Skip to main content
Cross Validated

Return to Question

Tweeted twitter.com/#!/StackStats/status/516975764265127936
correcting math
Source Link
edit approved Sep 30, 2014 at 14:09
Wrzlprmft
edit approved Sep 30, 2014 at 14:09
Wrzlprmft
  • 2.4k
  • 1
  • 20
  • 38

I have 50 variables in my dataset. I have correlated each variable against each variable, thus I have 50*50 = 1225$49·50/2 = 1225$ unique correlations when no variable is tested against itself. Now, suppose I want to correct the statistics for multiple comparison, and that for some reason (well, just for learning now) I want to use the Bonferroni correction. Let the threshold for significance be $\alpha = 0.05$. Is the corrected threshold $0.05/1200$ (where 12001225 is number of tests) or $0.05/49$ (because each variable was correlated with 49 other)?

I have 50 variables in my dataset. I have correlated each variable against each variable, thus I have 50*50 = 1225 unique correlations when no variable is tested against itself. Now, suppose I want to correct the statistics for multiple comparison, and that for some reason (well, just for learning now) I want to use the Bonferroni correction. Let the threshold for significance be $\alpha = 0.05$. Is the corrected threshold $0.05/1200$ (where 1200 is number of tests) or $0.05/49$ (because each variable was correlated with 49 other)?

I have 50 variables in my dataset. I have correlated each variable against each variable, thus I have $49·50/2 = 1225$ unique correlations when no variable is tested against itself. Now, suppose I want to correct the statistics for multiple comparison, and that for some reason (well, just for learning now) I want to use the Bonferroni correction. Let the threshold for significance be $\alpha = 0.05$. Is the corrected threshold $0.05/1200$ (where 1225 is number of tests) or $0.05/49$ (because each variable was correlated with 49 other)?

edited body
Source Link
log
log
  • 41
  • 1
  • 3

I have 50 variables in my dataset. I have correlated each variable against each variable, thus I have 50*50 = 12001225 unique correlations when no variable is tested against itself. Now, suppose I want to correct the statistics for multiple comparison, and that for some reason (well, just for learning now) I want to use the Bonferroni correction. Let the threshold for significance be $\alpha = 0.05$. Is the corrected threshold $0.05/1200$ (where 1200 is number of tests) or $0.05/49$ (because each variable was correlated with 49 other)?

I have 50 variables in my dataset. I have correlated each variable against each variable, thus I have 50*50 = 1200 unique correlations when no variable is tested against itself. Now, suppose I want to correct the statistics for multiple comparison, and that for some reason (well, just for learning now) I want to use the Bonferroni correction. Let the threshold for significance be $\alpha = 0.05$. Is the corrected threshold $0.05/1200$ (where 1200 is number of tests) or $0.05/49$ (because each variable was correlated with 49 other)?

I have 50 variables in my dataset. I have correlated each variable against each variable, thus I have 50*50 = 1225 unique correlations when no variable is tested against itself. Now, suppose I want to correct the statistics for multiple comparison, and that for some reason (well, just for learning now) I want to use the Bonferroni correction. Let the threshold for significance be $\alpha = 0.05$. Is the corrected threshold $0.05/1200$ (where 1200 is number of tests) or $0.05/49$ (because each variable was correlated with 49 other)?

Source Link
log
log
  • 41
  • 1
  • 3

About the Bonferroni correction

I have 50 variables in my dataset. I have correlated each variable against each variable, thus I have 50*50 = 1200 unique correlations when no variable is tested against itself. Now, suppose I want to correct the statistics for multiple comparison, and that for some reason (well, just for learning now) I want to use the Bonferroni correction. Let the threshold for significance be $\alpha = 0.05$. Is the corrected threshold $0.05/1200$ (where 1200 is number of tests) or $0.05/49$ (because each variable was correlated with 49 other)?