# Failures in Assumptions Flashcards Preview

## PS3021 Stats > Failures in Assumptions > Flashcards

Flashcards in Failures in Assumptions Deck (18)
1
Q

Why do assumptions matter in regard to inferential stats with non-normal distributions?

A

We assume there is a normal distribution, so any estimate of the p value is wrong.

2
Q

What is skewness a measure of?

A

The symmetry of a distribution.

3
Q

What is Kurtosis a measure of?

A

The “peakedness.”

4
Q

What is the third moment about the mean?

A

Skewness.

5
Q

What is the fourth moment about the mean?

A

Kurtosis.

6
Q

What inferential test of normality is appropriate for large sample sizes (2000+)?

A

Jarque-Bera test.

7
Q

What inferential test of normality is appropriate for small sample sizes?

A

Kolmorogov-Smirnoff test.

8
Q

What does the one sample case of Kolmorogov-Smirnov compare?

A

The sample distribution against H0/the normal distribution.

9
Q

What does the two sample case of Kolmorogov-Smirnov compare?

A

Two distributions and whether they come from the same population.

10
Q

Give one strength of the one sample case of Kolmogorov-Smirnov.

A

It is good for checking normality: calculate sample mean and SD and compare to normal distribution.

11
Q

What does Shapiro-Wilk examine?

A

The ratio of two estimates of variance: usually best estimator and corrected SS.

12
Q

What is a quantile?

A

A regular interval of a distribution.

13
Q

Describe a Q-Q plot for 2 data sets with equal n.

A

Use n quantiles, order values, and plot like a scatter plot.

14
Q

Describe a Q-Q plot for comparing with a normal distribution.

A

Use n quantiles, estimate sample mean and variance and plot days against the inverse of the cumulative for each quantile.

15
Q

What does Levene’s test examine?

A

Absolute differences from the mean.

16
Q

What characterises the SQRT(X) transform?

A

Removes linear mean-variance and skew-positive nature.

17
Q

What characterises the In(X) transform?

A

Removes mean-SD relationship and strongly reduced skew-positive nature.

18
Q

What characterises the 1/X transform?

A

It it the most extreme and causes the strongest reduction in skew-positive nature.