Here's a fun pet project I've been working on: udreamed.com/. It is a dream analytics app. Here is the RUclips channel where we post a new video almost three times per week: ruclips.net/channel/UCiujxblFduQz8V4xHjMzyzQ Also available on iOS: apps.apple.com/us/app/udreamed/id1054428074 And Android: play.google.com/store/apps/details?id=com.unconsciouscognitioninc.unconsciouscognition&hl=en Check it out! Thanks!
Dear Prof. Gaskin, thank you again for your great videos. In Information Systems, I see increasing skepticism towards PLS-based SEM among senior editors. The Journal of Management Information Systems (JMIS) or the International Journal of Electronic Commerce (IJEC) strictly do not review PLS studies anymore. Do you understand why?
There is a long debate on this and whether a journal accepts such studies depends entirely on the EIC. The two main arguments against PLS (currently) are that the lack of fit statistics allows for dubious and incomplete theories to be "supported" while alternative untested models may be more valid, and, formative constructs are not valid at the first-order level due to the composition of items being an incomplete and inaccurate representation of a construct. While both of these arguments are valid, they also can be overcome through due-diligence in PLS as thoroughly explained in A Primer on PLS (book).
Hello! I have a problem on the step ''Double click to import data''. When i click, it doesn't show the window to choose a file and to import my data. How can i resolve that?
Hello dear I can't calculate degree of freedom I saw some notes on smartpls website but I don't know what do you mean L in this equation. These notes on smart pls website about degree of Freedoom Assuming a multinormal distribution, the Chi² value of a PLS path model with df degrees of freedom approximately is (N-1)*L, whereby N is the number of observations and L the maximum likelihood function as defined by Lohmöller (1989). The degrees of freedom (df) are defined as (K² + K) /2 - t, whereby is the number of manifest variables in the PLS path model and t the number of independent variables to estimate the model implied covariance matrix. However, future research must clearly define how to determine the degrees of freedom of composite models, common factor models, and mixed models when using PLS-SEM.
Degrees of freedom are not calculated in SmartPLS because they are based on the covariance matrix (number of parameters not estimated subtracted from total number of possible parameters)
Dear Professor, Please could you let me know if this procedure of analysis is right for Disjoint 2 stage approach : 1. One of my predictor variables is an identity scale and authors have outlined different dimensions of identity saying that average or total score is not recommended. My understanding is that these should be treated as separate scales entirely (I've seen papers using them that way). The 3 variables that are to be predicted using these identity scales are established higher order constructs with differing lower order items. Hence I used Disjoint approach. In stage 1, I established the measurement model (CR,AVE,HTMT 0.90 criterion) and got the LVS. 2. My confusion is in stage 2: Can I use the LVS of the identity scales to model the separate identity scales with 1 item (the LVS obtained from the previous stage)? That means there will be 1 item with a loading of 1.000 for the identity scales in the 2nd stage. Is that the right way to go? My objective is to look at the structural relations using identity, rather than the sub-dimensional impact of the identity scales. Looking forward to your response. Regards, Navya
Hi professor! When i imported my data from a CSV file, it keeps showing scales as metric for some and they're supposed to be binary. There's no option to change the scale to binary. How can I solve this issue?
Thank you! I actually have another question regarding simple slope analysis. If my data is binary, how would I know if the effect of a variable on another is postive or negative as it does not show any mean on the chart. I'm fairly new to this, so I would really appreciate any help. Thank you!
@@lalitdulal6006 Look at the +1 SD for the slope of a binary predictor or binary outcome. The +1 indicates the presence of the value, and the -1 indicates the absence of the value. Or you could run a multigroup analysis instead.
Thanks a lot Prof. James, When i am importing the file, lot of lines of data appear in the 'indicator' header stating 'line0: indicator 'specify' is not unique' and many such lines below Also, when I select 'active data file' there is nothing which shows on the left tab of yellow 'indicator' Where do you think, is the error coming from.?
Hi Piu N, regarding the "line 0: indicator...." problem: I had the same problem, because some of my column headers carried the same names or words, which were repeated in another header (e.g. when you subquestions in your survey they often carry the same coding). I just changed the names of all headers that were not unique and then it worked. I hope that helps!
