Exploratory Factor Analysis (EFA): Concept, Key Terminologies, Assumptions, Running, Interpreting
Вставка
- Опубліковано 7 лип 2020
- This session explains in detail the concept of Exploratory Factor Analysis, The key assumptions of Exploratory Factor Analysis, and the rotation methods in Exploratory Factor Analysis. The video also highlights how to run EFA using Jamovi, how to interpret the results, and how to report EFA. Link to download Slides
businessresear...
To learn bow to use EFA in SPSS, watch
• Exploratory Factor Ana...
#ResearchCoach #ExploratoryFactorAnalysis #EFARotationMethods #Varimax #PrincipalComponentAnalysis #JAMOVI
Hi Dr. Latif, I know this video is more than 1 year old, but it is helping me now with my analysis. Thank you so much for posting this. Much appreciated.
Glad it was helpful!
If you are using SPSS, the following video may also be helpful.
ua-cam.com/video/Hag1xSdYce0/v-deo.html
Thank you sooooo much. This is the simplest explanation I have come across. Helped a tonne with my analysis.
Pleasure. Thanks for watching.
I have been following Fawad. HIs style is simple with understandable ppt. Thanks
Thanks. I am glad the sessions are helpful.
YOU ARE AMAZING MASHALLAH , IT IS VERYYYYY HELPFUL
Glad it was helpful!
amazing job, you are doing
Thank you
Very helpful video, as always...I have one query regarding interpreting and reporting of percentage of variance explained by each factor (when we are performing Principal Component Analysis using varimax rotation in SPSS) Should we report the extraction sum of squared loadings or rotated sum of squared loadings where variance is more evenly distributed among the factors or components??
Thanks. I am glad you liked it. Please report from Rotated Sum of Squared Loadings.
Sir plz help me, i have 5 factors but after doing efa i got 3 factors where eigen value is more than 1, what should i do? And is 0.75 eigen value acceptable as if i change the no of factors to 5 its all good but 2 factors are less than 1
There is no need for EFA if the scales used are already validated and adapted from existing research.
Amazing sir
Thanks. I am glad you liked it
Sir, how to resolve the cross loading problem. In the conceptual model, I have four factors in a construct (IV). When running EFA approx 80 percent of the items are loaded on more than two factors.
Please watch this video
ua-cam.com/video/mwO6st3GymE/v-deo.html
Thank you super video.
I have a question on the assumptions:
The assumption that the measurement errors are not correlated between them it is valid in reality?
For more details, this text may be helpful
www.bmj.com/content/313/7048/41#:~:text=Measurement%20error%20is%20the%20variation,quantity%20on%20the%20same%20individual.&text=In%20general%2C%20the%20correlation%20between,on%20the%20variability%20between%20subjects.
Amazing explanation thanks! however, the google drive link isn't working anymore. :(
Thanks. I will share the link to download slides soon.
Please visit following link to download
businessresearchcoach.wordpress.com/how-to-develop-and-validate-a-scale/
@@researchwithfawad Thank you so much. May Allah SWT bless you abundantly.
Sir, KMO is necessary when there is
EFA. For CFA, it is not necessary ?
For EFA yes, you need to assess measure of sampling adequacy.
Amazing video. Sir can you please run this test in spss also
Here it is
ua-cam.com/video/Hag1xSdYce0/v-deo.html
JazakAllah khairan sir. Trust me you are doing charity by making these videos. I was very worried . Shukuran Allah i came across your channel.
Thanks. I am glad you liked it
Sorry, I have both nominal and ordinal variable in my questionnaire, i took items from previous stuies and translate from english to french. Which one should i choose between EFA and CFA to check reliability and validity of my questionnaire? or should i use both of them? Thanks for your reply.
Hi.
If all of your items are from existing research and they are coming from different scales, plus you have added new dimensions, it is better to go for EFA. If the dimensions are already there and you have just added items from different sources, you can directly go for CFA.
As for having nominal scale variables, EFA will not be appropriate. As for likert scale, individual items in each scale are on ordinal scale but we can take them as numeric and perform EFA.
