202 Chapter 9 . Correlation and Simple Regression TABLE 9.1 Sample Size and Power as a Function of the Population Correlation Coefficient, for Alpha = .05 (Two-Tailed Test) Estimated Population Correlation (p) Power .50 .60 .05 .10 . 15 20 25 30 .35 .40 .45 70 .80 20 499 127 57 32 21 15 12 7 6 5 .30 823 207 92 53 34 24 18 14 11 7 00 O UT D 40 1163 292 130 73 47 32 24 18 15 12 9 NOT UID W .50 1533 383 171 96 61 42 31 24 18 15 11 .60 1956 489 217 122 78 54 39 30 24 19 13 9 .70 2463 614 272 152 97 67 49 37 29 23 16 11 OO 80 3149 785 347 194 123 85 62 47 36 29 19 13 10 90 4200 1047 463 258 164 112 81 61 47 37 25 17 12 The entries in the table indicate the sample size required to achieve the specified power, for the given correlation coefficient.A NP researcher randomly sampled 100 women aged 50-65 years and measured their minutes of exercise in the past week, BMI, and depression. Depression was measured using a Likert type scale consisting of 20 items. The summation score ranged 'orn 20 to 100 and the higher the score, the higher the level of depression. The Pearson correlation coefcients (r's) are summarized in the following table. For the analyses, statistical signicant level was set at u=0.05.| Table l: correlation among minutes of exercise, BMI and depression Exercise in past week BMI minutes Depression score 4130* 0.20 *p *i 0.05