@@konstantinsteiner1991 thanks a lot thanks @Konstantin Steiner.. i tried changing all variable values to numericals.. it didnt work, changed the header to simplified headers like AOC1, AOC2.. it didnt work, [now if i understand you correctly, even AOC1 &AOC2 Will clash, right?] someone recently told me that it could be that i dont have serial no .. &Another one said that delimited needs to be specified, I havent tried these two, so am not sure.. so 1. how do i keep the headers different coz they do have some words in the question which are similar. eg I do not feel a strong sense of "belonging" to my organization. I do not feel "emotionally attached" to this organization. both have common words in the headers 2.& then how do i refresh this so that it reflects in the PLS indicator correctly? possible to get your email ID?
@@piun6705 The variable names just have to be unique. So, AOC1 and AOC2 are fine. If you are still getting this error, make sure you're not violating another assumption, such as long variable names, scientific notation due to long decimal values, or variable names with spaces in them (or mathematical symbols).
Hello, I was watching most video smartpls and have a question. How I read clearly interpret data from smartpls? How I know which factors (e.g. VIF) meet enough conditional on analysis data? What does articles or books must read? Any advice from you. Tks
That can happen if you have a variable that does not have any variance (e.g., you included gender, but all respondents were male). It can also happen if using variables of vastly different scale sizes (e.g., 3 point scale and then income) on the same factor. It can also happen if you've include multinomial variables as indicators of a factor. These should be split up into binary dummy variables.
Dear Professor , I need your help I have downloaded smart pls and have availed 30 days free professional trial which is now expired. Now when I am again changing the license with new e-mails it is giving me license error . What should I do now?
Hello, sir, your video demonstration is awesome. Sir, can you plz guide me on how to get this PLS software without license? Moreover, sir when I put my data in the PSL it continues to show that continuing running no results are shown what should I do?
I cannot provide you an illegal copy of the software. As for your results not showing up, I would have to know more about the model and the data to better understand the issue.
@@Gaskination yes sir I understand I don't wanna illegal copy, sir I have to analyze my thesis with the mediating variable I just plot a diagram just like you on pls but when I used free student base it doesn't run and when I chose to free trial it also not work so what can I do?? Please guide me I have to draw measurement model and structure model with analysis
@@studyandentertainmentwitht2004 I think the student version is limited by the number of variables and the number of sample size. But the trial version should be full access. So, in this case, make sure your dataset is valid (green) and make sure the latent factors all turn blue. If that is all good, then make sure you do PLS, not PLS consistent.
Thank you for this video, I have a question,,, i uploaded data file, but few parameter are not visible in "INDICATORS", section.. can you please explain Sir.
Hi James it looks like SPSS has a PLS add-on, does it do the same job as SmartPLS (in principle, regardless the graphic)? Would you be interested to do a video for SPSS pls please?
While importing data file from excel with '.csv' format. Data has been imported but on indicator canvas showing 'No indicator to show' even I got green signal after importing data
same as for ordinal variables. If the moderator is continuous, you can either split it into low/high and use it as a grouping variable, or you can use it in an interaction.
I'm not a licensed distributor for SmartPLS. You should be able to find the license at their website: SmartPLS.com I know there is also a trial version.
Brahim, I would recommend looking on Google for a better explanation... Basically, a formative factor is made up of measures that do not co-vary. One example would be health, made up of... diet, exercise, non-smoking, etc. These measures do not co-vary and they come together to form the formative factor of health. However, a reflective factor is a factor made up of measures that are essentially the same question again and again. These measures DO co-vary. So, for diet you could measure if people eat vegetables, if they eat junk food, if they drink sugar-filled drinks, etc. This is a reflective factor. Hope that helps.
Hey Professor! Thanks so much for the video, i found it very helpful. I really appreciate your job. Sir, I've some queries related to my research work so i wanna keep in touch with you regarding your assistance. Would you like to share your email with me? it will be very grateful. thanks.
