A paper¹ revealed an association between owning a cat as a child and developing schizophrenia later in life. Many people enjoy cats as pets, so this conclusion has profound implications and could change pet ownership habits substantially. However, because of the chance for false positives (Type I errors) and potential problems with generalizability, good scientific conclusions rarely rest on a foundation of just one study. Because of this, significant results often require replication with follow up studies before they are truly trusted. If study results can be replicated, especially in a slightly different setting, they become more trustworthy, and if results can not be replicated, suspicions of a Type I error (significant results by random chance) or a lack of generalizability from the setting of the initial study may arise. In fact, the paper¹ actually provided three different datasets, all from different years (1982, 1992, and 1997) and with different choices for choosing the control group. The sample proportions for each dataset, with the sample sizes in the denominator, are given in the table below. Year 1982 Data (Analyzed in 2015) 1992 Data 1997 Data O Less credibility O More credibility Proportion of schizophrenics who owned cats as children 1075 2125 Round your answers to three decimal places. p-value for 1982 data = i 84 165 p-value for 1992 data = 136 262 =0.506 = 0.509 = 0.519 Proportion of controls who owned cats as children 2065 4847 65 165 220 522 Click here to access StatKey. ¹Torrey, E.F., Simmons, W., Yolken, R.H. (2015). "Is childhood cat ownership a risk factor for schizophrenia later in life?", Schizophrenia Research, June 2015, 165(1):1-2. = 0.426 (a) As we know, statistics vary from sample to sample naturally, so it is not surprising that the sample proportions differ slightly from year to year. However, does the relative consistency of the sample proportions affect the credibility of any single dataset? = 0.394 = 0.421 (b) Use StatKey or other technology to calculate the p-value for each dataset, testing the alternative hypothesis that the proportion of cat owners is higher among schizophrenics.