final quiz rm ch 9, 10, and 12 Flashcards
Reviewing three Casual Claims
Covariance
Temporal Precedence
Internal Validity
multivariate designs
longitudinal designs help address temporal precedence
multiple regression analyses help address internal validity
Longitudinal designs
measure the same variable in the same group of people at different times
cross sectional
two variables measured at the same point in time are correlated
cross-lag correlations
the earlier measure of one variable is associated with a later measure of another variable (looking at how people change over time) – help us establish temporal precedence
Regression results indicate whether a third variable explains the relationship
Criterion variables and predictor variables
- Using beta to test for third variables
- Predictor is independent variable
- In a multiple regression analysis, you will be studying three or more variables – you choose the variable you want to understand and that is called your criterion or dependent variable
- Where you see multiple regression output you will see a beta – in this example you will have one beta for exposure to sex on tv and another beta for age
- R just shares that there is a relationship between variables
- There is a negative relationship between the predictor variable and the criterion variable when other variables are constant
- No relationship when other predictors are controlled for
Smaller the beta the weaker the relationship and the bigger the beta the stronger the relationship (B will be what you see in graphs) – B is unstandardized coefficient and beta is standardized coefficient
B is just like beta and we think about the positives and negatives in the same way
o We can not compare two B values in the same table
o But we can compare two beta values in the same table since one is standardized and one is unstandardized
o When we think about B – unstandardized, Beta = standardized
o Beta and B tell us information similar to what R tells us
criterion
dependent
Using Beta to test for third variables continued
95% Cis and statistical significance of Beta
- Your P value of .05 is complementing your .95 from CI
- When the P is greater than .05 the associated CI does include zero
- When the P value is less than .05 the associated CI does not include zero
Higher beta, smaller beta
stronger relationship, weaker relationship
Betas are
helping us answer our research questions and is a standardized coefficient that we can compare in one table – we can’t compare beta’s between two different tables
regression in pop media articles
Controlled for
- Adjusting for
- Considering
Regression does not establish causation
Multiple regression – is not foolproof way to rule out all kinds of third variables
- We can’t control variables we don’t measure
- There could be some variables we didn’t think about measuring
Mediators versus third variables
Similarities
Both involve multivariate designs
Both can be detected using multiple regression
Differences
Third variables are external to the bivariate correlation (problematic)
Mediators are internal to the causal variable (not problematic).
Mediators versus moderators
Mediator- Why?
Moderator- For whom and when does this relationship exist?
Covariance
it’s about the results and differences between groups
Control groups, treatment groups, and comparison groups
Control group- no treatment condition (not a control variable). This condition has no treatment and compare it to one with a treatment
Treatment group- One or more treatment conditions
Placebo group- control group exposed to a treatment that is inert
Well-designed experiments establish internal validity
Design confounds- When a second variable varies systematically along with the independent variable and it gives us an alternative explanation for our results. Unable to support a causal claim. Only threatened if there is systematic variability which can’t be controlled
Ex: Students in the laptop group answer more difficult questions than the long-hand group. Can’t tell the cause because of a difference in the dependent variable
Selection effects- When the participants in one level of the IV are systematically different than the participants in another level of the IV
Random assignment avoids selection effects
Matched groups avoid selection effects. Ex: Matching IQ
Independent-groups design (between-subjects designs)
Different groups of participants that are placed at different levels of your IV
Posttest only
Equivalent groups. Simplest type of independent groups experiment where participants are randomly assigned to IV groups and are tested on your DV one time
Pre-test Post-test design
Participants are randomly to at least two different groups and are tested on the key DV twice (before and after)
Within-groups design
Each participant is presented with all levels of the IV
Repeated-measures design: Participants measured on the dependent variable more than one time after exposure to each level of the independent variable
Ex: Group tastes chocolate with confederate and say it tastes better than trying the same chocolate alone
Concurrent-measures Design- All participants are exposed to all levels of the independent variable are roughly the same time
Advantages of within-groups design
Participants in your groups are equivalent because they are the same participants and serve as their own controls
Within-groups designs require fewer participants than other designs.