Here's a fun pet project I've been working on: udreamed.com/. It is a dream analytics app. Here is the RUclips channel where we post a new video almost three times per week: ruclips.net/channel/UCiujxblFduQz8V4xHjMzyzQ
Also available on iOS: apps.apple.com/us/app/udreamed/id1054428074
And Android: play.google.com/store/apps/details?id=com.unconsciouscognitioninc.unconsciouscognition&hl=en
Check it out! Thanks!
You are the one , the world needs you
Thank you!!!! Very much!
Dear Prof. Gaskin, thank you again for your great videos. In Information Systems, I see increasing skepticism towards PLS-based SEM among senior editors. The Journal of Management Information Systems (JMIS) or the International Journal of Electronic Commerce (IJEC) strictly do not review PLS studies anymore. Do you understand why?
There is a long debate on this and whether a journal accepts such studies depends entirely on the EIC. The two main arguments against PLS (currently) are that the lack of fit statistics allows for dubious and incomplete theories to be "supported" while alternative untested models may be more valid, and, formative constructs are not valid at the first-order level due to the composition of items being an incomplete and inaccurate representation of a construct. While both of these arguments are valid, they also can be overcome through due-diligence in PLS as thoroughly explained in A Primer on PLS (book).
@@Gaskination thank you! :)
thank you, that is very useful
Thank you so much!
Hello! I have a problem on the step ''Double click to import data''. When i click, it doesn't show the window to choose a file and to import my data. How can i resolve that?
I'm not sure why that would happen... I would recommend reaching out to the SmartPLS support folks.
Thank you
Thanks Prof.
Prof. Gaskin is it possible to use dependent variable as dichotomous variable with smart PLS? does smart PLS works only on Likert scale?
Hello dear
I can't calculate degree of freedom
I saw some notes on smartpls website but I don't know what do you mean L in this equation.
These notes on smart pls website about degree of Freedoom
Assuming a multinormal distribution, the Chi² value of a PLS path model with df degrees of freedom approximately is (N-1)*L, whereby N is the number of observations and L the maximum likelihood function as defined by Lohmöller (1989). The degrees of freedom (df) are defined as (K² + K) /2 - t, whereby is the number of manifest variables in the PLS path model and t the number of independent variables to estimate the model implied covariance matrix. However, future research must clearly define how to determine the degrees of freedom of composite models, common factor models, and mixed models when using PLS-SEM.
Degrees of freedom are not calculated in SmartPLS because they are based on the covariance matrix (number of parameters not estimated subtracted from total number of possible parameters)
got data file for pls sem and able to do stuff
thanks a lot
Dear Professor,
Please could you let me know if this procedure of analysis is right for Disjoint 2 stage approach :
1. One of my predictor variables is an identity scale and authors have outlined different dimensions of identity saying that average or total score is not recommended. My understanding is that these should be treated as separate scales entirely (I've seen papers using them that way). The 3 variables that are to be predicted using these identity scales are established higher order constructs with differing lower order items. Hence I used Disjoint approach. In stage 1, I established the measurement model (CR,AVE,HTMT 0.90 criterion) and got the LVS.
2. My confusion is in stage 2: Can I use the LVS of the identity scales to model the separate identity scales with 1 item (the LVS obtained from the previous stage)? That means there will be 1 item with a loading of 1.000 for the identity scales in the 2nd stage. Is that the right way to go? My objective is to look at the structural relations using identity, rather than the sub-dimensional impact of the identity scales.
Looking forward to your response.
Regards,
Navya
I'm not sure. I've never done a disjoint analysis. So, I have no reference or experience with it. Sorry about that. Best of luck to you.
@@Gaskination I understand, Professor. Thanks a ton, I've learnt a lot from your videos. Have a great day :)
GOOD
Hello. How can we import and build models with categorical vairables (YES/NO) ??? are there any tips?
Only if they are binary, or split into dummy variables.
Dear Prof Gaskin,
Can I use SMART-PLS for normal data with huge samples (300 subjects)?
Definitely. 300 is a good size. Huge would be something greater than 800.
