Inhoud
Inferential statistics Test 2 micro lecture summaries .............................................................................. 3
550: Multiple regression addition ....................................................................................................... 3
The effect of a dummy and ratio on a scale variable....................................................................... 3
The effect of two ratio on another scale variable ........................................................................... 4
Studying residuals in a multivariate context ................................................................................... 5
In R ................................................................................................................................................... 6
553 Multiple regression interaction: ................................................................................................... 7
Interaction with a dummy ............................................................................................................... 7
Interaction/moderation with a nominal variable ............................................................................ 8
Interaction/moderation in R ............................................................................................................ 9
554: multiple regression non linearity .............................................................................................. 11
Introducing non linearity ............................................................................................................... 11
Theorising about non-linearity ...................................................................................................... 12
Non linearity in R ........................................................................................................................... 14
560: non normality of residuals and omitted variables..................................................................... 16
A general introduction and overview ............................................................................................ 16
About interpreting QQ plots .......................................................................................................... 17
Testing for normality (Shapiro-wilk test) ....................................................................................... 18
Transforming the y variable ........................................................................................................... 19
Checking and fixing normality in R ................................................................................................ 19
561: heteroscedasticity, non-equal variances and interaction effects .............................................. 21
What is homogeneity of errors? .................................................................................................... 21
Testing whether errors are heterogeneous ................................................................................... 22
Studying homogeneity in R............................................................................................................ 23
563: dealing with outliers, influential cases and multicollinearity .................................................... 24
A general introduction to leverage, outliers, influential cases and Cook’s distance ..................... 24
Finding special cases in your data set by using R .......................................................................... 26
Multi Collinearity ........................................................................................................................... 27
Detecting Multicollinearity ............................................................................................................ 28
510: a non-parametric alternative for means testing: Wilcoxon signed-rank test ............................ 29
Including non-parametric tests ..................................................................................................... 29
Ranks and signed ranks ................................................................................................................. 30
Wilcoxon signed-rank test ............................................................................................................. 30
Wilcoxon in R ................................................................................................................................. 31
,545: a non-parametric alternative for testing the effect of dummy or nominal variables ............... 32
Introducing the Mann Whitney Wilcoxon test .............................................................................. 32
The MWW test in R ....................................................................................................................... 34
Introducing the Kruskal Wallis test ................................................................................................ 35
The Kruskal Wallis test in R ............................................................................................................ 37
580: selecting a sufficient number of cases ...................................................................................... 38
Selecting a sample size .................................................................................................................. 38
Type I and type II error .................................................................................................................. 39
Power analysis ............................................................................................................................... 40
590: choosing between statistical tests – a general overview .......................................................... 42
An overview of statistical test........................................................................................................ 42
,Inferential statistics Test 2 micro lecture summaries
550: Multiple regression addition
The effect of a dummy and ratio on a scale variable
To know whether the error is normally distributed you look at the residuals, they should be normal
and equal across the line and between the groups. If they are not we need to adjust the way we
estimate these relationships.
, the first R-squared and F-test
then b-coefficients and the t-test
The effect of two ratio on another scale variable
If you see an N =… in a table of a study, this means it studies this number of papers, its not the
number of individuals.
We need to check the residuals, but this now becomes less easy using visual inspection. We need to
check the normality and equal variance of the residuals but we cannot simply do this by looking at
Inferential statistics Test 2 micro lecture summaries .............................................................................. 3
550: Multiple regression addition ....................................................................................................... 3
The effect of a dummy and ratio on a scale variable....................................................................... 3
The effect of two ratio on another scale variable ........................................................................... 4
Studying residuals in a multivariate context ................................................................................... 5
In R ................................................................................................................................................... 6
553 Multiple regression interaction: ................................................................................................... 7
Interaction with a dummy ............................................................................................................... 7
Interaction/moderation with a nominal variable ............................................................................ 8
Interaction/moderation in R ............................................................................................................ 9
554: multiple regression non linearity .............................................................................................. 11
Introducing non linearity ............................................................................................................... 11
Theorising about non-linearity ...................................................................................................... 12
Non linearity in R ........................................................................................................................... 14
560: non normality of residuals and omitted variables..................................................................... 16
A general introduction and overview ............................................................................................ 16
About interpreting QQ plots .......................................................................................................... 17
Testing for normality (Shapiro-wilk test) ....................................................................................... 18
Transforming the y variable ........................................................................................................... 19
Checking and fixing normality in R ................................................................................................ 19
561: heteroscedasticity, non-equal variances and interaction effects .............................................. 21
What is homogeneity of errors? .................................................................................................... 21
Testing whether errors are heterogeneous ................................................................................... 22
Studying homogeneity in R............................................................................................................ 23
563: dealing with outliers, influential cases and multicollinearity .................................................... 24
A general introduction to leverage, outliers, influential cases and Cook’s distance ..................... 24
Finding special cases in your data set by using R .......................................................................... 26
Multi Collinearity ........................................................................................................................... 27
Detecting Multicollinearity ............................................................................................................ 28
510: a non-parametric alternative for means testing: Wilcoxon signed-rank test ............................ 29
Including non-parametric tests ..................................................................................................... 29
Ranks and signed ranks ................................................................................................................. 30
Wilcoxon signed-rank test ............................................................................................................. 30
Wilcoxon in R ................................................................................................................................. 31
,545: a non-parametric alternative for testing the effect of dummy or nominal variables ............... 32
Introducing the Mann Whitney Wilcoxon test .............................................................................. 32
The MWW test in R ....................................................................................................................... 34
Introducing the Kruskal Wallis test ................................................................................................ 35
The Kruskal Wallis test in R ............................................................................................................ 37
580: selecting a sufficient number of cases ...................................................................................... 38
Selecting a sample size .................................................................................................................. 38
Type I and type II error .................................................................................................................. 39
Power analysis ............................................................................................................................... 40
590: choosing between statistical tests – a general overview .......................................................... 42
An overview of statistical test........................................................................................................ 42
,Inferential statistics Test 2 micro lecture summaries
550: Multiple regression addition
The effect of a dummy and ratio on a scale variable
To know whether the error is normally distributed you look at the residuals, they should be normal
and equal across the line and between the groups. If they are not we need to adjust the way we
estimate these relationships.
, the first R-squared and F-test
then b-coefficients and the t-test
The effect of two ratio on another scale variable
If you see an N =… in a table of a study, this means it studies this number of papers, its not the
number of individuals.
We need to check the residuals, but this now becomes less easy using visual inspection. We need to
check the normality and equal variance of the residuals but we cannot simply do this by looking at
