Like in all your videos, it is very useful that you provide a detailed information of the statistic application and interpretation (I liked the initial comparison, it is like MANOVA in reverse!). I also appreciate when you emphasize something that is important and requires extra attention.
Thank your for your informative video Dr. Todd Grande. We can calculate the formula with coefficients but to try our model on other applicants or patients how can I find the cut-off values for classification. Maybe you have already said the answer to my question but I couldn't catch that point.
It is clearly enough Dr. Thank you so much. But I d like to ask question. How to know the specificity or the accuracy of each predictor variable to predict group , both in original or cross validated grouped?
Thanks for this awesome explanation! Just a correction, I believe discriminant analysis can be done for more than 2 classes for the dependent variable. I have seen examples. Or am I wrong here? Please suggest.
Yes, ok for more than 2 groups and Mr. Grande also points out that this is possible. What Mr. Grande says cannot be is that discriminant analysis cannot be done with more than one dependent variable.
how do you determine the between-group matrices here? And do you have an example of min(k - 1, p) to identify the number of discriminant functions you should expect to emerge from the present DFA?
Hi Dr. Grande. I have got 2 functions in my descriminant analysis. How do i interprete the output. Which function to ne considered out of 2 to make an equation?
Is this analysis will be called as Linear Discriminant Analysis? I am confused with the term canonical discriminant analysis. Is it same as LDA? Can anyone answer please.
This was a tough one for me. I'm not sure I am understanding the difference between sensitivity with the successful students, and the specificity with the not successful. High sensitivity means few false negatives, and high specificity mean few false positives?
A colleague did a segmentation using a descriminant analysis in SPSS, the colleague left and the client now ask me the classification function coefficients and fisher's linear descriminant function. How can I generate these? I have the SPSS which include the variable with the new 6 segments. Any suggestions?
Dear Todd, Can you please tell me how I can analysis between different methods for its accuracy? For example, I want to see which of three different circumference measurements can accurately measure overweight?
Dr. Grande, I noticed that in the final dialogue box you chose Predicted group membership and Probabilities of group membership while in both Field (2009) and Mertler and Vannatta (2013) they chose discriminant scores and unchecked the other two. Can you explain the difference in data and possible the questions that would require one method or the other. Thanks.
+Elizabeth Adair Hi Lisa, I think the "Save" dialogue box allows you to create additional variables (hence Ds_1 and Ds1_1). In this case, the created variable allows you to see how the prediction compares with the actual outcome. I think it is not a necessary part of data analysis, but just an interesting way to visually see the prediction and outcome. Perhaps Dr. Grande can give additional clarification.
+Elizabeth Adair This is just a matter of personal preference. I like to view the predicted group membership and the probabilities of group membership, however, it is not critical to conducting a discriminant analysis.
Hello. What happen with the step wise? I tried with both ways (step wise and independent together) and the solution is different, which is the best solution? Or am I doing something wrong? Thanks 😎
Hi everyone, could someone give me some help? So you use the "standardised canonical discriminant function coefficients" box to show which variable best predicts the outcome. Is there a statitstical test you can conduct to see whether there is a significant difference in how much the predictor variables predict the outcome? I want to see whether a between-groups difference is larger for one dependent variable compared to another. (E.g. whether gender differences in height are larger than gender differences in weight). Is this the way to find that?
+Dr.Grande, I am studying LDA as part of Masters in Statistics. Kindly I would like to ask you whether you can send me the data used in this example please.
Thank you so much Dr. Todd. I am not familiar with discriminatory analysis, it gives me hint. What if we have dummy variables as independents, should we follow the same procedure in spss? For further analysis, can you please explain more when to re-classify (successful/not successful) according to SPSS output or keep the original classification? Thank you again.
First you show how you put text in place of numerical values? this was a big mistery for me. I am unable to do that. You are not showing your actual data sheet and data feeding in SPSS.
Like in all your videos, it is very useful that you provide a detailed information of the statistic application and interpretation (I liked the initial comparison, it is like MANOVA in reverse!). I also appreciate when you emphasize something that is important and requires extra attention.
You saved my life with your spss videos
Thank you very much! Very helpful. Simple and clear explanations! You have helped me with three of my doctoral assignments.
Thank your for your informative video Dr. Todd Grande. We can calculate the formula with coefficients but to try our model on other applicants or patients how can I find the cut-off values for classification. Maybe you have already said the answer to my question but I couldn't catch that point.
Dear Todd, this is a very useful simplification of the discriminant analysis. You made my life easy :) Thanks
You are welcome - thank you for watching -
This is very helpful...thank you for posting.
This is a tough one thanks for the helpful explanation
amazing tutorial. Please, do make the volume of your videos higher. I had to connect my computer to speakers to listen to what you were saying.
