Where to apply EFA (Exploratory Factor Analysis) and CFA (confirmatory factor analysis)?

Поділитися
Вставка
  • Опубліковано 9 січ 2025

КОМЕНТАРІ • 64

  • @the360degreeknowledge9
    @the360degreeknowledge9 2 роки тому +1

    Excellent sir one of the best i found in research methodology

  • @TheKothay1
    @TheKothay1 3 роки тому +1

    Wonderfully explained. Many confusions cleared. Many thanks

  • @atulhadap2176
    @atulhadap2176 2 роки тому

    Super explanation. You make it simple

  • @anuthomas3337
    @anuthomas3337 2 роки тому

    Excellent sir.. Well explained 👍

  • @subhasharora9718
    @subhasharora9718 3 роки тому +1

    You explained it very well. Thank you Sir.

    • @adilzia
      @adilzia  3 роки тому +1

      That's great. Keep watching and keep sharing

  • @sunshine-ik2vm
    @sunshine-ik2vm 3 роки тому +1

    very helpful video. thank you.

  • @dr.seemakhaneducation
    @dr.seemakhaneducation 3 роки тому +1

    Very well explained 👍

  • @Noor-dq6rl
    @Noor-dq6rl 2 роки тому +1

    very helpful thanks

  • @alara1684
    @alara1684 3 роки тому +1

    Sir, if I'm using two scales to measure a construct, both having different items, am I supposed to do EFA first

  • @pradeepkumarbhoi9323
    @pradeepkumarbhoi9323 3 роки тому

    Well explained

  • @amitavenkatesh5946
    @amitavenkatesh5946 Рік тому

    Sir if we are taking the variables from previous studies and modifying the statements should we conduct efa or directly conduct cfa

  • @anugrewal5693
    @anugrewal5693 2 роки тому +1

    sir if we take different dimensions from different authors in one questionnaire like for example for
    measuring green practices we take
    energy, water,air etc
    related statements in one questionnaire now can’t we use efa to check whether they correctly measure green practices ?

    • @adilzia
      @adilzia  2 роки тому

      yes, if u have used different items of the scale from previously developed scales u have to go for EFA followed by CFA.

    • @anugrewal5693
      @anugrewal5693 2 роки тому +1

      @@adilzia ok sir thank you
      But if we want to extract factor and then compare factor with demographic variables then is there any need to use cfa?

    • @adilzia
      @adilzia  2 роки тому

      @@anugrewal5693 that’s fine.

    • @adilzia
      @adilzia  2 роки тому

      @@anugrewal5693 whenever u change or use the same scale u have to go for CFA. This is a part of the validity check for any scale.

    • @anugrewal5693
      @anugrewal5693 2 роки тому +1

      @@adilzia okay sir cfa need to be used in combination with sem?
      But if we don’t have model ?

  • @shipra1305
    @shipra1305 2 роки тому +2

    Thank you for the video, it is really helpful.
    Sir can you please confirm if I can use CFA and qualitative data in a single descriptive study. I want to use a mixed method approach. So is it right if I use interviews and survey and use CFA for survey and thematic analysis for interviews. Also can this be called triangulation?
    Thanks in advance!

    • @adilzia
      @adilzia  2 роки тому +1

      yes, you can use it but for triangulation, you have to explain in the methodology section of the research that it includes different methods and theories.

  • @kanwalpreet
    @kanwalpreet 3 роки тому +3

    Sir, after your video, I have a doubt. I have identified factors affecting passengers satisfaction (public transport) from different papers and have accordingly framed statements/indicators for my questionnaire. Factors eg...comfort, punctuality and regularity, etc. So, no need for EFA now? Directly shall I go for CFA? I have almost modified all the statements. Some researchers have mentioned about comfort of seats in *comfort* factor and some in *tangibles* ,because of which I have almost modified almost majority statements under each factor..please advise the next step

    • @adilzia
      @adilzia  3 роки тому +3

      Dear Kanwalpreet, in my video i have very clearly explained that if u are only modifying the existing scale you should directly go for CFA. EFA is for generating the variables and not the items of the variables. In your case if the variables are same then go for CFA. If you have changed the model and generated new variables with major change, go for EFA. I hope this answers your question. Best of luck and keep researching.

