variable/column with your dummy code in your model. After you splitting your file in the different groups there are only e. g. 1 in the column. Therefore you have got an error message. (A second reason could be, that your model is not recursive.)
Here a very good article, this time in English, who also perform a group comparison with four groups in PLS:
Eberl, Markus (2010): An Application of PLS in Multi-Group Analysis, in: Esposito Vinzi, V./Chin, W./Henseler. J./Wang, H. (Eds.): Handbook of Partial Least Squares: Concepts, Methods and Applications, Berlin, pp. 487-514.
Best regards,
Christian
-
- PLS Junior User
- Posts: 3
- Joined: Fri Jan 07, 2011 2:08 pm
- Real name and title:
Hi Christian,
thank you for your quick answer. I actually just realized the same thing, to exlude the variable in total.
I am sorry for this stupid question, but do you maybe have a link to the article? I can only find it on sites on which I need to pay 25 dollars for the chapter.
Could you also recommend literature for reviewing the results?
Thank you very much in advance.
Kind regards,
Renee
thank you for your quick answer. I actually just realized the same thing, to exlude the variable in total.
I am sorry for this stupid question, but do you maybe have a link to the article? I can only find it on sites on which I need to pay 25 dollars for the chapter.
Could you also recommend literature for reviewing the results?
Thank you very much in advance.
Kind regards,
Renee
-
- PLS Junior User
- Posts: 3
- Joined: Fri Jan 07, 2011 2:08 pm
- Real name and title:
-
- PLS Expert User
- Posts: 248
- Joined: Sat Jul 25, 2009 1:34 pm
- Real name and title:
Hey,
Why not, but you should pay attention to the necessary sample size.
Because the book is very expensive (about 300,- Euro) I would recommend you to borrow the handbook from the library.
For checking your results you should read following paper on pp. 298:
Henseler, J. / Ringle, C. M. / Sinkovics, R. R.: The use of partial least squares path modeling in international marketing, in: Sinkovics, R. R. / Ghauri, P. N. (eds.), Advances in International Marketing, Vol. 20, Bingley 2009, pp. 277-320.
http://php.portals.mbs.ac.uk/Portals/49 ... cs-PLS.pdf
Best regards,
Christian
Why not, but you should pay attention to the necessary sample size.
Because the book is very expensive (about 300,- Euro) I would recommend you to borrow the handbook from the library.
For checking your results you should read following paper on pp. 298:
Henseler, J. / Ringle, C. M. / Sinkovics, R. R.: The use of partial least squares path modeling in international marketing, in: Sinkovics, R. R. / Ghauri, P. N. (eds.), Advances in International Marketing, Vol. 20, Bingley 2009, pp. 277-320.
http://php.portals.mbs.ac.uk/Portals/49 ... cs-PLS.pdf
Best regards,
Christian
-
- PLS Senior User
- Posts: 28
- Joined: Thu Jun 24, 2010 5:38 am
- Real name and title:
Hi Christian,
I have the same problem with Larissa, when splitting the data into male=1 and female=2, I have 250 rows, and after splitting into that gender group, the smartPLS prompted that the case is not sufficient.
I have 69% of male and 31% of female, out of 250 responses
Could you please briefly describe step-by-step guide to split the data. Thanks
I have the same problem with Larissa, when splitting the data into male=1 and female=2, I have 250 rows, and after splitting into that gender group, the smartPLS prompted that the case is not sufficient.
I have 69% of male and 31% of female, out of 250 responses
Could you please briefly describe step-by-step guide to split the data. Thanks
-
- PLS Expert User
- Posts: 248
- Joined: Sat Jul 25, 2009 1:34 pm
- Real name and title:
Hey,
Your sample size should be enough for splitting your sample.
One problem, which often occurs is, that after splitting the moderator variable e.g. sex for multiple group tests, this variable should not longer be a part in your model. In this variable are only 1 or 0 and therefore you get an error message.
What also can be helpful is to use a other weighting sheme.
I hope that helps!
Christian
Your sample size should be enough for splitting your sample.
One problem, which often occurs is, that after splitting the moderator variable e.g. sex for multiple group tests, this variable should not longer be a part in your model. In this variable are only 1 or 0 and therefore you get an error message.
What also can be helpful is to use a other weighting sheme.
I hope that helps!
Christian
-
- PLS Senior User
- Posts: 28
- Joined: Thu Jun 24, 2010 5:38 am
- Real name and title:
-
- PLS Expert User
- Posts: 248
- Joined: Sat Jul 25, 2009 1:34 pm
- Real name and title:
The problem for Larissa was, that her model included sex as latent variable. If you split your sample according to sex (1=male; 0=female) then only 1 or 0 would be in the column for sex. In other words after splitting for sex only the number 1 would be in the column sex for the male group and only 0 would be in this column for female. If you have an item/column with always the same number you will get an error message (single matrix occure).
In very complex models it could also be the case that your sample size is to small. Then it could be helpful to use another weighting scheme.
Christian
In very complex models it could also be the case that your sample size is to small. Then it could be helpful to use another weighting scheme.
Christian
Ok, I got an similar problem. I hope someone could please help me.
But I'm still not sure, if I understand it right.
in my case I've got 4 soccer player to be scored in several influence factors on their Marketvalue. So 2 of them are "superstars" and 2 are less good. So now I need to do a group analysis. That means, that I now want to compare the superstar player with the less good player, to evaluate which factors influences in which group more or less.
So what I understand was to first of all split the one big sample into two (one with the starplayer and one with the less good ones). Thats correct so far, right? And now, what exactly is what I have to do? Just calculate them in two different models (each with PLS, Bootstrap, Blindfolding) or is something else to do to make a group comparisson?
Thanks so much in advance for any help
Regards
Angel
But I'm still not sure, if I understand it right.
in my case I've got 4 soccer player to be scored in several influence factors on their Marketvalue. So 2 of them are "superstars" and 2 are less good. So now I need to do a group analysis. That means, that I now want to compare the superstar player with the less good player, to evaluate which factors influences in which group more or less.
So what I understand was to first of all split the one big sample into two (one with the starplayer and one with the less good ones). Thats correct so far, right? And now, what exactly is what I have to do? Just calculate them in two different models (each with PLS, Bootstrap, Blindfolding) or is something else to do to make a group comparisson?
Thanks so much in advance for any help
Regards
Angel
-
- PLS Expert User
- Posts: 248
- Joined: Sat Jul 25, 2009 1:34 pm
- Real name and title:
Hey Angel,
Your proceeding seems alright for me.
1.) Split your sample according the group (star players and bad players)
2.) Calculate each group in a separate model in SmartPLS
3.) Compare the different path difference with the formal by Chin or Henseler
For point 3 you have to calculate the path coefficient of every sub model. You also need the bootstraps results for the comparison.
Best regards,
Chrsitian
Your proceeding seems alright for me.
1.) Split your sample according the group (star players and bad players)
2.) Calculate each group in a separate model in SmartPLS
3.) Compare the different path difference with the formal by Chin or Henseler
For point 3 you have to calculate the path coefficient of every sub model. You also need the bootstraps results for the comparison.
Best regards,
Chrsitian