R Outputs

Question 1

Answer 1

two sample t- test

Are the mean ozone levels in gardens A and B equal?

Result: The mean ozone concentrations measured in garden A (3.11 pphm) and garden B (4.98 pphm) differ significantly (t-test, t=-5.3, df = 38, p ≤ 0.001***).

Question 2

What assumptions need to be met to use this?

Answer 2

linear Regression

What assumptions need to be met to use a linear regression?

Critical assumptions behind linear regression:
− Normality of errors = residuals are normally distributed
− Homescedasticity = constancy of variance
− Additionally, there should not be any strong outliers

What test can be used if the assumptions are not met?

• GLM

Question 3

Overview about statistical tests. Wann benutzt man was ?

Answer 3

Classical statistical tests for comparing two means depend on assumptions of normality and variance that have to be checked first.

Question 4

Common data transformations

Answer 4

Question 5

What is Multicollinearity?
What are the problems with Multicollinearity?

Answer 5

Multicollinearity: Correlation between 2 or more predictor variables.
* Interpretational problems: Variable effects can not be separated
* Estimation problems: parameter estimates not stable,
leads to inflated standard errors or coefficients
* Extrapolation problem: When making predictions to new data

Estimation: VIF, problem with multicollinearity when the VIF is close to 10

Question 6

GLM
Overall accuracy
Sensitivity
Specificity

Kappa
AUC

Answer 6

Overall accuracy: a+d/n
Sensitivity: a/a+c
Specificity: d/d+b

Kappa and AUC close to 1 -> excellent

Question 7

a b
c d

Answer 7

Observation
predic. a b
c d

a) true positive
b) false positive
c) false negative (prediction)
d) true negative

10
1
0