PSYC 3480 Lecture Notes - Lecture 2: Thalidomide, Sampling Bias, Random Assignment
Document Summary
You can generalize it beyond the sample, to the population. It measures what it is supposed to measure. A drug company made a hormone medication for women going through menopause. The d(cid:396)ug (cid:374)ot o(cid:374)l(cid:455) helped (cid:449)o(cid:373)e(cid:374)"s s(cid:455)(cid:373)pto(cid:373)s, (cid:271)ut it also helped thei(cid:396) hea(cid:396)ts (became a huge success, and a lot of women were taking it just for their hearts) An international study was conducted on this same drug, and the results showed that it should only be given to women with severe symptoms, and not to better their heart. There was a biased sample for the first study because the women were younger and healthier to begin with, so it was obvious why they would have had better heart results. The second study used random assignment and a more variable population of women (age, ethnicity) When the study was conducted properly, they found that the drug actually worsened heart conditions.