A survey was conducted in 295 people asking about arthritic pain on a visual analogue scale (can be treated as continuous) in 295 people. Sex was coded 1 = male 0 = female. Medication was coded 1 = on medication 0 = not on medication.

The output from a multiple linear regression computer program is shown in Table 9.8.

Table 9.8 Output from survey of 295 people R-squared = 0.0554 Adj R-squared = 0.0489 pain| Coef. Std. Err. t P>|t| [95% Conf. Interval]

+

sex| -5.285991 3.294272 -1.60 0.110 -11.76952 A medication| -9.489177 3.245583 B 0.004 -15.87688 -3.101475 _cons | 94.18235 5.853297 16.09 0.000 82.66236 105.7024

{adjusted r-squared=1-(1-r-squared)(N-1)/(N-k-1) where N is number of subjects and k is the number of predictors.}

(i) Deduce the values of A and B.

(ii) Is the effect of medication signifi cant? What is the assumption underlying this test?

(iii) What is the expected value of the pain score in a woman not on medication?

(iv) What is the expected value of the pain score for a man on medication?

(v) Is the model a good fi t to the data?