User Thomas Bilach

389 votes

9 answers

765k views

What is the difference between fixed effect, random effect and mixed effect models?

Andrew

5,668

asked Nov 19, 2010 at 0:03

283 votes

15 answers

532k views

What is the meaning of p values and t values in statistical tests?

hypothesis-testing p-value interpretation intuition faq

Sharpie

4,196

asked Jul 19, 2010 at 19:28

269 votes

10 answers

154k views

How would you explain covariance to someone who understands only the mean?

variance covariance intuition teaching

PhD

13.7k

asked Nov 7, 2011 at 19:41

224 votes

4 answers

418k views

How to interpret a QQ plot

r data-visualization inference qq-plot faq

JohnK

18.7k

asked Jun 5, 2014 at 10:44

207 votes

8 answers

464k views

In linear regression, when is it appropriate to use the log of an independent variable instead of the actual values?

regression distributions data-transformation logarithm faq

d_2

2,221

asked Jul 20, 2010 at 13:11

158 votes

3 answers

185k views

How are the standard errors of coefficients calculated in a regression?

r regression standard-error lm

ako

1,683

asked Dec 1, 2012 at 10:16

115 votes

21 answers

43k views

What's a real-world example of "overfitting"?

overfitting

user3851283

307

asked Dec 11, 2014 at 6:28

87 votes

14 answers

66k views

What is the meaning of "All models are wrong, but some are useful"

modeling

gpuguy

1,083

asked Apr 27, 2013 at 8:39

78 votes

7 answers

10k views

Do not vote, one vote will not reverse election results. What is wrong with this reasoning?

probability game-theory elections voting-system

Przemyslaw Remin

1,148

asked Feb 28, 2020 at 14:06

76 votes

6 answers

107k views

Why is it that natural log changes are percentage changes? What is about logs that makes this so?

regression logarithm mathematical-statistics

thewhitetie

967

asked Nov 4, 2016 at 15:07

74 votes

15 answers

11k views

Why would parametric statistics ever be preferred over nonparametric?

regression hypothesis-testing mathematical-statistics estimation nonparametric

en1

907

asked Jul 30, 2015 at 11:48

68 votes

2 answers

23k views

Is there a difference between 'controlling for' and 'ignoring' other variables in multiple regression?

regression multiple-regression

Siddharth Gopi

1,435

asked Dec 7, 2013 at 2:14

66 votes

4 answers

69k views

Does it make sense to add a quadratic term but not the linear term to a model?

regression polynomial

Henrik

13.4k

asked May 18, 2012 at 13:34

36 votes

1 answer

4k views

How exactly is the "effectiveness" in the Moderna and Pfizer vaccine trials estimated?

probability relative-risk

Not a statistician

363

asked Nov 16, 2020 at 23:58

30 votes

10 answers

3k views

What are the myths associated with linear regression, data transformations?

regression distributions normal-distribution data-transformation

Ritik P. Nayak

93

asked Feb 5 at 1:44

29 votes

3 answers

17k views

What does "all else equal" mean in multiple regression?

multiple-regression interpretation least-squares regression-coefficients controlling-for-a-variable

EconStats

835

asked Feb 3, 2014 at 20:08

28 votes

7 answers

60k views

Why is gender typically coded 0/1 rather than 1/2, for example?

data-transformation binary-data categorical-encoding units

Adhesh Josh

3,055

asked Oct 7, 2011 at 19:46

24 votes

3 answers

21k views

t-test on highly skewed data

t-test skewness permutation-test

Chris

585

asked Sep 12, 2013 at 20:53

23 votes

2 answers

3k views

R-squared is equal to 81% means what?

regression machine-learning r-squared

F.C. Akhi

635

asked Jul 23, 2020 at 18:00

21 votes

5 answers

6k views

When are Bayesian methods preferable to Frequentist?

bayesian frequentist

HFBrowning

1,136

asked Jun 17, 2014 at 0:24

19 votes

3 answers

12k views

Utility of the Frisch-Waugh theorem

regression econometrics least-squares projection decomposition

Anthony Martin

1,149

asked Mar 15, 2016 at 23:28

18 votes

4 answers

102k views

Split data into N equal groups

r distributions

Rajpal Kulhari

181

asked Mar 30, 2015 at 18:48

18 votes

5 answers

16k views

Fixed effect vs random effect when all possibilities are included in a mixed effects model

mixed-model random-effects-model fixed-effects-model many-categories

gung - Reinstate Monica

136k

asked Oct 21, 2014 at 19:01

18 votes

7 answers

5k views

Why do we worry about overfitting even if "all models are wrong"?

modeling overfitting

Gillian

361

asked Oct 31, 2019 at 13:27

16 votes

2 answers

14k views

P-values equal to 0 in permutation test

p-value permutation-test

user2886545

261

asked Jul 24, 2014 at 13:53

16 votes

1 answer

12k views

When is it necessary to include the lag of the dependent variable in a regression model and which lag?

autocorrelation panel-data fixed-effects-model trend lags

Mauricio Tec

273

asked Sep 9, 2013 at 15:23

15 votes

3 answers

23k views

Difference-in-differences with individual level panel data

econometrics panel-data fixed-effects-model difference-in-difference

greg

208

asked Aug 15, 2016 at 21:29

15 votes

3 answers

996 views

Which OLS assumptions are colliders violating?

least-squares causality assumptions dag

robertspierre

1,200

asked May 21, 2021 at 21:42

15 votes

7 answers

2k views

Why the p-value of t.test() is not statistically significant when mean values look really different

t-test mean p-value wilcoxon-mann-whitney-test group-differences

arshad

841

asked Apr 6, 2021 at 6:34

15 votes

2 answers

2k views

Significance test for two groups with dichotomous variable

statistical-significance t-test

Art

253

asked Mar 25, 2021 at 21:05

Stack Exchange Network

current community

your communities

more stack exchange communities

70 Bookmarks

What is the difference between fixed effect, random effect and mixed effect models?

What is the meaning of p values and t values in statistical tests?

How would you explain covariance to someone who understands only the mean?

How to interpret a QQ plot

In linear regression, when is it appropriate to use the log of an independent variable instead of the actual values?

How are the standard errors of coefficients calculated in a regression?

What's a real-world example of "overfitting"?

What is the meaning of "All models are wrong, but some are useful"

Do not vote, one vote will not reverse election results. What is wrong with this reasoning?

Why is it that natural log changes are percentage changes? What is about logs that makes this so?

Why would parametric statistics ever be preferred over nonparametric?

Is there a difference between 'controlling for' and 'ignoring' other variables in multiple regression?

Does it make sense to add a quadratic term but not the linear term to a model?

How exactly is the "effectiveness" in the Moderna and Pfizer vaccine trials estimated?

What are the myths associated with linear regression, data transformations?

What does "all else equal" mean in multiple regression?

Why is gender typically coded 0/1 rather than 1/2, for example?

t-test on highly skewed data

R-squared is equal to 81% means what?

When are Bayesian methods preferable to Frequentist?

Utility of the Frisch-Waugh theorem

Split data into N equal groups

Fixed effect vs random effect when all possibilities are included in a mixed effects model

Why do we worry about overfitting even if "all models are wrong"?

P-values equal to 0 in permutation test

When is it necessary to include the lag of the dependent variable in a regression model and which lag?

Difference-in-differences with individual level panel data

Which OLS assumptions are colliders violating?

Why the p-value of t.test() is not statistically significant when mean values look really different

Significance test for two groups with dichotomous variable