Factor analysis is a pretty complicated concept for me to grasp, so I like that this is explained in video form. If I start to get lost, I can rewind and figure things out for myself, whereas if this were a regular lecture, it would be easy to get lost in such a difficult concept. Showing this in SPSS makes me feel like I would be able to do this myself, something I thought was way above my skill level.
I'm usually alright with stats, but this definitely way over my head! I did like the explanation of how the factor analysis breaks down the relationships between variables though.
Great, this is a very clear (you could say even slow) explanation of factor analysis, so it is perfect and easy to understand even if you are fairly new to the concept.
Great step by step presentation of using SPSS. It's a lot to learn and master using SPSS but Dr. Grande takes his time to explain the process very well. I have used SPSS before and do I know it be can be used to determine correlations of items. I definitely have to watch again and again in order to remember the process. Great work.
Thank you so, so much for sharing your knowledge and expertise. I have basically taught myself all of the stats required for my research using your videos. I love your delivery, clarity and depth of knowledge. Thank you!
Sir suppose I have develop a scale.....And done pilot study....Now exploratory factor analysis should be done with pilot data or with final data set??.Pls reply.
In this video, I understand what Dr. G. is taking about, but I would need quite a bid of practice to completely understand how to compute the data. My concern is if I would have to compute factor Analysis using SPSS. The step by step procedure was good.
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 I want to know how to draw box plot for given data sets comparative to standard value line. As example if I want to draw the box plot for Fluoride distribution in drinking water comparatively WHO standard of 1.5 mg/l , how can i draw that?
I am trying to do a moderation analysis by using process v3.3. I have 3 moderators, 3 in dependent variables and a dependent variable. But in the output it doesn't give me the conditional statistics. What could be the reason? And also for 1 independent variable it doesn't give me the result at all. In the out put it says there is an error and calculation stopped. Could anyone say what is the reason and what am I doing wrong? I will really appreciate.
Am I right in thinking that this is an example of Exploratory Factor Analysis (because nowhere does he tell SPSS that 1-5 are meant to be one factor etc)? And not Comfirmatory Factor Analysis.
Suppose on a 10 item scale (i.e a subscale of a larger scale), 4 components was extracted with diverse loadings (widely spread) among the items. Shall i retain the original items or remove those items that are alienated from others. Please
This video was very informative and helped to refresh my memory about how to interpret the data. Using 100 participants helped provide a better understanding of the different potential outcomes as well. I also appreciated the definition of the term factor loadings as I was not familiar with that term.
Im doing a study on employee dissatisfaction level and early retirement. there are main contributory factors and under which several questions. example factor A got 6 questions and factor B got 4 questions like wise. Is the method you shown here applicable to analyze my questions
I appreciate the step by step teaching going on in this video. The topic is difficult for me, but I feel confident that it I could accomplish it with the use of this video as a directive. Great video!!!
I am not sure if this a stupid question but our professor wants us two report the factor structure. Is the correlation matrix the same as the factor structure?
Edit: I saw your recent videos, Audio was loud enough on those vids. Why you English guys speak on so low volume (decibel)? I turned to 100% volume still can't get the volume. Some other RU-vid channels are louder enough.
Thanks for the video. This will certainly be helpful for my exam tomorrow. But, I have one question... not sure if its too obvious or "too wrong"... after I group the variables in factors how do I assign values to these factors, as on my survey I had collected data for the variables, not the factors? Or is my line of thinking absolutely wrong in this concept?
This video was very clear and easy to understand. I appreciate the in depth explanation of factor analysis with a real example/data set that we can work through while understanding. It was easy to see the real world application for using this function. This was my first experience with spss, and have mainly used excel. This is the first time seeing a correlation matrix like this, although it was easier to understand when dr G went through and explain what to look for when reading the matrix. A good reminder about what constitutes a strong and weak correlation. I also learned about the determinant which I never knew before. Good video!
The section in the book on Factor Analysis is very very brief. Dr. G does an excellent job providing detail explanation about Factor Analysis, and applying it with SPSS.
I appreciate the explanation on factor analysis but having it explained in SPSS is overwhelming for me because I am not familiar with SPSS. Maybe it would make more sense if I was familiar with the program.
Woah. Didn't realize you made these types of videos. Currently using SPSS to run a few analyses on different factors that are correlated with a person's expectancy of success so that I can include the information on a report.
Hi. My topic is on A Study of Budget Slack and the Influence of Budget Participation on Budget Slack. I would like to do hypothesis testing using factor analysis. My hypothesis are: H1 : Managers' perceptions of budgetary participation and distributive fairness are positively related. H2 : Managers' perceptions of budgetary participation and procedural fairness are positively related. H3 : Distributive fairness is positively related to managerial trust H4 : Procedural fairness is positively related to managerial trust H5 : Managerial trust is positively related to budget goal commitment H6 : Budget goal commitment is negatively associated with managers' propensity to create slack. May I know what type of factor analysis is suitable? Exploratory or confirmatory factor analysis? May I know any other statistical methods that are appropriate to conduct hypothesis testing? Thanks.
9:00 became interesting for me I am having the issue where the determinant is ,000. but dont know if that is below ,00001. And if the items are highly uncorrelated with eachother this will mean that I have a problem as it seems that I have a low reliability with these items. Furthermore, I am not getting any info on KMO and Barlett's test. I did selected it in descriptives.
please continue with how we can generate factors scores. I'm struggling with Refined and Non-refined method. how the results from different methods can be used for? Refined methods include Regression, and.... can be seen in SPSS. Where the non-refined methods be defined in SPSS? Thank you very much.
