PALPSYU3349 Week 6

What are the three main concepts relating to outliers?

• Leverage, Standard Deviation, Distance

• Distance, Leverage, Influence

• Power, Leverage, Standard Deviation

• Gaps, Distance, Influence

Which of the following is incorrect? Omitting an outlier will usually result in:

• Increased R²

• Changed slopes (b's) and y-intercept (α)

• More error (ε)

• More precise (narrower) confidence intervals

If a standardised residual for a specific data point equals - 1.5, what can we tell about point in relation to the regression line?

• The residual is located at y = - 1.5

• The residual is 1.5 standard deviations above the line

• The residual is 1.5 standard deviations below the line

• The data point belonging to the residual is unusual and probably an outlier

What are the two main standardised measures of influence?

• Y values and standardised residuals

• DFFITS and DFBETAS

• Error and distance

• Unstandardised residuals and Pearson's correlation

What is not something you can do when your assumption of normality is violated?

• Use bootstrapping

• Transform your data

• Use non-parametric tests

• Delete your outliers and run your analysis as normal

When transforming your data, what variable do you transform?

• All of the predictors

• The independent variable of interest

• The dependent variable

• All of your variables (DV and IVs)

Which statement about bootstrapping is false?

• Bootstrapping involves redrawing samples from our sample data

• The bootstrapping method makes very specific assumptions about the population distribution

• Using bootstrapping can make us more confident in our results (less probability of having made a type 1 error)

• Bootstrapping is a more preferred approach (to dealing with violations of normality) than transformation