Hi professor! When i imported my data from a CSV file, it keeps showing scales as metric for some and they're supposed to be binary. There's no option to change the scale to binary. How can I solve this issue?
Don't worry. This won't actually change the analysis or estimates. So, continue without worrying about this adjustment.
Thank you! I actually have another question regarding simple slope analysis. If my data is binary, how would I know if the effect of a variable on another is postive or negative as it does not show any mean on the chart. I'm fairly new to this, so I would really appreciate any help. Thank you!
@@lalitdulal6006 Look at the +1 SD for the slope of a binary predictor or binary outcome. The +1 indicates the presence of the value, and the -1 indicates the absence of the value. Or you could run a multigroup analysis instead.
Thanks a lot Prof. James,
When i am importing the file, lot of lines of data appear in the 'indicator' header stating 'line0: indicator 'specify' is not unique' and many such lines below
Also, when I select 'active data file' there is nothing which shows on the left tab of yellow 'indicator'
Where do you think, is the error coming from.?
Hi Piu N, regarding the "line 0: indicator...." problem: I had the same problem, because some of my column headers carried the same names or words, which were repeated in another header (e.g. when you subquestions in your survey they often carry the same coding). I just changed the names of all headers that were not unique and then it worked. I hope that helps!
@@konstantinsteiner1991 thanks a lot
thanks @Konstantin Steiner.. i tried changing all variable values to numericals.. it didnt work, changed the header to simplified headers like AOC1, AOC2.. it didnt work, [now if i understand you correctly, even AOC1 &AOC2 Will clash, right?]
someone recently told me that it could be that i dont have serial no .. &Another one said that delimited needs to be specified, I havent tried these two, so am not sure..
so
1. how do i keep the headers different coz they do have some words in the question which are similar. eg
I do not feel a strong sense of "belonging" to my organization.
I do not feel "emotionally attached" to this organization.
both have common words in the headers
2.& then how do i refresh this so that it reflects in the PLS indicator correctly?
possible to get your email ID?
@@piun6705 The variable names just have to be unique. So, AOC1 and AOC2 are fine. If you are still getting this error, make sure you're not violating another assumption, such as long variable names, scientific notation due to long decimal values, or variable names with spaces in them (or mathematical symbols).
Prof. Gaskin, do you have a video for making graphic of moderation effect in SmartPLS3
Yes. here it is: ruclips.net/video/PnPfOGtl-lc/видео.html
Hello, I was watching most video smartpls and have a question. How I read clearly interpret data from smartpls? How I know which factors (e.g. VIF) meet enough conditional on analysis data? What does articles or books must read? Any advice from you. Tks
The best book on SmartPLS is the one by Joe Hair: A primer on partial least squares
Prof. I tried to calculate using the PLS algo, but there's an error of Single Matrix. How can I solve that ? Thank you in advance.
That can happen if you have a variable that does not have any variance (e.g., you included gender, but all respondents were male). It can also happen if using variables of vastly different scale sizes (e.g., 3 point scale and then income) on the same factor. It can also happen if you've include multinomial variables as indicators of a factor. These should be split up into binary dummy variables.
Waaw so easy to learn can you please help me i want to learn more
Here is a whole playlist: ruclips.net/p/PLnMJlbz3sefKTL7KGy_JIYTSpFXizxW1X
Dear Professor , I need your help
I have downloaded smart pls and have availed 30 days free professional trial which is now expired. Now when I am again changing the license with new e-mails it is giving me license error .
What should I do now?
I'm not sure. This is something you'll need to contact SmartPLS support for: www.smartpls.com/support
@@Gaskination thanks for your reply
Hello, sir, your video demonstration is awesome. Sir, can you plz guide me on how to get this PLS software without license? Moreover, sir when I put my data in the PSL it continues to show that continuing running no results are shown what should I do?
I cannot provide you an illegal copy of the software. As for your results not showing up, I would have to know more about the model and the data to better understand the issue.