Hello, I was watching your video and somehow started thinking about another issue unrelated to factor analysis but thought I would ask you advice. I have survey results that include a PRE intervention condition and a post intervention condition. It was supposed to be a paired samples t-test but when the results came back there was no way to pair up the scores. So I no longer have a paired samples test but I also don't have an independent samples test as the groups are not independent. What test would you suggest? Overall, the people that answered the PRE and POST survey are the same people just not paired up. They are matched in other ways - they are all nurses and also all fall under another category (filed of study). My dependent variable is ordinal (likert scale). What are your thoughts. I have already run a sign test, Wilcoxon signed ranks test, Chi-Square test of independence, and a paired samples t-test. All of which produce the same results in terms of significance levels. Thanks, Jeff
The video was very clear and easy to follow. I did watch it several times and I am still not 100% percent sure if I have truly grasped the whole concept. It would be interesting to have a hands on approach which might help make the information click.
Update: thanks for the like Dr. Todd. I've learned a lot from you videos about the Big Five Personality. However, I would still prefer if you spoke a bit louder and showed more enthusiasm while explaining to make the learning experience more engaging and enjoyable. Thanks! A very good explanation but I wish if you spoke a bit louder.
Pretty much agree with everyone else is that factor analysis is an interesting process and definitely useful, but I hit a block with the fact that I have never used or even played around with SPSS. I do find it pretty easy to understand the Excel videos though
I tried to adapt the knowledge from you video to my data, firstly tested factorability with KMO and Barlett`s test of sphericity, which indicated that Exploratory factor analysis is possible. However my factor structure with absolutely messy. Meaning I had lots of Cross loadings, low correlations and table results were different than scree plot suggestion. Could you please make a video, showing how to deal with something like that, with no perfect data. Anyway thank you so much for your videos!
This must be the third or fourth time I watched this video in the past 1 year and I think I am finally beginning to understand factor analysis. So, we use it to validate our self-made constructs (which I totally avoid).
Hello Dr Todd Grande! I just conducted this Factor Analysis and I ended up with 3 components. However, I need to measure the effects on this variable, and measuring the effect this variable has on another- so having 1 would be a lot better. Do you know what I can do to fix this without deleting too many items?
Once you have a suitable factor solution and you wish to look at correlations I'm told you need to create standardised factor scores that take only positive values. The method I used to do this was adding a constant to each factor, and taking the fact score away from the constant when the factor was measuring in a "negative light" so for example a factor measuring antisocial tendencies was transformed to measuring sociability. However, when I produced correlations with age none were significant. Have I missed a step or made a mistake?
After discovering a correlation between items above 0.8, what would be the correct way to deal with this 'too high' correlation? Average the 2 out into a new item?
Hi Dr. Grande. First things first. I thought you should know that I got a 1st on my quantitative research because of your videos! So thank you very much for putting these up, with them I wouldn't have understood the numbers that my quantitative tests gave me in my project so THANK YOU! Secondly is there any way you could explain to me what Factor Analysis is. I've tried everything to get it and I'm not sure at all to be honest. What does it mean to see how the factors load together. What is a factor exactly?
Can I ask a question? What if theres no structure matrix result in factor analysis data what are we going to do to interpret our data. Thank you so much
"Grief is not passive but active, grief reveals and challenges while it deals with the facts, it lends itself to truth in a way no other emotion can, it identifies all of the participants in tragedy and allows them their role in the universe." Quote unknown. thanks
It means that there is high degree of multicollinearity in your database. Have a look at the correlation coefficient matrix, seek rho value of more than 0.8 - You may need to remove those variables (with rho of more than 0.8) before running in factor analysis a second time. Conversely, you dont want to have very low rho values close to zero either as this would indicate no correlation at all and as mentioned earlier deselect variables with high multicolinearity with rho above 0.8.
Hello Dr I am really pleased to see your informative lectures. I have a question regarding the cut off value for correlation matrix determinant which is said to be 0.0001 in this video. How you can determine that cut off value. Second, your video is lacking information on the interpretation of the correlation (Sig-1 tailed) test scores. if 0.05 is the P value than the assumption we have is the P value less than 0.05 we will reject the null hypothesis (A=B) which is in contrast to the correlation results. for example two of my variables have a value of 0.754 and P value 0.000. With correlation, this shows two variables are 75% resembling each other and P value is saying they are significantly not equal. Please elaborate it for my understanding. Thanks Salman
I struggle with using the SPSS, excel etc. to do factor analysis. Dr. G. makes it look easy but I don't think I could do it on my own yet. Very interesting to watch but complicated.
thanks for the video! do i need to take the logs from the original data before starting? and then but the log at "item 1" etc or do i just use the original data that i collected?
There were several things in this video that I would need to explore further to understand the analysis of the results in this video. I think I have a basic understanding of what a factor analysis is but I feel like I would need an entire class to go over the concepts in this video especially pertaining to the different results and what they mean.
dr grande. i have a sample size of about 500, and had three constructs on the survey instruments. the factor analysis is all over the place. i can share the spss file with you. what is the way ahead to get to the factors? i can share the screen shots if i can have your mail ID