Hi Dr Gaskin, I have a problem here, I had recently followed your method to create measurement model and after connecting the latent variable, supposingly the connected latent variables will turn to blue, right? But the color remain the same as bef5, didn't turn to blue for the connected latent variables. May I know why is it so?
Hi Dr. Gaskin Thanks for your videos, they are proving to be really helpful. Can you let me know why income is classified as ordinal? Further, could you explain your interpretation of ordinal vs metric data? For Likert scale data, you always opt for ordinal yet I see many other researchers leaving it at metric (even when using SmartPLS 4 which offers the option for metric, binary, and cat) and the model seems to fit and run fine? I'm assuming those opting for metric may assume that their ordinal scale satisfies quasi metric data in that the scale options has equal distance (as a 1-5 or 1-7 likert scale would). SmartPLS seems to allow this yet it must have an impact on results?
In this dataset, income is measured as ranges, rather than as an actual number. Ordinal implies approximately equal interval discrete values spanning a limited spectrum. Likert scales are a good example of this from strongly disagree to strongly agree. Metric implies a continuous value, such as a true (exact) value for income, or age, or golf score. In SmartPLS, the distinction does not change much except in edge cases.
Hi Prof. Gaskin, thank you for your informative videos. After importing my data, I am unable to select the type of scale for some of my variables; it shows up as "none." As a result, those variables are not displayed as part of the indicators for my model. Could you please explain what might be causing this issue?
Check whether they are scientific notation in Excel, rather than plain numeric values. Also check that there is a value in the top row for that variable.
I want to know "can we analyze panel data through SmartPLS?" I mean, if we have 5 years data (2015-2020) for 100 companies. Do you have video on this ?
Dr, I seem to have a problem with the importing file part. Everytime I tried to do it, only file type csv and txt can be import. I also cannot see file type Sav, xls, and xlsx. How do I overcome this problem?
Make sure you have the latest version of SmartPLS. If you are on a Mac, this might be creating the issue (though I'm not sure). To create a csv from an sav file, open it in SPSS and then export as csv.
Thank you James. suppose we have 3 main variables (independent variable), (dependent Variable) and (Mediating variable). each variable is measured by some already established and validated scale. and each scale has about 4 Dimensions with about 3 statements for each dimension. and all items were scored on a 7-point Likert scale 1=Strongly Disagree to 7=Strongly Agree. How would be the correct analysis process? Based on items level or based on dimension average means level ?
@@Gaskination Thanks. could you please clarify how to model it at the item level with higher order constructs? Additionally, how would be the process if the dependent variable is measured by a seven Item scale (7 dimensions with one Item for each dimension)? Thanks
@@anisalshargabi1244 Here is are videos on modeling higher order constructs: www.youtube.com/@Gaskination/search?query=higher%20order%20construct Sounds like the DV is a composite formative variable. Here is a video on that: ua-cam.com/video/Qx_7pN7eAes/v-deo.htmlsi=ym1kTXZBLxzFqZA3
Hi Professor, may I ask how can I speed up changing the properties of data as fast as you did in this vedio? I can only change it one by one click which takes so long time.
Hello and thanks for this useful tutorial. I was wondering if I could have the file that you imported in PLS. well I want to practice with real data and I don't have any database of survey. Thanks
Hello Sir, i have log transformed firm size for my research, that is an independent variable. Somehow i am unable to see it as an indicator in the smartpls 4 tool. The concerned data holds a range,, for eg. 100-499 that is log transformed. It is showing as text and i am unable to change the variable type. Can u please help. Thank you.
You can open the dataset in Excel and check the column to see if there is any issue with the numbers. You can change the data type in Excel to be numeric. You can also check to make sure it is a number, rather than scientific notation. If there are too many decimals, you can reduce them in Excel (there is a button for this) and then copy and paste (values only) into a new variable, then delete the old variable. Then save and reimport it into SmartPLS.
Observed is the actual, despite the scale size. So, if I have a five point likert scale, the scale min and max is 1 and 5, but the observed min and max depends on the lowest and highest values someone responded. If no one put 1, then that won't be the observed min.
You are correct. The best form to measure income is continuous (metric). Some measure it in a quasi-ordinal scale though: e.g., 10-20k, 21-50k, 51-90k, 91-150k, 150+k (not truly ordinal, but close)
That is strange. Try saving the csv as an xls or xlsx file. If that doesn't fix it, check whether there are any unreasonably long cell values (such as a paragraph or a decimal that never ends).
If you are referring to low AVE or low CR values, then you can try to improve them by dropping the lowest loading item. If several of the items load poorly, then you might consider whether you are using formative measurement instead of reflective.
Excellent.
love your work Sir.
Great Respect from Pakistan
Thanks for your time and teaching Sir.
Thank you Sir for this informative video. (Dr. Fahad)
Thank you Prof. for this incisive video
Thank you, Dr. Gaskin.
Extremely Helpful
Thank you Sir, great help
thank you, i can't find such excellent videos in China
Thanks Dr. James Gaskin
Thank you James
Thanks you Dr.
Thank very much!
thank you !
Thank youuu!
Hi Dr Gaskin, I have a problem here, I had recently followed your method to create measurement model and after connecting the latent variable, supposingly the connected latent variables will turn to blue, right? But the color remain the same as bef5, didn't turn to blue for the connected latent variables. May I know why is it so?
All latent variables must be connected and must have at least one linked measure.