Order effects:
When exposure to one level of the independent variable influences reactions to other levels of the independent variable. Applies to any design
Practice effect: When our participants are either getting better or worse at a task from practice or fatigue. (also known as fatigue effect)
Carryover effect: When one condition carries over to the next
Avoiding order effects
Two types of counterbalancing
Presenting the levels of independent variables in different orders
Full counterbalancing: all possible condition orders are presented
Partial counterbalancing; Only some of the possible condition orders are used
What a mediating variable it is, what question your mediator asks?
Think about regression analysis, or what it means when you add extra variables to your regression analysis.
all coefficients are changed
Be able to understand what a matched group design is
What are things you lok for in multiple regression analysis? What values?
Coefficient estimate for the regression. Tell us what the correlation coeffeicei
This particular variables significantly predicted above and beyond all of the other variables there
Or does not
TYPES OF BETA VALUES:
Standardized betas: compare the values in the tables compare values with all the IV. Cannot compare standardized betas with other standardized betas in other tables Two different tables
Unstandardized betas: we know the relationship between the DV and IV but you can not compare these in a table. You can not compare those.
Standardized betas
compare the values in the tables compare values with all the IV. Cannot compare standardized betas with other standardized betas in other tables Two different tables
Unstandardized betas
we know the relationship between the DV and IV but you can not compare these in a table. You can not compare those. tell us if the IV is a significant predictor of the dependent variable
What is a selection effect?
When the participant in an IV are systematically different than the participants in another level of the independent variable
Avoiding selection effects with random assignment
there should be no systematic differences between the groups
Avoiding selection effects with matched groups
take two people and match them on a particular characteristic and put them in two equal groups
Interaction
the effect of one independent variable depends on the level of the other independent variable
Crossover interaction
it depends, It depends on.
Spreading interaction
“especially when”
Deliciousness of sammie vs presence of bacon
Factorial designs Study
Two or more IVs, two or more factors
Cell: number of IVs will impact the number of cells we have, one the phone and not on phone, young drivers vs old drivers, 4 cells, 2 factors
Participant variable: can’t change the variable. IV without being a true IV
Ex: age
Study either manipulated or participant variables
Repeated measures design
One group à taste chocolate with confederateà rate chocolate -> taste chocolate alone à rate chocolate (confederate – someone apart of research study but participant doesn’t know that)
Participants measured on the dependent variable more than one time after exposure to each level of the independent variable
Concurrent measures design
o Where all of your participants are exposed to all levels of independent variable at the same time
o One group – exposed to male and female faces and then dependent variable is who they looked at longest
Advantages of within groups designs
Participants in your groups are equivalent because they are the same participants and serve as their own controls.
2. Within groups designs require fewer participants than other designs
Internal validity: controlling for order effects
Exposure to one level of independent variable influences reactions to other levels of independent variables
- When being exposed to one condition affects how participants respond to other conditions
Practice effects
one type of order effect when our participants are either getting better at a task because they have practiced it or they are getting worse at a task because of fatigue (also known as fatigue effect)
Carryover effect
Some sort of contamination carrying over from one condition to the next.
Factorial designs study two independent variables
Cell
Participant variable
o Ex. On cell phone, not on cell phone + younger drives, older drivers à younger drives on cell phones, younger drives not on phones OR older drivers on cell phones, older drivers not on phones
o We use our factorial designs to study independent (manipulated) variables or participant variables
o Some variables you cannot manipulate – such as age – which then are considered participant variables
Main effect
is there an overall difference?
