Chapter 6: Research Strategies and Validity (Part 1)

What are the five main research strategies in behavioural sciences?

Non-Experimental Design Strategies (no manipulation)

1. Descriptive: Describes individual variables (rather than a relationship between variables)

2. Correlational: Investigates relationships between variables. It does not explain the relationships.

3. Nonexperimental: Demonstrates a relationship between pre-existing grouping variables, without explaining the relationship. (no causality)

Experimental Design Strategies (manipulation)

4. Quasi-experimental: Experimenter actively manipulates a variable without randomization. (non-randomized) So, uncontrolled confounds may account for the results. Attempts to answer cause-and-effect questions.

5. Experimental: Experimenter actively manipulates a variable with random assignment. (randomized)

What is the primary goal of the Descriptive Research Strategy?

To obtain a snapshot or description of specific characteristics of a specific group. Detailed observation to describe understudied phenomena.

What are the three criteria required to establish causality (John Stuart Mill)? (Correlational Research Strategy)

1. Covariation: The two variables change together. BUT covariation does not equal causation.

2. Temporal Precedence: The cause must happen before the effect.

3. Exclusion of Alternative Explanations: Ruling out other factors (often via experiments).

What is the "Third-Variable Problem" in Correlational Research?

When an apparent relationship between two variables (A and B) is actually explained by a third variable (C) that is correlated with both.

E.g. low self-esteem might be the third variable linking anxiety and depression.

What are the similarities between Correlational and Nonexperimental Research?

Both are designed to demonstrate that a relationship exists between two variables, but neither tries to explain the relationship.

How does a Correlational study differ from a Nonexperimental study in terms of data collection?

Correlational: Uses one group of participants and measures two or more variables for each person.

Nonexperimental: Compares two or more pre-existing groups on specific scores.

Define the three types of Nonexperimental designs based on time:

Cross-sectional: Measures variables at a single point in time.

e.g. Data are created and collected at present, right now.

Prospective: Looks forward in time; follows a group into the future.

e.g. data is collected longitudinally after the study begins.

Retrospective: Looks back in time; uses historical data or recollections.

e.g. uses data created or archived before the study design.

In nonexperimental designs, what is the primary difference between a Cohort and a Case-Control study?

Cohort Study: Groups are defined by Exposure (e.g., exposed vs. unexposed to a risk factor).

Case-Control Study: Groups are defined by Outcome or illness (e.g., cases with the illness vs. controls without it).

What are some examples of Quasi-Experimental designs?

One Group Pretest-Posttest: Measuring a group before and after an intervention.

Pretest Measurement → Intervention → Posttest Measurement

Posttest-only with Control Group: Comparing a treatment group and a control group only after the intervention.

Treatment group → Intervention → = Measurement

Control group →→→→→→→→→

Pretest and Posttest with Control Group: Measuring both treatment and control groups before and after an intervention.

Treatment group: Pretest Measurement → Intervention → Post.

Control group: Pretest Measurement → Standard Treatment → Posttest Measurement

What are the two "defining features" of an Experimental Design?

1. Active manipulation of a variable.

2. Random assignment of participants to conditions.

• It reduces systemic confounds

• Allows alternative explanations to be ruled out

• provides the strongest basis for causal inference.

What are the limitations of an Experimental Design?

• Experiements may have limited external validity.

• Not all questions can be studied experimentally.

• Ethical and practical constraints often apply.

Define Research Strategy

Strategy: The general approach determined by the research question and goals of a research study.

Define Research Design

Design: Decisions on how to operationalize constructs and select participants. It specifies whether the study will involve groups or individual participants, will make comparisons within a group or between groups and how many variables will be included in the study.

Define Research Procedures

Procedures: The exact, step-by-step description of how the study will be executed.