Quasi-Experimental Designs – Lecture Review

These question-and-answer flashcards cover definitions, design structures, advantages, disadvantages, threats to validity, and illustrative examples for quasi-experimental methods, including single-case designs, field studies, natural experiments, classic nonequivalent-group patterns, time-series analyses, and developmental research strategies.

Which four core criteria must a true experimental design satisfy?

Random assignment of participants, the presence of a control group, active manipulation of the independent variable, and experimental control (holding extraneous variables constant).

Why can’t causal inference be made when those criteria are missing?

Without random assignment, control groups, manipulation, and tight control, alternative explanations for an observed effect cannot be ruled out, so causality cannot be firmly established.

What is a quasi-experimental design?

A research design that lacks at least one of the true-experiment criteria (usually random assignment) due to real-world constraints but still attempts to test causal relationships as rigorously as possible.

Give two reasons researchers use quasi-experimental designs instead of true experiments.

Ethical limitations (manipulating variables may be harmful or illegal) and practical limitations (cannot randomly assign in natural settings or large-scale events).

Name the three broad families of quasi-experimental designs covered in the lecture.

Single-case (N=1) designs, field studies & natural experiments, and developmental (age-based) designs.

What key characteristic distinguishes single-case experimental designs from simple case studies?

Single-case designs employ systematic, repeated measurement of a dependent variable before, during, and after an intervention to demonstrate functional (causal) relationships; case studies generally do not involve such systematic manipulation or measurement.

In a single-case B-A design, what does "A" represent and what does "B" represent?

"A" is the baseline phase (pre-intervention) and "B" is the post-intervention phase during which the independent variable is introduced.

How many observations are recommended per phase in a basic A-B single-case design?

At least five observations (measurements) in each phase to establish stability and detect change.

List three descriptive statistics examined in single-case A-B graphs.

Level (mean response), trend (direction of change), and variability (within-phase fluctuation).

What major internal-validity threats still affect simple A-B single-case designs?

History, maturation, instrumentation, and repeated‐measurement (testing) effects.

What does an A-B-A (reversal) design add beyond a simple A-B design?

It removes the treatment in the second "A" phase to see whether behavior returns to baseline, strengthening causal inference by showing the effect can be turned on and off.

Why is the ABAB (A-B-A-B) design considered more powerful than A-B-A?

It replicates the treatment effect a second time, reducing the likelihood that changes are due to chance or a one-time external event, and can alleviate ethical concerns by ending with the treatment in place.

When are reversal designs NOT appropriate?

When treatment effects are irreversible or long-lasting, or when withdrawing a beneficial treatment is unethical.

What alternative single-case strategy is used when reversals are impossible or unethical?

The multiple-baseline design, which staggers the introduction of the intervention across behaviors, settings, or participants without removing it.

State two advantages of single-case designs.

High sensitivity (no between-subject variability) and strong experimental control over one participant or setting.

State two disadvantages of single-case designs.

Low external validity (results may not generalize) and potential ethical issues when manipulating single participants.

What is a field study in quasi-experimental research?

A study conducted in a natural setting where the researcher observes or sometimes manipulates variables with only limited experimental control, resulting in high ecological validity but lower internal validity.

Describe the Israeli traffic-incentive example in one sentence.

Drivers’ baseline travel patterns were recorded for six months, then monetary incentives were offered for eighteen months to reduce rush-hour trips, providing a large-scale field experiment on congestion.

How does a natural experiment differ from a field experiment?

In a natural experiment, the researcher does not manipulate the independent variable; instead, naturally occurring events (e.g., policy changes, disasters) create conditions comparable to experimental manipulation.

Give an example of a natural experiment listed in the lecture.

Assessing chronic stress exposure by comparing pregnant women living in a frequently shelled region to those living in an untouched area.

Which true-experiment criterion is typically still present in a natural experiment?

The existence of a comparison (control) group created by the external event, though groups are not randomly assigned.

How can researchers partially compensate for lack of random assignment in natural experiments?

By matching participants in treatment and control groups on key variables (age, health, SES) and by increasing sample size to reduce variability.

Why is replication across settings vital for natural experiments?

Because internal validity is limited, consistent findings in multiple contexts bolster the argument that the observed relationship is causal.

What is the One-Group Pretest-Posttest design, and what major threat does it face?

A design that measures the same group before and after treatment with no comparison group; it is especially vulnerable to history, maturation, and testing threats.

Explain the Nonequivalent Control Group Posttest-Only design in one sentence.

Two non-randomly assigned groups are compared after one receives treatment and the other does not, without pretest data.

What additional information does the Nonequivalent Control Group Pretest-Posttest design provide over the posttest-only version?

Pretest scores allow assessment of initial group differences and of changes over time, aiding in ruling out selection effects.

What is an Interrupted Time-Series design?

A quasi-experimental design that collects a long sequence of observations before and after a treatment (O1 O2 O3 X O4 O5 O6) to detect interventions’ impact while modeling underlying trends.

Which internal-validity threats are partly addressed by time-series designs?

Selection, regression to the mean, and attrition, because each participant serves as their own control across many observations.

Name one statistical complication common in time-series data.

Autocorrelation: successive observations are highly correlated, which can mask the effect of the intervention (multicollinearity).

What is a Control Series design?

An interrupted time-series that includes a nonequivalent control series measured over the same intervals, providing stronger evidence against history or instrumentation threats.

Why can’t age be manipulated experimentally in developmental research?

Age is a subject variable determined by time; researchers cannot randomly assign participants to different ages.

Describe a longitudinal developmental design.

The same individuals are measured repeatedly at different ages to observe intra-individual change over time.

Give one advantage and one disadvantage of longitudinal designs.

Advantage: Detects individual developmental trajectories. Disadvantage: subject attrition (loss of participants) and potential selective dropout.

Describe a cross-sectional (cohort) developmental design.

Different age groups are measured at one point in time and compared, providing quick age differences without long follow-up.

What is the major confound in cross-sectional studies and why?

Cohort effects: differences between age groups may stem from cultural or historical factors, not developmental change.

What is a sequential (cross-sequential) developmental design?

Combines longitudinal and cross-sectional approaches by following multiple birth cohorts over time, reducing limitations of each individual method.

Across all quasi-experimental designs, what is the common trade-off researchers accept?

Lower internal validity in exchange for practicality, ethics, and/or higher external (ecological) validity in real-world contexts.