PSYC 218 Chapter 15: Chapter 15.docx
Document Summary
Degrees f freedom for tests of no difference between dependent means: when the score of one member of the pair is specified, that of the other is not completely free to vary, so df will be n (number of pairs) 1, e. g. you have 20 subjects i. e. 10 pairs, the df = 9. Assumptions when testing a hypothesis about the difference between two dependent means: the assumption of homogeneity of variance is not required for the test between two dependent means, the hypothesis testing assumes random sampling with replacement, assumes a normal distribution of x y or d, central limit theorem sample of larger than 25 will approximate a normal distribution even if it is skewed, if you have a sample smaller than 25 and you believe the population may not be normal then you use non parametric statistical test for t.
