HSS 2381 Chapter Notes - Chapter 7: Repeated Measures Design, Null Hypothesis, Squared Deviations From The Mean
10 Dec 2016
School
Department
Course
Professor
Document Summary
Analysis of variance/anova; used to compare 3+ group means when t-tests not appropriate. Samples are used to draw conclusion about populations; statistical inference. Dependent variable: interval or ratio level (ex: heart rate, scores on self-esteem scale) Independent variable: nominal (3 racial/ethnic groups) or ordinal level with small # of categories (normal weight, overweight, obese) The null hypothesis: group means are equal. For example: h0: 1 = 2 = 3 = 4. The alternative hypothesis: at least some of the group means are not equal: Basic assumptions: random sampling from the populations, dependent variable is normally distributed in the populations, variances in the populations are equal. Robustness of anova: robust to violation of normality assumption if n per group > 20, robust to homogeneity assumption if ns are similar. Partitioning variance: isolates (cid:862)reasons(cid:863) (cid:455) ple"s sco(cid:396)es (cid:373)a(cid:455) differ from each other; key anova concept. Individual variability: some of the reasons causing people to differ. Independent variable: another cause of ple differing.
