Psych 300B: Final Exam Review (Interpreting Effects)

What are the 3 ways to communicate outcomes of multifactorial designs

1. Hypothesis testing - judge based on statistical/practical significance

2. Visual display - patterns of behaviour across level of factor in an image

3. Verbal Description - 3 main categories

What are the 3 main verbal descriptions of a relationship in a multifactorial design

1. Additive vs non-additive

2. Independence vs dependence

3. Difference between differences

When should a graphic figure be used to represent data

When have ≥4 cell means, and if interaction is significant (if not use table)

T or F: only hypothesis testing can confirm the presence of a significant treatment effect

T, graphs and tables only provide a description

How can we test for possible significance just using our cell/marginal means

Compare the individual means, if they are not equal we may not have significance, if they are equal we will not have significance

For a 2 factor design, how many unique statistical outcomes are there

Potentially 8,

For the following questions, let A (a1 and a2) be the row and B (b1, and b2) be the column

What would no main effects or interactions look like in a 2 factor design

means of rows are equal, means of columns are equal and diagonals are equal

a₁ = a₂

b₁ = b₂

a₁b₁ + a₂b₂ = a₁b₂ + a₂b₁

What would row main effects only look like in a 2 factor design

Lines are parallel and fully overlapping but not flat, possible main row effect

a₁ ≠ a₂

b₁ = b₂

a₁b₁ + a₂b₂ = a₁b₂ + a₂b₁

What would column main effects only look like in a 2 factor design

Lines are parallel to each other on the x axis and flat, may be column effect

a₁ = a₂

b₁ ≠ b₂

a₁b₁ + a₂b₂ = a₁b₂ + a₂b₁

What would interaction effects only look like in a 2 factor design

Lines have a point of intersection

a₁ = a₂

b₁ = b₂

a₁b₁ + a₂b₂ ≠ a₁b₂ + a₂b₁

What would two main effects, no interaction look like in a 2 factor design

Line s parallel to each other but different slopes

a₁ ≠ a₂

b₁ ≠ b₂

a₁b₁ + a₂b₂ = a₁b₂ + a₂b₁

What would row main effect and interaction look like in a 2 factor design

Lines will intersect, but slopes are inverted

a₁ ≠ a₂

b₁ = b₂

a₁b₁ + a₂b₂ ≠ a₁b₂ + a₂b₁

What would column main effect and interaction look like in a 2 factor design

Lines intersect, but slopes are inverted

a₁ = a₂

b₁ ≠ b₂

a₁b₁ + a₂b₂ ≠ a₁b₂ + a₂b₁

What would two main effects and interaction look like in a 2 factor design

Lines intersect and have different slopes

a₁ ≠ a₂

b₁ ≠ b₂

a₁b₁ + a₂b₂ ≠ a₁b₂ + a₂b₁

How can we determine if interactions are being hidden by main effects

By calculating eta-squared, allows us to see where variability between groups results from

What does it mean if an interaction is not significant

Changes across levels or conditions are consistent, so effect is generalizable

Effects of factor are additive

Have independence

Diagonal differences = 0

What does it mean if an interaction is significant

The influence on performance by one factor is not consistent across levels of the other factor

Non additive

Dependence

Diagonal differences ≠ 0

What do additive and non-additive mean

Additive - no interaction between two factors, change is uniform

Non-additive - statistically significant interaction between two factors, change accumulates over time

What do independence and dependence mean

Independence - each factor effects behaviour without influencing the other (non-significance)

Dependence - effect of one factor is influenced by the other (significance)