    • @kanwalpreet
      @kanwalpreet 3 роки тому

      @@adilzia yes sir. Thanks a lot. Means a lot when you are helping people like us🙏

    • @kanwalpreet
      @kanwalpreet 3 роки тому +1

      @@adilzia sir you are from which domain?

    • @adilzia
      @adilzia  3 роки тому +1

      @@kanwalpreet I am basically from Marketing and Research Method is my interest. These days I am working on the systematic and bibliographic literature review using various software.

  • @sandarshikabandara6368
    @sandarshikabandara6368 2 роки тому +1

    Sir, Can I use a questionnaire as it is, which is already created to identify factors effecting on some thing, in another country? and what analyzing method should I use?

    • @adilzia
      @adilzia  2 роки тому

      yes, u can use it. just u have to give reference to the author who has created that questionnaire.

  • @aroojnajmussaqib4632
    @aroojnajmussaqib4632 3 роки тому +1

    For CFA, dont we need EFA matrix table ? even if we translated the scale.

    • @adilzia
      @adilzia  3 роки тому

      no need. use the same constructs and perform the CFA

  • @A.m.0.0.9
    @A.m.0.0.9 2 роки тому +1

    What if we add a new construct in the existing model?? Is it still CFA??

    • @adilzia
      @adilzia  2 роки тому +1

      If you add only one variable it is still CFA for the same model.

    • @A.m.0.0.9
      @A.m.0.0.9 2 роки тому

      Thankyou sir!

  • @aashitiwaryofficial3168
    @aashitiwaryofficial3168 2 роки тому

    Nice explanation sir, thank you. A question comes to me is that what if very few literature is available on any topic or let's say that some information is there but not from the concerned geographical area we want to research, in that case what will be the research type?

    • @adilzia
      @adilzia  2 роки тому

      It's great that you liked the explanation. in your case, it will be Descriptive research. see 2:29. Don't forget to share it so that more and more people can get benefit from it.

  • @kepo364
    @kepo364 3 роки тому +1

    Hi,
    There has been a popular mental health syndrome coined that has several components, and are associated with sexual abuse survivors . I would like to verify that the components are corrrect and can stand, and that these are a problems that are more associated with sexual abuse survivors than other types of abuse. Can we use CFA? or can you recommend other statistical methods? all variables are continuous, but we can also categorize. i was contemplating between CFA and Latent Profile Anakysis (LPA)

    • @adilzia
      @adilzia  3 роки тому

      Dear ...., i don't have the details of your research questions and the associated objectives of your research. But generally, we use EFA to generate the variables and CFA to confirm that these variables are measuring what they should measure. In your question, I understand that you want to test if sexual abuse is associated with health syndrome. in this case, first of all, decide the Dependent and independent variables and based on the objectives use regression analysis.

  • @AmartyaGhosh
    @AmartyaGhosh Рік тому +1

    Thank you very much Sir for explaining the difference between EFA and CFA with such clarity. I have a question i.e., if we take items and constructs from existing studies that are already validated - do we still need to do an expert validation of the questionnaire or can we directly proceed for CFA? Would greatly appreciate your response, Sir. Best Regards.

    • @adilzia
      @adilzia  Рік тому +1

      if you take items and constructs from existing studies that are already validated you just need to apply CFA.
      Happy researching.

    • @AmartyaGhosh
      @AmartyaGhosh Рік тому

      @@adilzia Thank you very much Sir. Take care.

  • @deepthiamith5957
    @deepthiamith5957 3 роки тому

    Sir
    Video is very good
    Concepts were clear
    I am a Research scholar planning to develop EFA and CFA model for QWL and WLB
    Can u help me sir

  • @quisarr.3721
    @quisarr.3721 2 роки тому +1

    Thanks a lot sir.Your videos are so helpful.They are easy to understand.This video cleared many of my doubts.
    Now with due respect I want to ask a question.
    My phd topic has three constructs and i developed the dimensions of one construct(mediator one)based on the themes generated from interviews.These dimensions are supported with the literature.The items of other two constructs are already developed.
    Should i go for EFA for this paricular construct which i developed myself.
    Or should i go for CFA .THANKS IN ADVANCE

    • @adilzia
      @adilzia  2 роки тому

      Apologies for replying late. U have to see if the new model has majority of items of ur own, u have to go for CFA. Otherwise CFA will do. Basically CFA is for the old model or the items of the model with only less than 50% off items.

