PSYC3010 – Mediation & Indirect Effects Lecture

0.0(0)
studied byStudied by 0 people
learnLearn
examPractice Test
spaced repetitionSpaced Repetition
heart puzzleMatch
flashcardsFlashcards
Card Sorting

1/54

flashcard set

Earn XP

Description and Tags

A comprehensive set of question-and-answer flashcards covering key definitions, paths, statistical tests, examples, assumptions, and advanced topics related to mediation and indirect effects as presented in the PSYC3010 lecture.

Study Analytics
Name
Mastery
Learn
Test
Matching
Spaced

No study sessions yet.

55 Terms

1
New cards

What does a mediation model seek to explain in regression?

How a predictor (X) exerts an indirect effect on a criterion (Y) through a mediator (M).

2
New cards

In mediation terminology, what is an "indirect effect"?

The influence of X on Y that is transmitted through the mediator M (X → M → Y).

3
New cards

According to Baron & Kenny (1986), what is Path A?

The relationship between the predictor and the mediator (X → M).

4
New cards

According to Baron & Kenny (1986), what is Path C (the total effect)?

The relationship between predictor and criterion before the mediator is entered (X → Y).

5
New cards

What is Path B in traditional mediation analysis?

The relationship between the mediator and the criterion controlling for the predictor (M → Y | X).

6
New cards

What is Path C′ (C-prime) in mediation?

The direct effect of X on Y after M is controlled (X → Y | M).

7
New cards

Which statistical test traditionally evaluates the significance of the indirect effect?

The Sobel Test.

8
New cards

What modern resampling method is often preferred over the Sobel Test for indirect effects?

Bootstrapping (typically 5,000–10,000 resamples).

9
New cards

State the four classic Baron & Kenny conditions for declaring mediation.

(1) Path A significant. (2) Total effect (C) significant. (3) Path B significant. (4) Direct effect (C′) weaker than C, and indirect effect significant.

10
New cards

How does ‘partial mediation’ differ from ‘full mediation’?

In partial mediation, C′ remains significant; in full mediation, C′ is no longer significant once M is included.

11
New cards

What three conditions are considered sufficient in contemporary mediation approaches (e.g., Hayes)?

Significant Path A, significant Path B, and significant indirect effect (X → M → Y).

12
New cards

Which SPSS procedure tests Path A in simple mediation?

A bivariate regression of M on X.

13
New cards

Which SPSS procedure tests Paths C, B, and C′ simultaneously?

A hierarchical multiple regression entering X at Step 1 and M at Step 2 predicting Y.

14
New cards

What does a 95 % bootstrapped confidence interval that excludes zero indicate?

The indirect effect is statistically significant.

15
New cards

In the exercise–self-esteem–happiness example, what served as the mediator?

Self-esteem.

16
New cards

What was the main finding of the Sobel Test in the exercise example?

Exercise had a significant indirect effect on happiness through self-esteem (p ≈ .046).

17
New cards

Define a suppressor variable in regression.

A variable that, when controlled, increases the magnitude of the relationship between a predictor and criterion.

18
New cards

How can a suppressor effect mask a theoretically positive X–Y relation?

Opposing indirect pathways (e.g., X → M → Y negative) cancel out the positive direct effect until M is controlled.

19
New cards

Give an everyday suppressor example mentioned in the lecture.

Conscientiousness positively predicts achievement, but performance anxiety (a negative mediator) can suppress this link.

20
New cards

Differentiate mediation from moderation in one sentence.

Mediation explains how or why X affects Y, whereas moderation indicates when or for whom X affects Y (changes the strength of X–Y).

21
New cards

Does a moderator (W) need to correlate with the predictor (X)?

No, there is no requirement for X–W association in moderation.

22
New cards

What graphical test is typically used to probe moderation but not mediation?

Simple slopes analysis of X on Y at low and high levels of W.

23
New cards

Name two complex extensions of simple mediation covered in the lecture.

Multiple mediation and sequential (serial) mediation.

24
New cards

What analytic family encompasses multiple, serial, and parallel mediations?

Path analysis.