@@researchwithfawad Thanks for your reply. So finally in my case i can only do EFA if i am right, as you mention the problem of having also nominal variables
Sir, during the analysis of variables I observed that principle component analysis is giving higher values in factor loadings and in total variance explained. Could you please explain it?
Higher values in comparison to?
@@researchwithfawad Sir, in comparison to "maximum likelihood".
AOA, Respected Sir, EFA of 79 line items scale shows their are 15 factors but all values are loaded in first column. i dont know why.... it means there is only one dimension of scale ? but i does not seem logical. kindly guide me how to handle this.... Jazak Allah
WAS. You will need to look at the Rotated Component Matrix. Select Varimax as Factor Rotation. Please watch this video, focus on the time when i select the options.
ua-cam.com/video/Hag1xSdYce0/v-deo.html
Hi sir! Thank you so much for this wonderful explanation. I am asking for the word templates that you used. I did not found it on the sides. If possible can you share it?
You can email me in kh.fawad83@gmail.com.
I will share the paper for reference
Thank you very much for this video prof. regarding KMO & Bartlett’s Values, are they need to be run under a pilot test? also, I'm a little confused regarding Varimax, some theses use it and some are not, could you please let me know the criteria for using Varimax? thank you so much.
Thanks. I am glad you liked it. No need to go for EFA in pilot study.
Please watch this video, for rotation.
ua-cam.com/video/Hag1xSdYce0/v-deo.html
@@researchwithfawad Thank you so much
Thank you Dr. Latif for a very nice presentation. While analyzing my data I had several of the factors with KMO value less than 0.5. Should I remove those factors?
I am glad it was helpful. Are you doing separate EFA analysis for different factors? If they are sub-dimensions for a single construct, perform single EFA with all items and see how many factors you can extract. Less than .50 is surely low. Improve your sample size.
Sir, Kindly suggest... In my research model, there are five main constructs ie. A,B,C,D,E. A is having four dimensions and each dimension has 4-5 items. B is also having four dimensions in which the individual dimension is having four subdimensions, and each subdimension is having 3-4 items. C,D,E are having 6-7 items. So should EFA is to be performed for each of the five constructs or for full model?
Also please suggest.. if Construct D is adopted and all others are adapted with the inclusion of new dimensions, how can we explain it for EFA and CFA.
Sir I want to ask. If my thesis proposal is ready, is the first step in this efa distributing the questionnaire to test its validity? Or how?
EFA is required if you are developing yoir own instrument. If you are using a scale that has been develop and validated by others, no need to gonfor EFA.
Sir, is it necessary to run EFA if my scale is a developed scale but I made some changes in the wording of questionnaire. 🙏🙏🙏
No need to run EFA just for changes in wording.
@@researchwithfawad thank you so much sir... God bless you🙏🙏🙏
Hello sir, Do we apply EFA and CFA in case of Formative scale ? I read somewhere we apply only for reflective scale.
EFA is not required for Formative Constructs.
@@researchwithfawad Thank you sir.
@@researchwithfawad Thank you sir for the reply. Do we still require a CFA ? If so then can it be done using SmartPLS?
CFA is for CB-SEM, in SmartPLS it's Confirmatory Composite Analysis CCA
Hi, can we merge two models, means taking some dimensions from one and some from other. Or it is mandatory to use one model only.
Yes, if you are developing a new scale.
@@researchwithfawad thanku so much. Is there any reference for this
For instance, see this paper where i have used multiple sources to develop a scale.
Latif, K. F. (2018). The development and validation of stakeholder-based scale for measuring university social responsibility (USR). Social Indicators Research, 140(2), 511-547.
professor is Eigenvalue 1.06 is acceptable
Yes, its greater than 1
@@researchwithfawad thankyou so much professor 💜
can you clarify why not use EFA in jamovi but PCA instead? thank you
Factor Analysis is considered different to PCA in JAMOVI. While SPSS adds PCA as part of factor analysis in SPSS. Since, the objective is to limit the large number of items into smaller factors. Hence, it was used. It is better you review difference between PCA and factor analysis.
@@researchwithfawad Got it. Thank you so much!
What if in EFA the item do not come under the desired construct.
Plz suggest
It is recommended that the item be deleted.