Quasi-Experimental Designs and Regression Analysis

These flashcards cover key concepts related to quasi-experimental designs and regression analysis, providing definitions essential for understanding the material.

Quasi-Experimental Design

A research design that lacks random assignment, leaving the researcher with less control over participant conditions.

True Experiment

A study with random assignment to conditions, allowing for claims of causality.

Causality

The relationship between cause and effect, which cannot be definitively claimed in quasi-experiments.

Quasi-Independent Variable

A variable that is not manipulated by the researcher but occurs for other reasons.

Internal Validity

The extent to which the design ensures that conclusions about cause and effect are valid.

Between-Subject Design

A design where different participants are assigned to different conditions.

Mixed Factorial Design

A design combining one true experimental variable with one quasi-experimental variable.

Longitudinal Design

A research method that involves repeated observations of the same variables over long periods.

Cross-Sectional Design

A study that observes multiple groups at one point in time, rather than over a time span.

Correlation vs. Causation

The principle stating that correlation does not imply causation, emphasizing that two variables can be related without one causing the other.

Confounding Variable

An outside influence that can affect the relationship between the independent and dependent variable.

Regression Analysis

A statistical process for estimating the relationships among variables, primarily used to predict the outcome of a dependent variable based on one or more independent variables.

P-Value

A measure that indicates the probability of obtaining the observed results if the null hypothesis is true, typically used to determine significance.

Incidence Rate

The number of new cases of a disease that occur in a specified period among a population at risk.

Attrition

The loss of participants during a study, which can affect the validity of findings.

Matching in Research

The practice of pairing participants based on specific characteristics to control for potential confounding variables.

Case-Control Study

A type of observational study comparing two existing groups differing in outcome based on some supposed causal attribute.

pattern and parsimony

the principle that the simplest explanation fitting all the data is usually the correct one

Nonequivalent control group design (between subjects)

comparing two groups that already exist where one gets a treatment and the other doesn’t

Interrupted time-series design (within subjects)

measuring participants repeatedly before, during, and after an event

cross-sequential design

a mix - following several age groups (cohorts) over time

selection threat

the groups were different to begin with

maturation threat

participants grew or changed naturally over time

history threat

an external event happened during the study and affected the results

what does regression allow us to do?

control for third variables statistically since we couldn’t do it physically through random assignment

how do you tell if linear regression is appropriate on a scatterplot?

if the relationship is NOT curved

what variables do you select for linear regression?

confounds (third variables)

limits of regression

regression cannot definitively prove causation because it lacks temporal precedence and cannot account for unknown third variables that weren’t measured

prevalence

total cases at a specific time

incidence

number of new cases over a period of time