Study Guide for the Midterm Exam

PSY 441 Experimental Design (Spring 2008)

Note: This document is intended as a rough guide to the topics that may be covered on the exam, and make the distinction between topics that should be learned on a conceptual level versus those that are skills-based. All material discussed in lecture or from assigned readings may be included on the exam.

You should understand and be able to define and identify by example:

Independent variable

Dependent variable

Confounding variable

Within-subjects/Between-subjects designs

Ways to control for confounding in an experiment (e.g., by design, by randomization)

Bivalent design

Multivalent design

Factorial design

Sources of variability in data:

- between groups (treatment)

- within groups (error)

Sum of squares

Mean squares

Partitioning variability (“component deviations”)

Degrees of freedom

Probability distribution of F (e.g., why is expected value 1 if null is true?)

Type I error (alpha)

Type II error (beta)

Power (1 - beta)

Assumptions of ANOVA

Familywise error rate

Planned comparisons

Bonferroni adjustment

Weighted sums comparisons

Coefficients of a weighted sums comparison

Orthogonal comparisons

Difference between planned comparison and post hoc tests

Main effects

Interaction effects / Simple effects

Interpretation of SPSS output for ANOVAs, planned comparisons, and t-tests

You should be able to calculate or determine a value for:

SS (within and between)

MS (within and between)

F

df (within and between)

orthogonality

main effects and interaction from condition means

use syntax code to determine simple effects

Sample exam questions:

• Given a set of data, you could be asked to calculate (SPSS) a one-way ANOVA, planned comparison (Bonferroni or weighted sum), a two-way ANOVA, and simple effects analysis
• You could be asked to interpret SPSS output (generated by you or that I generated ahead of time) for the tests listed above
• You could be asked to design an experiment for a particular research question or identify variables (e.g., IVs, DVs, confounding, etc.) within a described study
• Given condition means, you could be asked to evaluate possible main effects and/or describe possible interaction effects
• For a factorial ANOVA; calculate and interpret simple effects

Other sample conceptual (closed-book) questions:

• What aspects of experiments allow causation to be determined?
• Explain why F should be approximately 1.0 when H₀ is true.
• What does it mean for a pair of contrasts to be orthogonal?
• What is the major difference between planned comparisons and post hoc tests? How does this difference affect whether or not the test (i.e., the planned comparison test or the post hoc test) is significant?
• What is the difference between Type I and Type II error? What are some methods for controlling the two types of error?
• Describe the full partitioning of the SS_T term in a two-way ANOVA