SOME USEFUL FORMULAS

including the correlation formula that I use that is not in Gravetter & Wallnau:

Population: mean =  , variance = 2 , standard deviation =  , size = N

Population Variance = 2 = SS/N = (X-)2 / N

standard deviation = variance

Sample: mean = M , variance = s2 , standard deviation = s , size = n (or N), df = n-1 (or N-1)

Sum of Squares (of deviations from the mean): SS = (X-M)2 = (X-M)(X-M)

Variance: s2 = SS/df = (X-M)2 / (n-1)

Correlation (Pearson product-moment correlation coefficient): r = covxy / sxsy

Covariance: covxy = SP/(n-1) = (X-Mx)(Y-My) / (n-1)

Sum of Products (of deviations from the mean): SP = (X-Mx)(Y-My)

z = (X-) /  [for individual scores], or informally for descriptive purposes z = (X-M) / s

z = (M-M) / M = (M-) / M [for sample means] where M =  / n

t(df) = (M-M) / sM = (M-) / sM where sM = s / n

df = n-1

t(df) means "t on these df" NOT "t TIMES df"; e.g. t(24) means t on 24 df, not t*24

t(df).05 or t.05(df) means "the .05 cutoff value for t on these df"

Cohen's d = (M-) /  [for the population, with d sometimes written as Greek ]

Cohen's d = (M-) / s [for the sample]

r2 = t2 / (t2 + df)

r2 = R2 = 2 (eta-squared) [but NOT = 2 (omega-squared) as G&W p. 299 says, because that's calculated differently]

Probability Density Function for the Normal distribution: y = [1 / (22)] * e^[-(x-)2 / 22]

-know how to integrate that expression to arrive at various areas under the curve between given limits

-yes of course I'm kidding

Assume all tests are two-tailed using  = .05 unless the question states otherwise.