Hypothesis Tests And Confidence Intervals For Normal Linear Models Are Based On The As Sumption Of Normal Errors Recal 1 (214.82 KiB) Viewed 33 times
Hypothesis Tests And Confidence Intervals For Normal Linear Models Are Based On The As Sumption Of Normal Errors Recal 2 (92.16 KiB) Viewed 33 times
Hypothesis tests and confidence intervals for normal linear models are based on the as- sumption of normal errors. Recall that normality can be assessed using a normal Q-Q plot comparing the residuals to "ideal" normal observations. Consider the full model for economic dataset on 50 countries, savings, from the faraway package: > g <- 1m (sr data = savings) . 3 > summary (g) Your tasks:
Produce a normal Q-Q plot of the standardized residuals. Do the residuals look normal? Are histograms and boxplots suitable for checking normality? Explain your answer.
Join a community of subject matter experts. Register for FREE to view solutions, replies, and use search function. Request answer by replying!