I cannot thank you enough for this video! I have struggled with statistics and data analysis since I started uni, and since I've been writing up my results section for my dissertation, I've been having a non-stop panic. This video has been incredibly informative, helping me to understand not just what we're doing, but also why we're doing it, which is so important for embedding that understanding. The examples throughout were great, and really helped me to consolidate what I was learning. Dare I say I'm actually looking forward to analysing my data now! Big thanks!
More than just comprehensive and informative, this is an ASMR video for statistics. Speaking clearly, with empathy for the learner, in a soothing tone that helps you stay interested in a subject that is well, DRY. The upbeat, down to earth explanation helped me get through it. I needed to watch twice as there is a lot to learn…and I think I got it. Thank you, I can run EFA in JASP now.
Thank you so so much for this video. You have no idea how helpful and actually inspiring this lecture was for my own research! I am eternally grateful. I will thank you in my dissertation!
Hi there! For around 43.30 where you do the EFA... I have problems moving them into the 'Included Variable'. I can not move them there. Are there anything where I have done wrong? I did follow your steps carefully though. :(
Two ideas: make sure they are showing as "rulers" because the bars/circles are not allowed in that box. Second would be just to close jasp and try again?
Is there a way to generate new variables for factor 1, 2 and 3 (like SPSS) ? Or is it done manually? There is an option in SPSS in the dimension reduction box that allows users to 'save as variables'. I wonder if we can do something similar with JASP. Thanks
That depends on which version of excel you have, but here's a guide: support.office.com/en-us/article/use-the-analysis-toolpak-to-perform-complex-data-analysis-6c67ccf0-f4a9-487c-8dec-bdb5a2cefab6
Really nice tutorial and very understandable! The best I have found so far. I am only just starting out with EFA for developing a questionnaire. I noticed that the word file that accompanies this tutorial is not available on your personal website or through the link. Can I still find it somewhere? One minor correction: JASP does not seem to use Maximum Likelihood, but "MinRes" to estimate EFA. This is something I found in a forum though so I am not 100% sure if it's correct, but it seems to make sense based on the function description from the link underneath and that the psych package of which fa is a function is listed on JASP's website for packages that JASP depends on :). This might help some people explain why they get different results from SPSS EFA using ML versus EFA in JASP. www.rdocumentation.org/packages/psych/versions/1.8.12/topics/fa
I only recently realized that minres is the new default for psych, so that is correct. The files are now at github.com/doomlab/statsofdoom-files/tree/master/advanced/JASP - we are working on updating the links over the summer.
Great explanation and showing how to clean up the data. i have a question. at 21:30 you say that participants can miss 1-2 item and you can still fill them in. Is there a term for that? total x .05 = adjustable amount.
Hi Dr. Buchanan! Thanks for this great introduction to EFA. I am writing a paper for my experimental psych class and I wanted to cite your Github repository (in addition to Preacher and MacCallum). Is there a way you recommend I do this?
Sure and thank you! Maybe something like Buchanan, E.M. (2020). Statistics of DOOM. Retrieved from: github.com/doomlab/statsofdoom-files. I mostly cite these things as websites!
At lot of the material is from Preacher and MacCallum's repairing tom swift, but I also wrote a chapter for sage on this: www.aggieerin.com/pubs/buchanan%20sage%2014.pdf
Excelent and valuable overview of EFA and how to do it with JASP. However I am not so keen on some procedures you performed, mainly on regresion of total values (across all items) on RND variable. Firstly, you will never work with total accross all items, because in the first place you are retaining two or three factors, or scales so the total in advance is meaningless. As it is residuals on radnom variable.
I pulled that screening procedure from the Tabachnick + Fidell, and more on that set of data screening steps can be found in some of the data screening videos. Mainly it's a procedure to examine the variables you will use in the analysis, not just the ones you'd retain in the EFA.
Statistics of DOOM It is certainly desirable to screen each manifest variable before conducting analysis. But I do not see a point in making composite variable and regressing it to the random generated variable. What is use of it, screening for what?
I'd actually recommend screening all the variables regressed to the random variable but in Excel that can be excruciating to run - so, for this class assignment, I told them to do a composite variable. The data screening videos cover the whys and ideas of how screening work more.
I cannot thank you enough for this video! I have struggled with statistics and data analysis since I started uni, and since I've been writing up my results section for my dissertation, I've been having a non-stop panic. This video has been incredibly informative, helping me to understand not just what we're doing, but also why we're doing it, which is so important for embedding that understanding. The examples throughout were great, and really helped me to consolidate what I was learning. Dare I say I'm actually looking forward to analysing my data now! Big thanks!
Thanks for the kind words!
More than just comprehensive and informative, this is an ASMR video for statistics. Speaking clearly, with empathy for the learner, in a soothing tone that helps you stay interested in a subject that is well, DRY. The upbeat, down to earth explanation helped me get through it. I needed to watch twice as there is a lot to learn…and I think I got it. Thank you, I can run EFA in JASP now.
