PSYC211 Final Exam - Counterbalancing and Chi-Squares

Counterbalancing

A technique designed to evenly distribute the effects of potential confounds across the levels of the independent variable

Complete Counterbalancing

All treatment orders used

Incomplete Counterbalancing

Any condition/treatment appears in each position as often as every other condition/treatmen

Latin-Square Counterbalancing

Construct a square such that any condition/treatment appears once in each order position

Balanced Latin-Square

Ensures that each condition/treatment precedes and follows every other condition/treatment

Balanced Latin-Square FORMULA

BL-S, row 1 = 1, 2, n, 3, n - 1, 4, n - 2, etc…

Quasi-Experimental Design

Independent variables that cannot be controlled

Differential Research Design

Compare scores from two groups for preexisting differences

Posttest-Only Non-Equivalent Control Group Design

Preexisting groups are defined by exposure or no exposure to a treatment that occurred naturally

Time Series Design

Several pre-treatment observations help establish a stable baseline; used to study trends

Chi-Square Test

An inferential statistical test that tests if frequency differences occur due to chance; data must be in nominal form

Observed Frequencies (f_o)

Frequencies you observe in your sample

Expected Frequencies (f_e)

Frequencies you would expect given H₀

Goodness of Fit

Testing how well our observed frequencies (f_o) with the expected frequencies (f_e), given H₀

GoF and df FORMULA

df = C - 1

GoF Effect Size FORMULA

w = √χ²/N

• 0.10 (small effect size)

• 0.25 (medium effect size)

• 0.40 (large effect size)

Chi-Square Solving Steps

f_o

f_e

f_o - f_e

(f_o - f_e)²

(f_o - f_e)²/f_e

∑ (f_o - f_e)²/f_e (add all values)