Week 4

Question 0

Outlier

Answer 0

Score very different to the rest. Can influence mean and the error.

Question 1

3 contexts of bias

Answer 1

1. Things that bias parameter estimates (inc effect sizes)
2. Things that bias SE & CI
3. Things that bias test statistics and p-values

Question 2

What can bias estimates of parameters and drastically affect sum of errors?

Answer 2

Outliers

Question 3

Assumption of normality

Answer 3

Look at central limit theorem

Question 4

T-test limitations

Answer 4

1. Can only be applied to populations similar to rest sample
2. Sample and pop should be normal I.e. Bell curve
3. Same no of scores

Question 5

How big should sample be to kick in central limit theorem?

Answer 5

At least 30

Question 6

Homogeneity of variance

Answer 6

Audiologist. Spread of scores around mean is roughly the same

Question 7

Heterogeneity of variance

Answer 7

Scores are spread out around the mean. Scores different each occurance.

Question 8

Unbiased heterogeneity and homogeneity of variance

Answer 8

Least method of squares. Or weighted least squares

Question 9

Homogeneity of variance matters when…

Answer 9

U want to look at confidence intervals around model parameter estimates or test model significance

Question 10

P-p plot (probability-probability plot)

Answer 10

Graph to check normality. Plots cumulative probability of a variable against cumulative prob of particular distribution

Question 11

Homogeneity

Answer 11

The same

Question 12

Heterogeneity

Answer 12

Different

Question 13

Variance

Answer 13

Represents the spread of data

Question 14

Tests of normality

Answer 14

S-W test

K-S test