9. Mediation and Moderation Flashcards
Mediation occurs when the r____ between a p____ and an o____ can be explained by the r____ not a t____ variable: the mediator
relationship
predictor, outcome
relationship, third
Mediation is interested in the p____ between variables
pathways
Mediation doesn’t imply causation without t____ d____ pathways. However, mediation can be used in c____ and e____ designs
theoretically driven
correlational, experimental
Mediation pathways start with the t____ e____, which is the s____ r____ between predictor and outcome and doesn’t adjust for any other variables
total effect
simple regression
The total effect can be broken down into:
1. The i____ effect = the effect of the p____ on the o____, through the m____
2. The d____ effect = the effect of the p____ on the o____, after a____for the m____
- Indirect, predictor, outcome, mediator
- Direct, predictor, outcome, adjusting, mediator
Total effect = i____ effect (path ab) + d____ effect (path c)
indirect, direct
In theory, we could test step 2a (predictor –> mediator) with a simple regression. However, for step 2b (mediator –> outcome) we need to use a m____ regression where we adjust for our predictor. This is because our mediator and the outcome could be r____just because they are both r____ to the predictor.
multiple, related, related
Step 3 of mediation is checking whether our predictor still predicts outcome path c (d____ effect) after accounting for the i____ effect
direct, indirect
We can interpret the bs in the same way we would with a linear model:
1. Path __ tells us the effect for the predictor on the mediator
2. Path __ tells us the effect for the mediator on the outcome a____ for the predictor
3. Path __ (d____ effect) tells us the effect for the predictor on the outcome adjusting for the mediator
4. The i____ effect tells us w____ there is a mediator
5. The t____ effect tells us the effect of the predictor on the outcome, n____ a____ for the mediator
- a
- b, adjusting
- c, direct
- indirect, whether
- total, not adjusting
To see if we have a significant mediation, we look at the s____, the c____ i____ and the __-value of the indirect effect. If the c____ i____ do not contain __ and if the __-value is less than __, then we can say there is a m____ e____.
size, confidence intervals, p-value
confidence intervals, 0, p-value, .05, mediation effect
We can have different types of mediation:
1. P____ mediation - when the direct effect is r____ but still s____ - there’s both an i____ and d____ effect of the predictor on the outcome
2. F____ mediation - when the direct effect is reduced to n____-s____ - the effect of the predictor on the outcome goes e____ through the mediator
- Partial, reduced, significant, indirect, direct
- Full, non-significant, entirely
A moderator is a v____ that affects the r____ between two others. We can investigate whether the effect of our predictor on our outcome is the same for a____ p____ or whether It differs under different c____ depending on the value of the moderator.
variable, relationship
all people, conditions
Differences in moderation could be the p____ of an effect, the s____/s____ of the effect, or the d____ of the effect
presence, size/strength. direction
Moderation can be used in c____ or e____ designs with c____ or c____ variables
BUT.. Remember correlation does not imply c____. This holds in moderation as well.
correlational, experimental, continuous, categorical, causation
We run a moderation in the same way as a l____ m____ but with __ predictors
We have two m____ effects (i.e., ‘lower-order effects’), and one i____ effect (i.e., ‘higher-order effects’) or in our case the moderation.
linear model, 3, main, interaction
When using a linear model with multiple predictors normally, each of the individual effects (bs) are interpreted when other variables in the model are __
But for some variables a value of __ just doesn’t make s____.
In that case the interaction term in our moderation analysis means that the bs for the main effects are usually u____.
This is where the first step in the moderation process comes in.
0
0, sense
uninterpretable
Moderation process - for c____ variables only, we centre them by transforming them into d____ around a f____ point. In grand m____ centering, this ‘f____ point’ is the o____ m____ of that variable. We can then interpret our i____ effects at average levels of the other variable rather than __
continuous, deviations, fixed
mean, fixed, overall mean
individual, 0
Interaction and moderation have only a few key differences:
1. Interactions are mostly used with t____ c____ p____, while moderations are used with at least o____ c____ p____
2. Conceptually, in interactions both i____ variables are p____ but in moderations one variable is the p____ d____ the effect and the other is the m____ m____ that effect.
- two categorical predictors, one continuous predictor
- independent, predictors, predictor driving, moderator modifying
A model summary in moderation process tells us whether we have s____ m____ e____ of our predictor and moderator, and also if we have an i____ e____ (a moderation)
significant main effects
interaction effect
The Johnson-Neyman interval shows us a ‘z____ of s____’
zone of significance
The simple slopes analysis examines the r____ between predictor and outcome at l____, a____ and h____ values of the moderator (if c____)
relationship, low, average, high, continuous
The Interaction Plot v____ the simple slopes analysis
visualises
The differences with categorical moderators:
1. We can’t g____ m____ centre the moderator (but we should with our predictor)
2. We can’t get a J____ N____ Interval
3. Our simple slopes isn’t looking at l____, m____ and h____ values of our moderator anymore, and instead will use what g____ exist in our moderator
- grand mean
- Johnson Neyman
- low, mean, high, groups
