国外心理学-设计和分析课件【荐】.pdfVIP

Hypothesis Testing Psychology 3256 Introduction • There are three things we need to say that we have a causal relationship • Temporal precedence • Elimination of alternative explanations • Covariation • the first two can be easily dealt with through design Some made up data Group 1 Group 2 57 72 66 62 60 64 73 79 75 81 Mean 66.2 71.6 sd 7.85 8.56 Question... • Could these two sets of scores come from this same distribution by chance alone? • Well seems pretty likely to me We just made a decision • If the chance is relatively small, usually less than five percent, we say it is unlikely our difference happened by chance alone and we therefore have a statistically significant effect • Without significance we have no covariation of x and y and therefore no causation We set up two, mutually exclusive hypotheses H the null hypothesis • 0 H the alternative hypothesis a • Errors in Hypothesis Testing H True H True • 0 • a Do Not Reject Correct Type II Error H0 Decision Correct Reject H0