2. Obtain a matrix plot of the data, calculate the correlation matrix, ﬁt the regression,

obtain the ANOVA table, estimates of the parameters and their standard errors etc. 3. Do the diagnostics: plot the observed Y values against the ﬁtted Y-values, plot the

residuals against the independent variables, histogram of residuals, normal probability

plot of residuals etc. 4. If you believe there is some nonlinearity in the data from your analysis in steps 2 and

3, then include the nonlinear terms (such as squares), ﬁt the regression, obtain the

ANOVA table, estimates of the parameters along with their standard errors, plot of

observed against ﬁtted Y values, plot of residuals against the ﬁtted values, histogram of

residuals, normal probability plot etc. [There may be no need to do Box-Cox

transformations if you begin the data analysis after transforming the variables as

suggested] 5. If you believe some variables can be deleted from your model (either from step (2) if

you do not suspect nonlinearity or the model from step (4) if you suspect nonlinearity),

then use all subsets regression (if possible using the computer) and stepwise procedures

for model selection. If you have used both procedures (all subsets and stepwise), then

comment on the differences between the results, if any. 6. Summarize your findings. Brieﬂy discuss if further analysis is needed for this data.

7. Attach all the R codes in an Appendix of the report. Format - Reports should be typed. - The report should include a title page. - The main body of the report should contain no R code.

- All R code should be included in an appendix.