ST500                             Session 5

_Minitab: visit http site (home-page) _review assignment 1 (b) - spss
output _Regression  (Handout SAS Stat syntax sheet) * model building:
forward/backward/stepwise/r**2/cp/all subsets-best * coeffecients/pred
* diagnostics: =var / normal residuals /outliers /
        model specification error/influential obs/collinearity
_sas examples: ../data/regex.sas
            Assignment 2 - regression
email to me before class on Weds, Jan 10:  For sas (.log/.lst) for spss
(.out) and minitab (.LIS) files Please use a2a for sas, a2b for spss and
a2c for minitab analyses.  1. Run the following regression with sas using
'.../data./weight.dat' as data file. This is the same as assign1.dat,
except no -1 for missing data.  proc reg; model weight = height age/ p
cli influence vif;
    output out = resid p = pwt r = rwt student = student;
proc plot data = resid hpercent = 50;
    plot rwt*pwt student*pwt/vpos=20 vref=0;
2. Run the same regression + above diagnostics (influence vif) with
minitab + spss.  (no need to do the output statement + plot and you can
skip 'name' + 'sex' for minitab .)  3. Run one regression program with
only weight=height and the above diagnostics.  4. Looking at the plot
below, 3 new data files were made from weight.dat: ../data/chg1.dat,
../data/chg2.dat and ../data/chg3.dat where the indicated point in
the plot has been moved.  (Note each file has only the one indicated
point moved.)  Rerun the regression in 3. and show/indicate which
coefficients/regresslon diagnostics were affected by the changed data
point.  Why were the statistics affected? (Class discussion)