Inhoudsopgave
Lecture 11: Inference for more than two means........................................................................................... 1
Lecture 12: Inference for Two Way ANOVA................................................................................................ 24
Canvas 12.1: Exercises Two Way ANOVA ................................................................................................... 39
Lecture 13. More about multiple regression and Two-Way ANOVA ............................................................ 47
Lecture 11: Inference for more than two means
Assignment D about regression
- Explained variance = standardized coefficient → coincidence for only simple
regression
- Df can be calculated form the ANOVA table
- In this case of a simple regression, taking the square root of F = t-value b1 (t^2 = F) →
F = general test for whole model
1
,- Based on the residual plot we see there is not a real pattern. There is maybe a little
pattern but we have a small sample. If we had a bigger data set there would be
probably more variation. You can criticize the variation but you could also say the
variation is minimal
- If distribution was more peaked → dots would be closer to the regression line
- Unstandardized data is more related to the values of the data, which would be
better. However standardized shows more easily how many SD you are away from
the mean
2
, - Not meaningful to predict something about one person (individual level).
3
Lecture 11: Inference for more than two means........................................................................................... 1
Lecture 12: Inference for Two Way ANOVA................................................................................................ 24
Canvas 12.1: Exercises Two Way ANOVA ................................................................................................... 39
Lecture 13. More about multiple regression and Two-Way ANOVA ............................................................ 47
Lecture 11: Inference for more than two means
Assignment D about regression
- Explained variance = standardized coefficient → coincidence for only simple
regression
- Df can be calculated form the ANOVA table
- In this case of a simple regression, taking the square root of F = t-value b1 (t^2 = F) →
F = general test for whole model
1
,- Based on the residual plot we see there is not a real pattern. There is maybe a little
pattern but we have a small sample. If we had a bigger data set there would be
probably more variation. You can criticize the variation but you could also say the
variation is minimal
- If distribution was more peaked → dots would be closer to the regression line
- Unstandardized data is more related to the values of the data, which would be
better. However standardized shows more easily how many SD you are away from
the mean
2
, - Not meaningful to predict something about one person (individual level).
3
