Make your own free website on Tripod.com

Experimental Psychology

Review, Exam II

 

Chapter 9 – Using Statistics to Answer Questions

            Levels of measurement

            Measures of central tendency (for nominal, ordinal, and interval/ratio sets of data)

            Measures of variability (apply & understand SD and variance)

            Abscissa vs. ordinate axes

            Types of graphs – bar and histogram

            Standard normal distribution (characteristics)

            Normal curve – simulation of all frequency distributions with large enough sample

            Correlational relationships

            Degrees of freedom

            t-test – two uses

            Significance level; 1-tailed vs. 2-tailed tests

            Type I and Type II errors – how to minimize

            Effect size (correlations & t-tests)

           

Chapter 10 – Experiments with 2 Groups

            Structure of design

            Group assignment (random or matched); if neither of these is used, groups are non-equivalent

            Benefit of correlated groups designs - why they reduce error variability; know types (matched,

                        repeated measures, natural sets)

Statistic (t-test or ANOVA) = between groups/within groups (know relationship of between-

groups to within-groups)

            t-test (independent samples and correlated samples) – correlated samples should have positive

                        correlation; requires interval or ratio level of measurement

            Robustness of t-test (assumption of homogeneity)

            Computer output – interpretation; be able to identify the type of test and # participants

            Importance of internal validity; cause/effect relationships

 

Chapter 11 – Experiments with Multiple Groups

            Structure of design – know how to identify names (example: “2 x 3 ind. Grps.” or “one-way”

design)

            Group assignment

            Statistic = between groups/within groups

            One-way ANOVA (completely randomized and repeated measures) – requires interval or ratio

level of measurement

            Problems with repeated measures designs

            Mean squares (SS/df), F ratio (between grps MS / within grps MS), and Post hoc tests (identify

where the difference is)

            Asymptomatic

            Computer output – interpretation