From the discussed article [1]: "The hypotheses implicitly posit increases or decreases in T levels in relation to an individual’s baseline T. As such, we comp

From the discussed article [1]: "The hypotheses implicitly posit increases or decreases in T levels in relation to an individual’s baseline T. As such, we compared the participants’ two baselines (samples 1 and 6), to assess which of the two would most accurately represent their true baseline T levels, and hence to determine from which baseline we would compare their hormonal changes. Upon arriving to the laboratory, the initial social encounter with the researchers can act as a differential stressor yielding an increase in participants’ T levels. Thus, in such instances the final sample might be a better indicator of an individual’s true baseline than the first (K. E. Wynne-Edwards, personal communication, October 31, 2005). A paired t-test comparing the two baselines confirmed this to be true: the T levels of sample 6 were significantly lower than those from sample 1 (Msample6 = 2.20, SD = .16 vs. Msample1 = 2.26, SD = .21; t-stat = 2.32, df = 30, p < .05). Therefore, to test our hypotheses, we used the baseline value that stemmed from sample 6." So, the interpretation of the experiment can be swung either way: it looks like using the other baseline would have shown that the crappy car depressed testosterone, and the Porsche had no effect. [1] http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6WP2-4WNGD8X-1&_user=209690&_coverDate=11%2F30%2F2009&_rdoc=4&_fmt=high&_orig=browse&_srch=doc-info(%23toc%236978%232009%23998899997%231548116%23FLA%23display%23Volume)&_cdi=6978&_sort=d&_docanchor=&_ct=10&_acct=C000014438&_version=1&_urlVersion=0&_userid=209690&md5=60743f9b036c4b930d75ef8c2c8d3d57 http://www.sciencedirect.com/science?_ob=ArticleURL&_udi...