@@Gaskination yes sir I understand I don't wanna illegal copy, sir I have to analyze my thesis with the mediating variable I just plot a diagram just like you on pls but when I used free student base it doesn't run and when I chose to free trial it also not work so what can I do?? Please guide me I have to draw measurement model and structure model with analysis
@@studyandentertainmentwitht2004 I think the student version is limited by the number of variables and the number of sample size. But the trial version should be full access. So, in this case, make sure your dataset is valid (green) and make sure the latent factors all turn blue. If that is all good, then make sure you do PLS, not PLS consistent.
@@Gaskination thanks sir
Thank you for this video, I have a question,,, i uploaded data file, but few parameter are not visible in "INDICATORS", section.. can you please explain Sir.
You mean something like "mean" or "max" is not there? I'm not sure what would cause that. Make sure to check your data for completeness.
@@Gaskination can you please tell me your email ?
@@fahadsardar765 google me and you'll find it behind a captcha
Hi James it looks like SPSS has a PLS add-on, does it do the same job as SmartPLS (in principle, regardless the graphic)? Would you be interested to do a video for SPSS pls please?
I think it does PLS regression, but not modeling (like with mediators and moderators). I think it also doesn't accommodate latent factors.
@@Gaskination yes right
While importing data file from excel with '.csv' format. Data has been imported but on indicator canvas showing 'No indicator to show' even I got green signal after importing data
That is odd. I'm not sure why that would happen unless your variables are all string values (like words instead of numbers).
You can click the diagram of your project first, the indicators will appear after that.
Hi, I am having the same problem. can you suggest specifically?
@@citraang I had the same problem! Two years later, it´s still the solution. Thank you.
how to save in notepad ?
Would it be possible to have the database with which you work in the example? It would be useful to practice while I watch the video. Thanks
Here you go: drive.google.com/open?id=1k_ylcLKDicbxydC6iVDljbvGe6BRx5z-
@@Gaskination Thank you Professor. You are the best!
i am trying to import data through a csv file. but it says that too many rows. can you please help
hmmm.... I wonder if you have a student or trial version with a limited number of rows. Try reducing it to 99 rows to see if that is the issue.
Hi james , Do you have any videos about how to do moderated mediation for continuous variables in SmartPLS3
same as for ordinal variables. If the moderator is continuous, you can either split it into low/high and use it as a grouping variable, or you can use it in an interaction.
Which one is the best version in smartpls 2, 3,3.0 I am confused
I don't understand your question. Version of what?
3
Thanks Prof... So, why my data any missing when I import to new project?
Sorry, I don't understand your question.
Sir can I get Smart PLS key for 3 version please, I want to do analysis of my thesis with the help of this please
I'm not a licensed distributor for SmartPLS. You should be able to find the license at their website: SmartPLS.com I know there is also a trial version.
@@Gaskination thank you so much 😊
Hii james..how to import model and data from the backup file that previously had been tested
You should be able to just go to 'file', 'import project from backup file', then navigate to the backup file (of type .splsp or .zip) to open.
@@Gaskination oh...thank you so much😊
please give me the tutorial for putting the collected in excel
Most people just use a survey service like Qualtrics. This will automatically enter survey responses into a spreadsheet for you.
Hello james, plaese what do you mean by formative factors
Brahim, I would recommend looking on Google for a better explanation...
Basically, a formative factor is made up of measures that do not co-vary. One example would be health, made up of... diet, exercise, non-smoking, etc. These measures do not co-vary and they come together to form the formative factor of health.
However, a reflective factor is a factor made up of measures that are essentially the same question again and again. These measures DO co-vary. So, for diet you could measure if people eat vegetables, if they eat junk food, if they drink sugar-filled drinks, etc. This is a reflective factor. Hope that helps.
thank you very much Sam
what about a marketing mix (price product place and promotion) is it a formative or reflective factor? thanks!
how to change data external ????
You can add as many datasources as you would like. Each project can have multiple data sources.
Hey Professor! Thanks so much for the video, i found it very helpful. I really appreciate your job. Sir, I've some queries related to my research work so i wanna keep in touch with you regarding your assistance. Would you like to share your email with me? it will be very grateful. thanks.
if you google me, you'll find my email. It is on my profile after you prove you're not a robot.