Chapter 1: Research and Methods Flashcards
Exam review CH1
Predictor Variables
- What you believe impacts later thoughts, feelings and behavior
- Only an independent variable if it is manipulate
- What we tink is the cause
Outcome Variable
- What you are interested in measuring differences/change in
-What we think is the effect
Requirements for causal relationship
- Co-variation
- Temporal Precedence
-Elimination of plausible alternatives
Co-variation
Outcome must vary as a function of the predictor (change in A -> change in B)
Temporal precedence
Measurement of predictor must be before measurement outcome (makes sure the outcome is affecting the predictor (study -> test scores not test score ->study)
Elimination of plausable alternatives
- Experiments
- Experimental control
- Measurement of third variables
Variables we cannot manipulate
- Infeasable
- Unethical
Infeasable
Variables that we cannot manipulate (race, gender, ses, etc)
Unethical
Certain events, such as stressful events that should not be induced because it may cause distress
Multivariable design
This approach involves examining multiple variables simultainously to understand how they interact and influence one another
Hierarchical Linear Model (HLM)
statistical model used to analyze multiple levels of nested culture
Example of Hierarchical linear model (HLM)
Study on how student performance (test scores) influenced by individual characteristics (study hours) and school characteristics (school funding)
Structural Equation Model
Allows for testing complex relationships among observed and latent variables
Example of structural equation model (SEM)
study of the relationship of grades, SES and motivation, though motivation is not directly measureable. The relationships we hypothesize are
SES -> Motivation
Motivation -> Grades
SES -> Grades
Define observable traits of motivation and link to SES, or estimate the strengths of both and then test how well the model its this data
Cross sectional VS Longitudinal Measurement
Cross sectional: sectional multiple cohots measured at one time point
Longitudinal Measurement: one cohort measured at multiple time points
Cross sectional VS Longitudinal
Longitudinal: 1) Hard to separate personal aging from cohort effects 2) Expensive and logistically different 3) Results take many years 4) Practice effects of repeated measures 5) Selective attrition
Cross Sectional: 1) Hard to separate from historical effects 2) Results do not reflect within-person changes
Cohen’s D VS Pearson’s R
Cohen’s D: Small 0.2, Medium 0.4, Large 0.6
Purpose to measure the magnitude of the difference between two groups (treatment vs. control) Measure of the effect size of the difference btwn two means
Pearson’s R: Small 0.1, Medium 0.3, Large 0.5
Purpose to measure the strength and direction of the linear relationship, Tells you how one variable can predict another in terms of liner relationships, measure of correlation
Confound
Outcome variable that causes both predictor and outcome variable
Mediator
- Mechanism through which predictor variable impact outcome variable
- “Middle Step” in sequence of psychological variables
Moderator
- Changes the magnitude or direction of the association between predictor and outcome
What do confounds and mediators do
Help explain the relationship between the predictor and variable
Difference between confound and mediator
- Confound is external to the causal process
- Mediator is intrinsic to the causal process
Mediators
Carry influence between predictor variable and outcome variable
Often cognitive or affective link between two behavior variables
Moderators
Change how the predictor variable affects the outcome variable
- ‘‘What might make this effect stronger or weaker?”
-“What might change the direction of this effect?”
Measured via statistical interaction
Can be any kind of variable