Your comment gave me a good laugh - glad to be of help!
Thank you so so much for this video. You have no idea how helpful and actually inspiring this lecture was for my own research! I am eternally grateful. I will thank you in my dissertation!
Glad to help!
From one lecturer to another - love your approach...and landing :)
Thank you!
That was an amazing lecture.. Please come up with more samples. Thank you very much!
Great lecture. Thank you Dr. Doom! :-)
Thank you!
Hi there! For around 43.30 where you do the EFA... I have problems moving them into the 'Included Variable'. I can not move them there. Are there anything where I have done wrong? I did follow your steps carefully though. :(
Two ideas: make sure they are showing as "rulers" because the bars/circles are not allowed in that box. Second would be just to close jasp and try again?
Is there a way to generate new variables for factor 1, 2 and 3 (like SPSS) ? Or is it done manually? There is an option in SPSS in the dimension reduction box that allows users to 'save as variables'. I wonder if we can do something similar with JASP. Thanks
Yes! It’s in a newer version than this video: jasp-stats.org/2018/12/18/how-to-compute-columns-in-jasp/
If Pca is caused by questions, would You give us an example?
How do I get my excel to display the "data analysis" button?
That depends on which version of excel you have, but here's a guide: support.office.com/en-us/article/use-the-analysis-toolpak-to-perform-complex-data-analysis-6c67ccf0-f4a9-487c-8dec-bdb5a2cefab6
Really nice tutorial and very understandable! The best I have found so far. I am only just starting out with EFA for developing a questionnaire.
I noticed that the word file that accompanies this tutorial is not available on your personal website or through the link. Can I still find it somewhere?
One minor correction: JASP does not seem to use Maximum Likelihood, but "MinRes" to estimate EFA. This is something I found in a forum though so I am not 100% sure if it's correct, but it seems to make sense based on the function description from the link underneath and that the psych package of which fa is a function is listed on JASP's website for packages that JASP depends on :). This might help some people explain why they get different results from SPSS EFA using ML versus EFA in JASP.
www.rdocumentation.org/packages/psych/versions/1.8.12/topics/fa
I only recently realized that minres is the new default for psych, so that is correct.
The files are now at github.com/doomlab/statsofdoom-files/tree/master/advanced/JASP - we are working on updating the links over the summer.
@@StatisticsofDOOM Thank you! That was really useful!
Great explanation and showing how to clean up the data. i have a question. at 21:30 you say that participants can miss 1-2 item and you can still fill them in. Is there a term for that? total x .05 = adjustable amount.
Not really - but a good citation would be Tabachnick and Fidell (2012).
thanks for uploading this video.. are you explaining Uniqueness in your video?
You'll have to be more specific on where or what you are talking about re: uniqueness.
Hi Dr. Buchanan! Thanks for this great introduction to EFA. I am writing a paper for my experimental psych class and I wanted to cite your Github repository (in addition to Preacher and MacCallum). Is there a way you recommend I do this?
Sure and thank you! Maybe something like Buchanan, E.M. (2020). Statistics of DOOM. Retrieved from: github.com/doomlab/statsofdoom-files. I mostly cite these things as websites!
Clear, do you have SPSS application for the same problem?
Not sure I do since I've switched to free software. The same concepts apply!
I can not move the variables into the box of variables. Can you please help to advice me how to solve this problem?
It's probably the wrong "type" - click on the circles/ruler icon next to the variable name and change it to the appropriate type.
Thank you for the video! How do we know the cutoff score is 3?
You are using an outlier criterion of p
Very informative video.Is there any reference for the materials discussed here..I want to cite this for my thesis work..Please do let me know
At lot of the material is from Preacher and MacCallum's repairing tom swift, but I also wrote a chapter for sage on this: www.aggieerin.com/pubs/buchanan%20sage%2014.pdf
Excelent and valuable overview of EFA and how to do it with JASP. However I am not so keen on some procedures you performed, mainly on regresion of total values (across all items) on RND variable. Firstly, you will never work with total accross all items, because in the first place you are retaining two or three factors, or scales so the total in advance is meaningless. As it is residuals on radnom variable.
I pulled that screening procedure from the Tabachnick + Fidell, and more on that set of data screening steps can be found in some of the data screening videos. Mainly it's a procedure to examine the variables you will use in the analysis, not just the ones you'd retain in the EFA.
Statistics of DOOM
It is certainly desirable to screen each manifest variable before conducting analysis. But I do not see a point in making composite variable and regressing it to the random generated variable. What is use of it, screening for what?
I'd actually recommend screening all the variables regressed to the random variable but in Excel that can be excruciating to run - so, for this class assignment, I told them to do a composite variable. The data screening videos cover the whys and ideas of how screening work more.
why 2.99 when you were doing the outlier
Can you explain your question better? I'm not sure I'm following? Maybe because the cut off score is 3, so filter for greater than 2.99?