I thought this video did a good job of helping me understand discriminant analysis and the steps to conduct the test.
nice video with relevant explanations
Thank you so so much! Extremely helpful!
Very important and helpful!
Thanks for your reassuring vids Doc.
It is clearly enough Dr. Thank you so much. But I d like to ask question. How to know the specificity or the accuracy of each predictor variable to predict group , both in original or cross validated grouped?
Thanks for this awesome explanation! Just a correction, I believe discriminant analysis can be done for more than 2 classes for the dependent variable. I have seen examples. Or am I wrong here? Please suggest.
Yes, ok for more than 2 groups and Mr. Grande also points out that this is possible. What Mr. Grande says cannot be is that discriminant analysis cannot be done with more than one dependent variable.
Thanks a lot for a very good tutorial
You're welcome!
You're the best
Would this be a good analysis to use when trying to test variables from an intersectional framework?
how do you determine the between-group matrices here? And do you have an example of min(k - 1, p) to identify the number of discriminant functions you should expect to emerge from the present DFA?
Hi Dr. Grande. I have got 2 functions in my descriminant analysis. How do i interprete the output. Which function to ne considered out of 2 to make an equation?
Do all predictors need to be in the same direction? e.g., a higher score should theoretically lead to a greater likelihood of success?
Thanks Dr. Grande.
understood very well, thank you!
hi dr todd, is there a way that we can tell the most significant variable using discriminant analysis.
Is this analysis will be called as Linear Discriminant Analysis? I am confused with the term canonical discriminant analysis. Is it same as LDA? Can anyone answer please.
I have got a similar question. What is the difference between Canonical, Stepwise, Direct discriminant function analysis?
This was a tough one for me. I'm not sure I am understanding the difference between sensitivity with the successful students, and the specificity with the not successful. High sensitivity means few false negatives, and high specificity mean few false positives?
thank you Dr. Grande.
8:40 What would be considered too high of a correlation to continue with the DA?
thanks a lot from the core of my heart
thanks.. this helped me in my presentation
A colleague did a segmentation using a descriminant analysis in SPSS, the colleague left and the client now ask me the classification function coefficients and fisher's linear descriminant function. How can I generate these? I have the SPSS which include the variable with the new 6 segments. Any suggestions?
Thank you for the explanation! Can you share the Datasheet you used to run this?
Dear Todd, Can you please tell me how I can analysis between different methods for its accuracy?
For example, I want to see which of three different circumference measurements can accurately measure overweight?
Dr. Grande, I noticed that in the final dialogue box you chose Predicted group membership and Probabilities of group membership while in both Field (2009) and Mertler and Vannatta (2013) they chose discriminant scores and unchecked the other two. Can you explain the difference in data and possible the questions that would require one method or the other. Thanks.
+Elizabeth Adair Hi Lisa, I think the "Save" dialogue box allows you to create additional variables (hence Ds_1 and Ds1_1). In this case, the created variable allows you to see how the prediction compares with the actual outcome. I think it is not a necessary part of data analysis, but just an interesting way to visually see the prediction and outcome. Perhaps Dr. Grande can give additional clarification.
+Elizabeth Adair This is just a matter of personal preference. I like to view the predicted group membership and the probabilities of group membership, however, it is not critical to conducting a discriminant analysis.
Hello. What happen with the step wise? I tried with both ways (step wise and independent together) and the solution is different, which is the best solution? Or am I doing something wrong? Thanks 😎
Thanks, and do you have a paper related to this video for more details and to use it a a reference for the provided information
Can u explain how to use Quadratic Discriminant Analys in spss?
Hi everyone, could someone give me some help?
So you use the "standardised canonical discriminant function coefficients" box to show which variable best predicts the outcome. Is there a statitstical test you can conduct to see whether there is a significant difference in how much the predictor variables predict the outcome?
I want to see whether a between-groups difference is larger for one dependent variable compared to another. (E.g. whether gender differences in height are larger than gender differences in weight). Is this the way to find that?
+Dr.Grande, I am studying LDA as part of Masters in Statistics. Kindly I would like to ask you whether you can send me the data used in this example please.
The Grandayy we really need
thanks a lot
Thank you so much Dr. Todd. I am not familiar with discriminatory analysis, it gives me hint. What if we have dummy variables as independents, should we follow the same procedure in spss? For further analysis, can you please explain more when to re-classify (successful/not successful) according to SPSS output or keep the original classification? Thank you again.
First you show how you put text in place of numerical values? this was a big mistery for me. I am unable to do that. You are not showing your actual data sheet and data feeding in SPSS.
Where can I get the data Sir!!
can I get this data set
How to apply the test in SPSS discriminent analysis ???? guys give me a answer
please share the data as well
Speak a bit lounder and with enthusiasm I started feeling sleepy after half of the video 😴