25
New cards

How does structural equation modeling (SEM) extend path analysis?

By using latent variables measured by multiple indicators to account for measurement error.

26
New cards

List the five key assumptions of multiple regression reviewed.

Linearity, normality of errors, homoscedasticity, independence of errors, and lack of multicollinearity.

27
New cards

What is homoscedasticity?

Equal variance of residuals across all levels of the predicted values (Ŷ).

28
New cards

Why can’t simple survey-based mediation establish true causality?

Because variables are not experimentally manipulated; alternative causal orders or third variables may explain associations.

29
New cards

What experimental design helps strengthen causal claims in mediation?

An experimental-causal-chain design manipulating X in Study 1 and M in Study 2.

30
New cards

In the exercise example, which regression step tested the total effect?

Step 1 of the hierarchical regression (X predicting Y alone).

31
New cards

What happens to the X coefficient from Step 1 to Step 2 if partial mediation exists?

It decreases but remains significant.

32
New cards

What value did the bootstrapped indirect effect (IE) have in the example?

IE = .15 with 95 % CI [.03, .34].

33
New cards

Why is multicollinearity problematic in regression?

Highly correlated predictors inflate standard errors, making it hard to isolate individual predictor effects.

34
New cards

Which assumption ensures residuals do not display systematic clustering?

Independence of errors.

35
New cards

What statistical symbol is commonly used for unstandardised regression slopes?

b.

36
New cards

What symbol denotes standardised regression coefficients?

β (beta).

37
New cards

In mediation figures, what does asterisks beside path coefficients denote?

Levels of statistical significance (*p < .05, **p < .01, ***p < .001).

38
New cards

Name two reasons statisticians critique the strict Baron & Kenny steps.

(1) Total effect (C) may be unnecessary; (2) suppressor effects or distal mediators can yield meaningful mediation even when C is non-significant.

39
New cards

What sample size issue often plagues the Sobel Test?

It assumes normality of the indirect effect and can be under-powered in small samples.

40
New cards

How many participants were in the example exercise study?

30 participants.

41
New cards

Which correlation is required to begin testing mediation?

A significant correlation between predictor (X) and mediator (M).

42
New cards

What does Path C′ being “weaker than Path C” indicate?

The mediator accounts for part of the X–Y relationship (evidence of mediation).

43
New cards

Give an example of a positive mediator pathway for social media use → happiness discussed in class.

Increased social connection or social support (open-ended answer acceptable for students).

44
New cards

Give an example of a negative mediator pathway for social media use → happiness.

Upward social comparison leading to lower self-esteem (open-ended answer acceptable).

45
New cards

Why must we report either b or β consistently across paths in mediation?

To maintain interpretational consistency and allow comparison across paths.

46
New cards

What regression output section provides R² changes when adding the mediator?

Model Summary (Change Statistics) in hierarchical regression.

47
New cards

Under what circumstance might mediation be claimed even if X–Y is non-significant?

When indirect effects are significant but opposing pathways suppress the direct relationship.

48
New cards

What is the practical implication of suppressor effects for researchers?

They highlight the need to measure multiple mediators to uncover hidden mechanisms.

49
New cards

What two lecture periods are there no PSYC3010 classes according to the schedule?

Week 3 (no lecture/tutorials) and Week 10 (Labour Day break).

50
New cards

Which statistical program was used throughout the tutorial demonstrations?

SPSS.

51
New cards

Why are partial and semi-partial correlations usually omitted in basic mediation output?

Because mediation interpretation focuses on regression coefficients (b or β), not unique variance partitions.

52
New cards

What does ‘complex mediation’ allow researchers to do beyond simple mediation?

Test multiple mediators simultaneously and compare their indirect effects.

53
New cards

What is the ultimate goal of including figures when reporting mediation?

To visually depict the mediation model with path coefficients for clarity.

54
New cards

What should you do if the regression assumptions are violated?

Consider data transformations, different analytic techniques, or address outliers to satisfy assumptions.

55
New cards

Where can students post PSYC3010 questions outside class time?

On the course Padlet or via the lecturer’s email.