Key Concepts in Correlational Research and Design

These flashcards cover key terms and definitions related to bivariate correlation, research design, and validity in psychological research.

Bivariate correlation

An association that involves exactly two measured variables.

Moderate effect size for $r$ according to Cohen's guidelines

An $r$ value of approximately $.30$ (or $-.30$).

Visual representation for two quantitative variables

A scatterplot.

Tight cluster of dots in a scatterplot

Indicates a strong association.

Graph typically used to show means when one variable is categorical

A bar graph.

Statistical test for significant differences between two groups

A $t$ test.

Validity interrogated by measurement quality in associations

Construct validity.

The strength of an association or relationship between variables

Effect size.

Effect of sample size on 95% Confidence Interval (CI)

A larger sample size leads to a narrower, more precise CI.

Meaning of statistically significant association based on CI

The Confidence Interval does not include zero.

Likelihood of observed result under null hypothesis

$p$-value.

Definition of an outlier in research

An extreme score that stands out from the rest of the sample.

Impact of restriction of range on correlation coefficient

It makes the correlation appear smaller than it actually is.

Cause of Pearson $r$ showing near-zero correlation in curvilinear relationships

The positive and negative parts of the curve cancel each other out in linear calculation.

Three criteria for establishing causal inference

Covariance, temporal precedence, and internal validity.

Directional problem in causality

Temporal precedence.

Validity type threatened by third-variable problem

Internal validity.

Variable that changes the strength or direction of a relationship

A moderator.

Definition of multivariate design

A research study that involves more than two measured variables.

Cross-sectional correlation in longitudinal design

The correlation between two different variables measured at the same point in time.

Longitudinal correlation that measures stability over time

Autocorrelation.

Cross-lag correlation

The correlation between an earlier measure of one variable and a later measure of another variable.

Usefulness of cross-lag correlations in longitudinal research

They help address the directionality problem and establish temporal precedence.

Variable predicted in multiple regression

Criterion variable.

Meaning of standardized coefficient (beta or $eta$) in regression

The strength and direction of a relationship while controlling for other predictor variables.

Controlling for a variable in regression analysis

To account for its effects to isolate the unique impact of the main variable of interest.

Conclusion when relationship disappears after controlling for a third variable

The third variable explains the original relationship.

Definition of parsimony in research

The degree to which a theory provides the simplest explanation for a phenomenon.

Difference between a mediator and a third variable

A mediator is internal to the causal pathway, whereas a third variable is an external nuisance.

Definition of a quasi-experiment

A study where researchers select an IV and DV but cannot use random assignment.

Description of a nonequivalent control group posttest-only design

Outcomes are compared between a treatment and non-random control group only after the intervention.

Characterization of an interrupted time-series design

A single group is measured repeatedly before, during, and after an event or 'interruption'.

Strongest quasi-experimental design for ruling out alternatives

Nonequivalent control group interrupted time-series design.

Internal validity threat caused by participant dropout

Attrition threat.

Selection effect in quasi-experimental research

When the participants at one level of the IV are systematically different from those at another level.

Threat involving external event affecting everyone at the same time as treatment

History threat.

Definition of regression to the mean

When extreme outcomes become less extreme over time due to the exhaustion of random factors.

Difference in data representation between small-$N$ and large-$N$ designs

Small-$N$ designs present data for each individual separately rather than as group averages.

Timing of treatment in stable-baseline design

Only after a stable pattern of behavior has been established through repeated measurements.

Definition of a multiple-baseline design

A study where the introduction of an intervention is staggered across different subjects, behaviors, or settings.

Effect during reversal period of a reversal design (ABAB)

The treatment is taken away to see if the problem behavior returns to baseline levels.

Major disadvantage of small-$N$ studies regarding external validity

The few participants used may not represent the general population well.

Definition of replicable research

Testing the same hypothesis as an original study but using different operational procedures.

Definition of conceptual replication

Replication study that confirms original findings while adding new variables to test additional questions.

What is $p$-hacking?

Analyzing data in various ways until a $p$-value of just under $.05$ is found.

Meaning of HARKing

Hypothesizing After the Results are Known.

Definition of preregistration

The practice of sharing hypotheses and study designs publicly before data collection.

Definition of a meta-analysis

A quantitative technique that mathematically averages effect sizes from multiple studies on a topic.

File Drawer Problem in meta-analysis

Studies with null or opposite results are rarely published, potentially overestimating effect sizes.

Definition of ecological validity

The extent to which a laboratory study's findings generalize to real-world settings.

Type of validity prioritized in theory-testing mode

Internal validity.

Research mode focusing on generalizability of findings

Generalization mode.

WEIRD in research context

Western, Educated, Industrialized, Rich, and Democratic populations.

Problem with WEIRD participants in psychology research

They are not representative of the whole world's population, limiting generalizability.

Definition of experimental realism

The extent to which a laboratory study feels real to participants, causing them to behave naturally.

Historical bias in participant inclusion in psychology

White samples were often treated as the 'default' or universal standard.

Impact of hair texture on Black participants in EEG research

Participants are often preemptively excluded because equipment is not designed for diverse hair types.

Way to address underrepresentation in psychological research

Diversifying the researchers and engaging communities directly.