JX

Research Methods: Correlational vs Experimental

OVERVIEW

  • Summarize the major characteristics of correlational studies.

  • Summarize the major characteristics of experiments and how they differ from correlational studies.

  • Explain major research strategies.

  • Identify the primary ethical principles used to guide research.

  • Assignment: Evaluating Developmental Claims through Research.

INTRODUCTION: SAD BEIGE TOYS

  • Context: There is a debate about whether filling children’s environments with neutral, beige toys contributes to an aesthetically pleasing home versus enriching environments with color.

  • Critics’ claim: Beiges/neutral environments may be harmful because color and environmental richness support development.

  • Media examples (from slides):

    • "BEIGE MOMS" ARE OUT OF CONTROL; POV: You're a beige mom.

    • TikTok videos with titles like "Sad Beige Moms" discussing overconsumption and painting toddler toys beige.

    • Videos with hundreds of thousands of views and hashtags such as #sadbeigemoms, #beigemoms, overconsumption.

    • Others discuss prioritizing aesthetics over development; clips titled Influencer Insanity, etc.

  • Implication for research literacy: Popular media narratives can frame questions about environmental color and child development; need for systematic evidence.

THE SCIENTIFIC METHOD

  • Step 1: Identifying a specific question of interest.

  • Step 2: Formulating an explanation.

  • Step 3: Carrying out systematic research that supports or refutes the explanation.

RESEARCH QUESTION

  • Are colorful environments important for child development?

  • Are monochromatic beige environments harmful for child development?

  • Do children who grow up in colorful environments have better developmental outcomes?

  • Do children who grow up in monochromatic backgrounds have worse developmental outcomes?

DEFINING OUR PARAMETERS

  • Independent variable (IV): environment type — beige vs colorful.

  • Dependent variable (DV): developmental outcomes.

  • Questions to specify DV: what outcomes exactly? how to operationalize them?

  • Age range: what ages to include? which ages might be more susceptible to environment type?

DIFFERENCE BETWEEN CORRELATIONAL AND EXPERIMENTAL RESEARCH

  • Correlational approach examines the relationship between two variables without manipulating them.

  • Experimental approach manipulates an independent variable to observe its effect on a dependent variable, with random assignment and control conditions.

RESEARCH QUESTION (conceptual framing)

  • Correlational framing: Do beige environments correlate with developmental differences?

    • Example as stated: Do children raised in beige environments show developmental differences from those raised in colorful environments?

  • Experimental framing: Does placing children into beige environments causally change developmental trajectory?

CORRELATIONAL STUDIES

  • A correlation is a statistical relationship between two variables (how they vary together).

  • Positive correlation: As one variable increases, the other also increases.

    • Example: more hours studying → higher GPA.

  • Negative correlation: As one variable increases, the other decreases.

    • Example: more stress → less sleep.

CORRELATION AND CAUSATION

  • Problems with inferring causation from correlation:

    • Third-variable problems (confounds): a separate variable may influence both X and Y.

    • Directionality problem: unclear which variable influences the other.

  • Visual aid mention: CDF (Cumulative Distribution Function) used to illustrate data distributions (example figures included in slides).

  • Example cues in slides: Annual Income by Educational Level and related distributions (illustrative of how variables may co-vary and mislead causal inferences).

CORRELATION AND CAUSATION (continued) / CONTROLLING CONFOUNDING FACTORS

  • Identifying potential confounds through common sense and review of relevant literature.

  • Statistically control for potential confounds (e.g., partial correlations, multiple regression, etc.).

  • Temporal precedence considerations to address directionality (leading into longitudinal designs).

  • Longitudinal studies as a way to address temporal precedence.

TYPES OF CORRELATIONAL STUDIES

  • Naturalistic Observation: researchers observe behavior without interference.

  • Ethnography: researcher aims to understand values and attitudes by acting as a participant for a period of time.

  • Case Studies: in-depth interviews and testing with one individual or a very small group.

  • Survey Research: collecting self-report data from a sample.

  • Psychophysiological methods: test for associations between physiological measures and behaviors.

ETHICAL AND PRACTICAL CONSIDERATIONS IN CORRELATIONAL RESEARCH

  • Exploring naturally occurring relationships.

  • Generating hypotheses for future experiments.

  • Rationale for using correlational methods when experimentation is not feasible or ethical.

EXPERIMENTAL STUDIES

  • Random assignment of participants to different experiences.

  • Independent variable: the manipulated factor.

  • Dependent variable: the outcome measured.

  • Treatment group(s): receiving the treatment.

  • Control group: no treatment or a placebo.

EXPERIMENTAL STUDIES / DESIGN TYPES

  • Between-Subject Design: different participants in each condition.

  • Within-Subject Design: same participants experience all conditions.

  • Note on slide wording: The slide text appears to mislabel between-subjects vs within-subjects (with order/randomization notes). Standard definitions:

    • Between-Subject: each participant is assigned to one condition only.

    • Within-Subject: the same participants experience all conditions; order is randomized.

CHALLENGES WITH SAMPLING

  • Recruitment can be difficult.

  • Small sample sizes reduce statistical power (ability to detect an effect).

  • Convenience sampling reduces diversity and generalizability.

  • Financial costs for compensating participants can be a constraint.

THREATS TO EXPERIMENTAL VALIDITY

  • Confounding variables: extraneous factors that vary with the IV.

  • Demand characteristics: participants guess the hypothesis and alter behavior accordingly.

  • Experimenter bias: researchers unintentionally influence results.

  • Attrition: dropout rates differ across groups.

  • Order effects: in within-subjects designs, earlier tasks influence later ones.

  • Floor effects: participants perform at the lowest level regardless of condition.

  • Ceiling effects: participants perform at the highest level regardless of condition.

LONGITUDINAL RESEARCH

  • Definition: the same individuals are tested multiple times as they age.

  • Purpose: measure development by tracking changes over time.

CROSS-SECTIONAL RESEARCH

  • Definition: individuals at different ages are compared at the same time.

  • Purpose: measure developmental change across age groups without following individuals over time.

MEASURING DEVELOPMENTAL CHANGE

  • Longitudinal Research: same individuals across multiple time points.

  • Cross-Sectional Research: compares different age groups at a single time point.

  • Disadvantages of longitudinal studies:

    • Attrition: participants drop out over time.

    • Studies take longer to publish.

    • Testwise or practice effects can inflate performance.

  • Disadvantages of cross-sectional studies:

    • Cohort effects: individuals born in the same period may share experiences unrelated to age.

    • Selective participation and dropout: certain individuals in one age group may be more likely to participate or quit.