o When you are thinking about the main effect you think about the overall effect of one independent variable on your dependent variable, averaging other levels of the other independent variable
overall effect of one of those IVs on the DV
We compare our marginal means and if there is a significant difference between the marginal means there is a main effect
Confidence interval that includes zero
not significant because it indicates no relationship
2 by 2
means there are 2 independent variables but 2 levels of independent variables
2 by 1
means we have 2 independent variables (there are two things on either side of the “by”) and 2 levels of the first independent and 1 level for the second independent variable
3 by 3 by 3
3 independent variables and each has 3 levels
2 by 3 by 4
3 independent variables
one has level 2, one has level 3 and one has level 4 \
ndependent – groups factorial designs
Both IVs are studied as independent groups
- A 2 x2 independent groups factorial design has four groups or cells
- This is known as a between subject’s design
Interactions: Is there a difference in differences? Options 1 and 2
We need a p value; confidence interval corresponding to p value and get other information to be able to differentiate differences in tables
- If it is statistically significant, we can identify that by the table
Interactions: is there a different in differences? Line and bar graphs
When lines are parallel there is no interaction but when they cross it is statistically significant because they overlap
- Interactions will either be a crossover interaction or spreading interaction
- Interactions from bar graphs – connect the bars with a line to determine the interactions
Interactions are more important than main effects
When a study shows both a main effect and an interaction, the interaction is almost always more important
- There may be real differences in marginal means, but the more exciting part is the interaction
Independent – groups factorial designs
Both IVs are studied as independent groups
- A 2x2 independent groups factorial design has four groups or cells
BLANK BY BLANK, how many variables, how many levels are in each
Ex: cell phone 2 x 3
Do we increase main effects if we increase levels? NO # of main effects is not affected by levels but by variables
How to figure out how many cells? Multiple the levels
PUT SMALLER NUMBER IN THE FRONT
2x2x2 : THREE WAY DESIGN
3 variables: with two levels each
Two age conditions
Two traffic conditions
Two cell phone conditions
the difference of differences of differences
INcreasing the number of Levels of an Independent Variable
Blank X black
Main effects and interactions from a three-way design
Why worry about all these interactions?
Adding the third one to
Within-groups factorial design
All the people are participating in every level of the independent variable
- Needs less participants than independent groups
Mixed factorial designs
One IV is manipulated as independent-groups and the other is manipulated within-groups
- This design is intermediate between the within groups design and the independent groups design in terms of number of participants
MOST OUTCOMES IN PSYCHOLOGY STUDIES
ARE NOT MAIN EFFECTS BUT RATHER INTERACTIONS
Adding more variables to congressional analysis?
TO account for as many third variables as possible all at once
Indicate an experiment was done?
One of the variable is manipulated control group comparison group
Going to the mall ask about positive attitudes with shopping?
Study, questionnaire no comparison group
Convenience sampling
Dependent came before iv
Independent groups’ design is known as what?
Between-subjects design: REQUIRES MORE PARTICIPANTS
NEED DIFFERENT PEOPLE FOR EACH LEVEL OF IV
Everyone experiences different conditions
COMPARED TO WITHIN GROUPS
Same groups experiences all the same levels of the IV
CONCURRENT MEASURES DESIGN
Participant is experiencing all levels at the same time
Monkey’s studies HArry Harlow’s
Baby monkey choose wire monkey or cloth monkey
Monkey’s preference
COKE VS PEPSI
MEDIATING VARIABLE
Telling you why this variable impacts those other variable why IV impacts DV
Why the relationship exists
I don’t know about this
YOu have variable A and the relationship between C above A and C you have variable B. B needs to have an arrow to A and C. B is the mediator.
Think about regression analysis, or what it
means when you add extra variables to your regression analysis.
changes the size or sign of the coefficients
Within groups study between group studies which one of those would be affected by practice effects
WITHIN GROUPS WOULD BE EFFECTED, WHEN you have the same participants who are experiencing all levels of of what they or you are asking they may get tired. Which cold influence your results
Repeated measures design
a type of within groups design where your participants are measured on DV more than one time and how to identify it
FIgure out how much the number of cells for your factorial design
BETWEEN GROUP 16 people per cell and there are 8 cells
2x4 = 8 cells
How many IVs and how many DVs
What is the difference between the interaction and the main effect?
Main effects are looking at the differences between 1 of IV
Interactions looks at the difference of differences in the IVs and how they interact
Tells up about moderating variable
How can we determine the number of main effects you are looking for in a factorial analysis?
Looking at the number of IVs, if we have 4 IVs how many effects do we have 4.
Can we have a situation where in a In our factorial analysis where our Main effect is not significant but our interaction is significant
MAIN EFFECT IS NOT SIGNIFICANT AND INTERACTION CAN BE
Crossover interactions yes
There is a relation between the IV and 1 of the DV
Popular media would talk about a multiple regression finding
be able to identify multiple regression in a popular media sources, depends…. accounting for….. Even when we control for or even when we think about, our variable of interest still impacts In light of, controlling for
Controlling for, accounting for, in light of
When you are conducting a study and your participants kind of suspect the purpose of the study, they act upon that and they give you what you are looking for. What is this called?
demand characteristic