Excellent! I have not ever studied this topic in any one of my statistical courses in two previous degrees. In fifteen minutes I understood everything! :) Thanks a lot!
Hi Phil, thanks for this. I was wondering whether you'd be able to give me some more resources about which statistics should be run here with and without weightings i.e. descriptives rather than inferential. I am trying to run an ordinal logistic regression and from your video on that topic you also have to weight cases by frequencies to this, so I'm unsure how you would also then weight it by the factor I'm wanting to weight initially. Thanks.
3:10 "I don't like Excel." Ha Ha. You're not the only one. Microsoft works too hard trying to figure out what you want before you do it. Anyway, I might add that the mean of age*weight also gives the weighted mean age, just in case anyone wants to know how the math is done. Good vid!
Let me know if you find an answer to this question. I am looking around for a method of weighting multiple variables. For example, I would like to weight my data by Race/ethnicity and gender. aei11@txstate.edu
Phil I get a warning message when I run the weighted calculation that says on at least one case the value of the weight variable is zero, negative or missing...such cases are invisible to statistical procedures which need positively weightless cases. I've dummy coded gender 1/0 f/m what does this mean?
Thanks for this useful video! One question:I gathered a sample and it is not representative for the population. At least for sex and age. First I made a weight variable for sex and the results after weighting slightly differ, not so much actually. So, can I create an extra weight based on age and use it both?
can you tell me if one can perform a kruskal wallis test/mann whitney u test using weighted data? I have 3 groups, G1 n=368, G2 n=160 and G3 n=95. These are my ordinal groups, my variables are sperm count, motility and volume. I want to know if I can weight the data, then rank it to perform the test. Or do I randomly select 95 subjects from the first two groups to match my third group, and perform the test (unweighted)......help please
50/50 is what the % split should be for Males/Females in the population. So, he's merely merely weighting his Male/Female sample up/down to match the 50/50 % split in the population.
thank you for sharing , like how relatable you are I was laughing whole time at your side comments
I really appreciate.. This is a free source, we see it if wana learn something or we choose to pass.. :-)
Excellent! I have not ever studied this topic in any one of my statistical courses in two previous degrees. In fifteen minutes I understood everything! :) Thanks a lot!
Hi Phil, thanks for this. I was wondering whether you'd be able to give me some more resources about which statistics should be run here with and without weightings i.e. descriptives rather than inferential. I am trying to run an ordinal logistic regression and from your video on that topic you also have to weight cases by frequencies to this, so I'm unsure how you would also then weight it by the factor I'm wanting to weight initially. Thanks.
3:10 "I don't like Excel." Ha Ha. You're not the only one. Microsoft works too hard trying to figure out what you want before you do it. Anyway, I might add that the mean of age*weight also gives the weighted mean age, just in case anyone wants to know how the math is done. Good vid!
Amazing, thank you very much
What if I have to use a combination of weights?
Let me know if you find an answer to this question. I am looking around for a method of weighting multiple variables. For example, I would like to weight my data by Race/ethnicity and gender. aei11@txstate.edu
Thanks, I appreciated this video!
Phil I get a warning message when I run the weighted calculation that says on at least one case the value of the weight variable is zero, negative or missing...such cases are invisible to statistical procedures which need positively weightless cases. I've dummy coded gender 1/0 f/m what does this mean?
man u r a wizard
Thanks for this useful video! One question:I gathered a sample and it is not representative for the population. At least for sex and age. First I made a weight variable for sex and the results after weighting slightly differ, not so much actually. So, can I create an extra weight based on age and use it both?
can you tell me if one can perform a kruskal wallis test/mann whitney u test using weighted data? I have 3 groups, G1 n=368, G2 n=160 and G3 n=95. These are my ordinal groups, my variables are sperm count, motility and volume. I want to know if I can weight the data, then rank it to perform the test. Or do I randomly select 95 subjects from the first two groups to match my third group, and perform the test (unweighted)......help please
excellent tutorial. Thanks
where did you get 50?
50/50 is what the % split should be for Males/Females in the population.
So, he's merely merely weighting his Male/Female sample up/down to match the 50/50 % split in the population.
Great! But you may want to use a bit tighter prepared script in the future... A little too much back-and-forth. Fun though :).
😂 fuck this was slow...
prep before recording mate.. 😝
I disagree...I thought it was a great explanation.
I got lost.
OneMileyCyrusFanVlog I fell alseep