User dsimcha

349 votes

Accepted

ROC vs precision-and-recall curves

answered Feb 14, 2011 at 18:11

201 votes

The Two Cultures: statistics vs. machine learning?

answered Feb 14, 2011 at 22:09

84 votes

Accepted

What do you call an average that does not include outliers?

answered Feb 2, 2009 at 14:24

70 votes

Is normality testing 'essentially useless'?

answered Sep 18, 2010 at 2:32

15 votes

Why would all the tests for normality reject the null hypothesis?

answered Oct 6, 2011 at 18:27

12 votes

What skills are required to perform large scale statistical analyses?

answered Mar 6, 2011 at 21:38

8 votes

If the t-test and the ANOVA for two groups are equivalent, why aren't their assumptions equivalent?

answered Feb 4, 2011 at 1:29

8 votes

Accepted

Testing statistical software

answered Feb 24, 2011 at 14:15

6 votes

When conducting a t-test why would one prefer to assume (or test for) equal variances rather than always use a Welch approximation of the df?

answered Sep 21, 2010 at 1:36

6 votes

Accepted

Computing chi-square for large tables with some expected cell counts equal to zero

answered Feb 21, 2011 at 2:54

4 votes

Is it possible to prove a null hypothesis?

answered Feb 10, 2011 at 15:29

3 votes

Big-O Scaling of R's cmdscale()

answered May 17, 2011 at 18:24

1 vote

GLMNET or LARS for computing LASSO solutions?

answered Feb 10, 2011 at 19:19

0 votes

What are disadvantages of using the lasso for variable selection for regression?

answered Mar 7, 2011 at 0:38

Stack Exchange Network

14 Answers

ROC vs precision-and-recall curves

The Two Cultures: statistics vs. machine learning?

What do you call an average that does not include outliers?

Is normality testing 'essentially useless'?

Why would all the tests for normality reject the null hypothesis?

What skills are required to perform large scale statistical analyses?

If the t-test and the ANOVA for two groups are equivalent, why aren't their assumptions equivalent?

Testing statistical software

When conducting a t-test why would one prefer to assume (or test for) equal variances rather than always use a Welch approximation of the df?

Computing chi-square for large tables with some expected cell counts equal to zero

Is it possible to prove a null hypothesis?

Big-O Scaling of R's cmdscale()

GLMNET or LARS for computing LASSO solutions?

What are disadvantages of using the lasso for variable selection for regression?