This is a great video for me. I am working withn SPS in a statistics course. I am not versed in mathematics or statistics. My course is challenging. I am going cite some information from the three videos on Cronbach's Alpha. I can use the youtube url, but I can't find anything that referencves your name! I feel sort of stupid!
The value is negative due to a negative average covariance among items. This violates reliability model assumptions. You may want to check item codings.
my study involves two groups of participants completing the same questionnaire - will I have to conduct a Cronbach alpha test two times for each group? or overall With the complete data set responses? thank you!
hellllo hope your study 8 months ago went well.... i'm also doing two groups completing the same questionnaire (parents and their children), wondering what tests you decided to use? ahaha I have so far used wilcoxon test & mcNemar's test (which i simplified it to just agree/disagree for). Cheers!
Imagine there were two items in a questionnaire. The questionnaire is a 5-point Likert scale with 1 = strongly disagree to 5 = strongly agree. Q1. I like stats. Q2. I hate stats. Say someone dislikes stats. They might score a 1 for Q1 and 5 for Q2. For people that dislike stats, as they score low on Q1, they tend to score high on Q2, giving a negative correlation. Vice versa for people who like stats, but the correlation will still be negative. To get around this problem, before calculating Cronbach's alpha, use reverse scoring. Reverse scoring does something like say for Q1, a score of 1 is converted to a 5; 2 to 3, etc.
Unidimensionality implies the presence of only one factor in the data (and is determined with factor analysis); coefficient alpha (consistency) assumes unidimensionality (it can't test for it). Coefficient alpha simply represents a ratio of true score variance (reliable/consistent) variance to total variance. I can't say for sure what people mean by homogeneity, in this context.
Nope. Still don’t get it.
@Herschel Brammer that is wrong. If you don't trust her- don't be with her... trust and privacy is important.
I want to (and will) kill myself.
What a lovely voice you have 🦉
Thanks! Super helpful video, my stats professor is terrible
it added consistent approach to my knowledge.
Thank you so much for these videos (part 1 and part 2), they are very helpful
Thank you again and again !
Thank you very much! May I have the video's references? Furthermore, are reliability tests necessary for demographic questions?
thanks for simple explanation
Can you explain why it is not a measure of homogeneity?
it is is it not?
i think is the measure of variance
This is a great video for me. I am working withn SPS in a statistics course. I am not versed in mathematics or statistics. My course is challenging. I am going cite some information from the three videos on Cronbach's Alpha. I can use the youtube url, but I can't find anything that referencves your name! I feel sort of stupid!
Don't feel stupid: www.how2statsbook.com/p/about.html
Great explanation - thanks
Can you please show the derivation of the alpha- of the formula
The value is negative due to a negative average covariance among items. This violates reliability model assumptions. You may want to check item codings.
How is Cronbach's Alpha different from Cohen's Kappa in terms of Nominal Data?
my study involves two groups of participants completing the same questionnaire - will I have to conduct a Cronbach alpha test two times for each group? or overall With the complete data set responses? thank you!
hellllo hope your study 8 months ago went well.... i'm also doing two groups completing the same questionnaire (parents and their children), wondering what tests you decided to use? ahaha I have so far used wilcoxon test & mcNemar's test (which i simplified it to just agree/disagree for). Cheers!
Thank you so much
This is a simple explanation? I didn't understand anything that was being said. I haven't studied statistics though, so maybe that is a prerequisite.
thank you!
Thanks for this! :-)
Why there is negative value in Cronbach's alpha? How to correct it?
Imagine there were two items in a questionnaire. The questionnaire is a 5-point Likert scale with 1 = strongly disagree to 5 = strongly agree. Q1. I like stats. Q2. I hate stats. Say someone dislikes stats. They might score a 1 for Q1 and 5 for Q2. For people that dislike stats, as they score low on Q1, they tend to score high on Q2, giving a negative correlation. Vice versa for people who like stats, but the correlation will still be negative. To get around this problem, before calculating Cronbach's alpha, use reverse scoring. Reverse scoring does something like say for Q1, a score of 1 is converted to a 5; 2 to 3, etc.
how to tackle this issue
How is consistency different from homogeneity and unidimensionality?
Unidimensionality implies the presence of only one factor in the data (and is determined with factor analysis); coefficient alpha (consistency) assumes unidimensionality (it can't test for it). Coefficient alpha simply represents a ratio of true score variance (reliable/consistent) variance to total variance. I can't say for sure what people mean by homogeneity, in this context.
Homogeneity and unidimensionality get more at validity as well, I believe.
what is your name? and how can i use this as a reference
???
...Gilles?
explanation starts at 03.10
Uses 5 other terms in the first minute that I didn't understand either.
Not simply explained at all