  • @abhijatshukla4592
    @abhijatshukla4592 2 роки тому +1

    To apply CFA for descriptive research ,should we first conduct EFA?

    • @adilzia
      @adilzia  2 роки тому

      Dear @abhijat u have to understand the reason behind your analysis. There are different reasons behind CFA and EFA. we use EFA for factor reduction and CFA for model fit.

    • @adilzia
      @adilzia  2 роки тому

      I am replying to your email question.
      There are basically two types of hypotheses in your research.first related to influence or impact and second related to the difference between the two groups of respondents. The t-test and ANOVA are applied for the differences in the two or more groups. Multiple regression can be applied for the impact or influence. You need to adjust ur hypothesis. Talk to your research guide and streamline the hypothesis. Hope This answers your question.
      Take care.

  • @TeachersGurukul
    @TeachersGurukul 3 роки тому +2

    Dear Sir, Thanks for wonderful explanation A query Kindly reply.. I am trying to develop a Questionnaire for assessing the level of Happiness of school teachers... Gone through previous established scales, Shortlisted a few factors/ Domains on the basis of interviews and previous happiness/ wellbeing scales... Shortlisted a few specific statements in each domain.... Should I go for CFA directly or EFA required.... Would be great help if replied asap...Thanks

    • @adilzia
      @adilzia  3 роки тому +3

      In the video, I have clearly explained that if u are only modifying the existing scale you should directly go for CFA. EFA is for generating the variables and not the items of the variables. In your case, if the variables are the same, go for CFA. If you have changed the model and generated new variables with major changes, go for EFA. I hope this answers your question. Best of luck and keep researching.

    • @TeachersGurukul
      @TeachersGurukul 3 роки тому +1

      @@adilzia Thanks a lot for prompt response... Just want to add that variables/ domains are shortlisted after expert panel discussion (From different questionnaires/ scales of measuring wellbeing/ happiness) i.e. not based on single questionnaire instead from different standardized scales and then statements shortlisted in each domain from different scales mentioned to measure that particular domain/ variable as domains/ variables are overlapping as well.. Kindly comment

    • @adilzia
      @adilzia  3 роки тому +1

      @@TeachersGurukul exactly what do you want me to comment on? It's a procedure that you want to follow to make a questionnaire. that's fine.

  • @kossonouprunelle7576
    @kossonouprunelle7576 3 роки тому +1

    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.

    • @adilzia
      @adilzia  3 роки тому

      see the video again. I have explained and discussed the conditions to apply EFA and CFA. to summarise you can see that how many variables are newly identified. if you think that the concept is altogether novel, go for EFA. if you are just updating the existing constructs in the scale, go for EFA.

    • @kossonouprunelle7576
      @kossonouprunelle7576 3 роки тому

      @@adilzia okay, thanks a lot

  • @tajammalrasool812
    @tajammalrasool812 3 роки тому +1

    Respected Sir, I developed questionnaire for my PhD research. I took factors from literature and developed own item statements from literature. I was applying EFA for factor Loading... So i think i should apply CFA... Is it?

    • @adilzia
      @adilzia  3 роки тому

      Yes. EFA followed by CFA.

  • @Dr.RajKumar-ue4uk
    @Dr.RajKumar-ue4uk 3 роки тому +1

    RESPECTED SIR, CAN WE APPLY EFA TO THE DATA COLLECTED THROUGH OPEN-ENDED QUESTIONNAIRE. IF YES THEN HOW WE CAN APPLY

    • @adilzia
      @adilzia  3 роки тому

      dear @raj kumar, it is not the questionnaire on which the EFA is applied, rather it is the data collected on the basis of a questionnaire is evaluated using EFA or other tools. In your case, if you have collected the data in open-ended questions then you have to convert that qualitative data into quantitative data to apply EFA. I hope it answers your question.

    • @Dr.RajKumar-ue4uk
      @Dr.RajKumar-ue4uk 3 роки тому

      @@adilzia thank your sir, got your answer

  • @SunakshiPhdscholar
    @SunakshiPhdscholar Рік тому +1

    Hello Sir I need help in my data analysis part can you please help out pls share your mail id if you can help me

    • @adilzia
      @adilzia  Рік тому

      sorry i don't provide coaching.