User user1205901 - Слава Україні

69 votes

9 answers

58k views

How can a distribution have infinite mean and variance?

asked Mar 27, 2014 at 4:49

40 votes

3 answers

3k views

When is it appropriate to use an improper scoring rule?

asked Apr 21, 2016 at 6:14

32 votes

2 answers

7k views

Are 50% confidence intervals more robustly estimated than 95% confidence intervals?

asked Nov 27, 2016 at 6:28

29 votes

2 answers

15k views

Why is the normality of residuals "barely important at all" for the purpose of estimating the regression line?

asked May 17, 2015 at 4:44

28 votes

5 answers

3k views

Why do these statements not follow logically from a 95% CI for the mean?

asked Apr 24, 2014 at 12:08

28 votes

5 answers

51k views

Why do we say the outcome variable "is regressed on" the predictor(s)?

asked Apr 15, 2016 at 2:25

25 votes

2 answers

123k views

Can a statistical test return a p-value of zero?

asked Mar 18, 2014 at 20:10

23 votes

3 answers

46k views

Why does the t-distribution become more normal as sample size increases?

asked Aug 2, 2014 at 6:31

22 votes

2 answers

20k views

What are Bayesian p-values?

asked Sep 7, 2015 at 2:10

22 votes

5 answers

181k views

What is the difference between descriptive and inferential statistics?

asked Oct 5, 2013 at 4:59

19 votes

3 answers

54k views

When doing a t-test for the significance of a regression coefficient, why is the number of degrees of freedom $n-p-1$?

asked Apr 9, 2014 at 6:11

18 votes

2 answers

3k views

How to choose between the different Adjusted $R^2$ formulas?

asked Mar 25, 2012 at 3:12

18 votes

3 answers

50k views

Why use a certain measure of forecast error (e.g. MAD) as opposed to another (e.g. MSE)?

asked Dec 13, 2012 at 21:58

18 votes

4 answers

2k views

Is variance a more fundamental concept than standard deviation?

asked Jun 9, 2016 at 3:16

17 votes

1 answer

8k views

What's the difference between asymptotic unbiasedness and consistency?

asked May 30, 2016 at 11:35

17 votes

2 answers

2k views

Why is Pearson's ρ only an exhaustive measure of association if the joint distribution is multivariate normal?

asked Jul 21, 2015 at 21:00

16 votes

1 answer

3k views

What does it mean to explain variance?

asked Apr 8, 2014 at 15:33

15 votes

1 answer

27k views

What's the difference between regression coefficients and partial regression coefficients?

asked Apr 8, 2014 at 11:06

15 votes

2 answers

924 views

ICC as expected correlation between two randomly drawn units that are in the same group

asked May 4, 2016 at 9:57

14 votes

2 answers

928 views

What is the Frequentist definition of fixed effects?

asked May 9, 2020 at 11:59

14 votes

2 answers

5k views

Is multilevel modelling simpler, more practical, or more convenient using Bayesian methods or frequentist methods?

asked May 11, 2014 at 6:14

14 votes

1 answer

15k views

Why do we divide by $n-1$ when calculating sample correlation?

asked Aug 10, 2015 at 22:39

13 votes

3 answers

13k views

Is every correlation matrix positive definite?

asked Nov 21, 2015 at 11:33

13 votes

4 answers

750 views

How can I best deal with the effects of markers with differing levels of generosity in grading student papers?

asked Nov 13, 2014 at 1:03

12 votes

1 answer

2k views

Is it a mistaken idea to use standardized coefficients to assess the relative importance of regression predictors?

asked Apr 6, 2016 at 11:48

12 votes

4 answers

9k views

Why is it not OK to do a Pearson correlation on proportion data?

asked Mar 31, 2014 at 3:50

12 votes

2 answers

1k views

When (if ever) is it a good idea to do a post hoc power analysis?

asked Sep 29, 2013 at 12:39

12 votes

2 answers

3k views

Regression: why test normality of overall residuals, instead of residuals conditional on $\hat{y}$?

asked Apr 24, 2016 at 0:30

12 votes

3 answers

985 views

What does it mean for something to have good frequentist properties?

asked Dec 7, 2016 at 6:17

11 votes

3 answers

4k views

Intuition behind the formula for the variance of a sum of two variables

asked May 15, 2018 at 12:42

Stack Exchange Network

267 Questions

How can a distribution have infinite mean and variance?

When is it appropriate to use an improper scoring rule?

Are 50% confidence intervals more robustly estimated than 95% confidence intervals?

Why is the normality of residuals "barely important at all" for the purpose of estimating the regression line?

Why do these statements not follow logically from a 95% CI for the mean?

Why do we say the outcome variable "is regressed on" the predictor(s)?

Can a statistical test return a p-value of zero?

Why does the t-distribution become more normal as sample size increases?

What are Bayesian p-values?

What is the difference between descriptive and inferential statistics?

When doing a t-test for the significance of a regression coefficient, why is the number of degrees of freedom $n-p-1$?

How to choose between the different Adjusted $R^2$ formulas?

Why use a certain measure of forecast error (e.g. MAD) as opposed to another (e.g. MSE)?

Is variance a more fundamental concept than standard deviation?

What's the difference between asymptotic unbiasedness and consistency?

Why is Pearson's ρ only an exhaustive measure of association if the joint distribution is multivariate normal?

What does it mean to explain variance?

What's the difference between regression coefficients and partial regression coefficients?

ICC as expected correlation between two randomly drawn units that are in the same group

What is the Frequentist definition of fixed effects?

Is multilevel modelling simpler, more practical, or more convenient using Bayesian methods or frequentist methods?

Why do we divide by $n-1$ when calculating sample correlation?

Is every correlation matrix positive definite?

How can I best deal with the effects of markers with differing levels of generosity in grading student papers?

Is it a mistaken idea to use standardized coefficients to assess the relative importance of regression predictors?

Why is it not OK to do a Pearson correlation on proportion data?

When (if ever) is it a good idea to do a post hoc power analysis?

Regression: why test normality of overall residuals, instead of residuals conditional on $\hat{y}$?

What does it mean for something to have good frequentist properties?

Intuition behind the formula for the variance of a sum of two variables