@@Gaskination Thanks 🙏👍😊
Hi Dr. Gaskin
Thanks for your videos, they are proving to be really helpful. Can you let me know why income is classified as ordinal? Further, could you explain your interpretation of ordinal vs metric data?
For Likert scale data, you always opt for ordinal yet I see many other researchers leaving it at metric (even when using SmartPLS 4 which offers the option for metric, binary, and cat) and the model seems to fit and run fine? I'm assuming those opting for metric may assume that their ordinal scale satisfies quasi metric data in that the scale options has equal distance (as a 1-5 or 1-7 likert scale would). SmartPLS seems to allow this yet it must have an impact on results?
In this dataset, income is measured as ranges, rather than as an actual number. Ordinal implies approximately equal interval discrete values spanning a limited spectrum. Likert scales are a good example of this from strongly disagree to strongly agree. Metric implies a continuous value, such as a true (exact) value for income, or age, or golf score. In SmartPLS, the distinction does not change much except in edge cases.
Thanks
Hi Prof. Gaskin, thank you for your informative videos. After importing my data, I am unable to select the type of scale for some of my variables; it shows up as "none." As a result, those variables are not displayed as part of the indicators for my model. Could you please explain what might be causing this issue?
Check whether they are scientific notation in Excel, rather than plain numeric values. Also check that there is a value in the top row for that variable.
Hey James do we make both models I mean measurement model and structural model while performing CFA in smartpls4 Reply soon dear
Yes. In SmartPLS, the measurement and structural models are created simultaneously.
I want to know "can we analyze panel data through SmartPLS?" I mean, if we have 5 years data (2015-2020) for 100 companies. Do you have video on this ?
SmartPLS 4 does not include a feature for time series analysis.
Dr, I seem to have a problem with the importing file part. Everytime I tried to do it, only file type csv and txt can be import. I also cannot see file type Sav, xls, and xlsx. How do I overcome this problem?
Make sure you have the latest version of SmartPLS. If you are on a Mac, this might be creating the issue (though I'm not sure). To create a csv from an sav file, open it in SPSS and then export as csv.
Thank you James. suppose we have 3 main variables (independent variable), (dependent Variable) and (Mediating variable). each variable is measured by some already established and validated scale. and each scale has about 4 Dimensions with about 3 statements for each dimension. and all items were scored on a 7-point Likert scale 1=Strongly Disagree to 7=Strongly Agree. How would be the correct analysis process? Based on items level or based on dimension average means level ?
In SmartPLS, you can easily model it at the item level with higher order constructs.
@@Gaskination Thanks. could you please clarify how to model it at the item level with higher order constructs?
Additionally, how would be the process if the dependent variable is measured by a seven Item scale (7 dimensions with one Item for each dimension)?
Thanks
@@anisalshargabi1244 Here is are videos on modeling higher order constructs: www.youtube.com/@Gaskination/search?query=higher%20order%20construct
Sounds like the DV is a composite formative variable. Here is a video on that: ua-cam.com/video/Qx_7pN7eAes/v-deo.htmlsi=ym1kTXZBLxzFqZA3
Is it still allowed to download the result analysis in one single excel sheet?
Yes. From the Export menu. You can export to Excel or HTML.
Hi Professor, may I ask how can I speed up changing the properties of data as fast as you did in this vedio? I can only change it one by one click which takes so long time.
There should be a bulk changes option at the top
Thank you Professor@@Gaskination
How can I display t-values, p-values and indicator reliability ?
t-values and p-values for indicators and for structural paths are estimated when conducting a bootstrap.
Hello and thanks for this useful tutorial. I was wondering if I could have the file that you imported in PLS. well I want to practice with real data and I don't have any database of survey. Thanks
You can find several datasets on the homepage of statwiki.gaskination.com/
Hello Sir, i have log transformed firm size for my research, that is an independent variable. Somehow i am unable to see it as an indicator in the smartpls 4 tool. The concerned data holds a range,, for eg. 100-499 that is log transformed. It is showing as text and i am unable to change the variable type. Can u please help. Thank you.
You can open the dataset in Excel and check the column to see if there is any issue with the numbers. You can change the data type in Excel to be numeric. You can also check to make sure it is a number, rather than scientific notation. If there are too many decimals, you can reduce them in Excel (there is a button for this) and then copy and paste (values only) into a new variable, then delete the old variable. Then save and reimport it into SmartPLS.
my observed min and observed max is showing wrong when i import it i even tried to change the type, but no use. Any suggestions for this error?
Observed is the actual, despite the scale size. So, if I have a five point likert scale, the scale min and max is 1 and 5, but the observed min and max depends on the lowest and highest values someone responded. If no one put 1, then that won't be the observed min.
how can be the variable income ordinal? it's metric
You are correct. The best form to measure income is continuous (metric). Some measure it in a quasi-ordinal scale though: e.g., 10-20k, 21-50k, 51-90k, 91-150k, 150+k (not truly ordinal, but close)
Why I cannot import my Csv file ? It says unexpected error what should be the fix for it? thanks in andvance🙂
That is strange. Try saving the csv as an xls or xlsx file. If that doesn't fix it, check whether there are any unreasonably long cell values (such as a paragraph or a decimal that never ends).
If the result will not reliable. What can i do??
If you are referring to low AVE or low CR values, then you can try to improve them by dropping the lowest loading item. If several of the items load poorly, then you might consider whether you are using formative measurement instead of reflective.