A clinical trial is conducted to evaluate the efficacy of a new medicton to relieve pain in patients undergoing knee replacement surgery. In the trial patients are randomly assigned to receive either the new medication  or the standard medication.  After receiving the assigned medication, patients are asked to report thier pain on a scal of 0 -100 with higher scores indicative of more pain. Data on the primary outcome is:                 sample size                    Mean pain score                     Std Dev of pain score

  New medicaiton          60                                       30.31                                        7.52

 Standard Medication     60                                      53.85                                        7.44  

Because procedures can be more complicated in older patients, the investigators were concerned about confounding by age. For analysis patients wre classified into two age groups, less 65 yrs old and 65 yrs of age and older. The data is   Age< 65    sample size       mean pain score      std dev of pain

                       new medication                     40                         25.30                 2.46                                    Standard med                       25                         45.51                 1.83

                         Total                                    65                         33.07                 10.16

                        +65                                       20                         40.33                  2.16

                                                                      35                         59.80                  2.49

is age a confounding factor?Is there an effect modification by age using a multiple linear regression analysis? Is there a statistcally significant difference in mean pain scores between patiens assigned to the new meds as compared to the standarad meds?  run α = 0.05