Name: Score: 1. Suppose you are planning an experiment for a class project with a group of students and you have reached the point at which you need to decide how many subjects you will assign to each of the treatment conditions. Let's suppose that your group has decided to study the effects of various background noises on reading comprehension-specifically: silence, soft recorded conversation, light classical music, and musical selections recorded from a local "top-40" radio station. Your group expects to find a "large" effect as a result of these manipulations. A. Estimate the sample size you will need to achieve a power of .90, using a 5% level of significance. B. Suppose that you feel a "large" effect is not a realistic estimate and argue for using a "medium" effect to estimate sample size. What sample size will your group need now to achieve the same degree of power? C. Let's assume that the sample size calculated in part (B) is simply not possible to achieve by your group given the time available to complete the study. After much argument, a compromise is reached in which the group decides to set the estimated effect size equal to .10. What sample size do you now need? D. Some members of your group still feel that steps could be taken to reduce the estimated sample size some more. They suggest that the group could live with a small decrease in power--from .90 to .80. What sample size would you now need, still using an estimated effect size of .10? n'= $2 (1-A2) where n' is the estimated sample size we wish to determine, the quantity (phi) is a statistic that we will @A of the treatment effects. obtain from Table A-6 on p. 538, and @A2 is our estimate of the size If we use our example of the three different amounts of coffee, a-3 groups, so we would use the curve labeled 2. For this problem, let's assume that we want to set power = .80 and we are expecting a medium effect size. n' = (1.8)2 (1-.06) (3.24) (.94) = (3.24)(15.67) = 50.77 (.06) (.06) This means that we will need at least 51 subjects per condition to achieve power = .80 when the anticipated results are of medium size. Always round up to the next whole number since we are estimating number of people. Using Power to Estimate Sample Size Yayan to do. A small effect produces an estimated omega squared of at least .01. A medium effect produces an estimated omega squared of at least .06. A large effect produces an estimated omega squared of .15 or greater. dpw. Video 3d sample 1.00 8 .90 .60 Values of Power 60 To 80 50 24 12 9 8 7 6 5 4 3 2 df effect de the 40 80 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 26 Values of