Normality/Homogeneity of Variance Week 5 Flashcards
What does Kurtosis tell us?
The spread of the data around the mean
What is the assumption of homogeneity of variance?
Variability within each group is roughly the same
What is the assumption of normality?
Data comes from a population that is normally distributed
What are the six reasons for conducting exploratory data analysis?
- Checking for data entry errors
- Checking for outliers
- To find patterns that aren’t otherwise obvious
- Checking for and dealing with missing data
- Checking assumptions
- Obtaining a thorough descriptive analysis of the data
What are the measures of central tendency?
Mean, median and mode
What are the measures of variability?
Interquartile range, variance, standard deviation, range
What are the four ways normality is tested?
- Visual inspection of scatterplots and graphs
- Tests of normality
- Inspections of normality plots and detrended plots
- Skewness divided by SE skewness >1.96.
When would you use transformations?
- When homogeneity of variance is violated
- When the sample size is too small or skewed
List the main transformations.
- Log10
- SQRT
- Reciprocal
When would you use a log 10 transformation?
When data is extremely skewed
When would you use a SQRT transformation?
When data is moderately skewed
When would you use a reciprocal transformation?
When your data is shaped like an inverted U, reciprocal will flip the distribution to make it look more normal.
What test tests homogeneity of variance?
Levene’s test
Which options of the four testing for normality are influenced by sample size?
- normality testing
- skewness
When is normality a problem?
When the deviation is strong and the sample is small, below 20.
