/*********************  senicmv97a.sps  **********************/

include 'senicread97.sps'. /* Reads & Labels data
subtitle 'Multivariate regression example'.
Comment: We are interested in the conditions under which hospitals start
 monitoring their patients for hospital-acquired infection. The dependent
 variables are xratio and culratio. The first question is, if we control
 for infection risk, does hospital size (as represented by census & nbeds)
 help predict monitoring as represented by xratio and culratio?

manova xratio culratio census nbeds infrisk /* Variable List
           /  analysis = xratio culratio    /* DV list
           / design = pool(infrisk), pool(census nbeds) .
 /* Each block of vars (in parentheses) is controlled for all other blocks.

Comment: The F-tests for R-square change are exactly the univariate
         Fs from manova, for the pool(census nbeds) block.
regression variables = census nbeds infrisk xratio culratio
           / statistics = default change
           / dependent = xratio culratio
           / enter infrisk
           / enter census nbeds .

Comment: This shows 3 blocks. Each is controlled for the other two.
manova xratio culratio census nbeds infrisk stay /* Variable list
           /  analysis = xratio culratio         /* DV list
           / design = pool(infrisk), pool(census nbeds), pool(stay).