Psychology Research Methods Flashcards

Thinking Critically About Correlation and Causation

• Core Principle: Correlation indicates a statistical relationship or association between two variables, enabling prediction, but correlation does not prove causation.
• Correlation and Cause-Effect Fallacy:
  • Example (Mental Illness and Smoking): Mental illness correlates with smoking, meaning individuals experiencing mental illness are more likely to be smokers.
  • Causation Implications: This correlation alone does not prove that smoking causes mental illness or that mental illness causes smoking.
  • Three Explanations for Correlated Variables:
    1. Variable A causes Variable B (e.g., smoking leads to mental illness).
    2. Variable B causes Variable A (e.g., experiencing mental illness makes an individual more likely to smoke).
    3. A third variable triggers both Variable A and Variable B (e.g., a stressful home life triggers both smoking and mental illness).
• Predictive Power of Correlations:
  • Negative Correlation Example: Self-esteem correlates negatively with depression, meaning it predicts depression. The lower a person's self-esteem, the greater their risk for experiencing depression.
  • Takeaway: Correlation suggests a possible cause-and-effect relationship, but experimental testing is required to prove causation.
• Retrieval Practice Scenarios:
  • RP-9 (Basketball Coach Scenario): A school basketball coach claims she saved her team's winning streak by yelling at the players after an unusually bad first half. An alternative explanation for performance improvement is regression to the mean (a performance returning to its normal statistical baseline after an extreme outlier event).
  • RP-10 (Marriage and Male Hair Loss): The length of marriage positively correlates with hair loss in men. This does not mean marriage causes hair loss or that bald men make better husbands. The underlying third variable driving both factors is aging.

Social Media and Youth Mental Health: A Case Study

• Context and Social Trend:
  • Depression rates among teenagers, particularly girls, have increased drastically.
  • Youth hospitalizations for attempted suicide or self-injury have more than doubled.
  • This trend coincided with the widespread introduction and availability of smartphones and social media access.
• Hypothesized Mechanisms (John Jonathan Haidt, Jake Rush, Jean Twenge):
  • Screen time directly displaces crucial face-to-face social interactions and sleep.
  • Social prompts encourage unfavorable online comparisons with peers who appear to travel, party, and enjoy life more.
• Evaluating the Hypothesis via Research Designs:
  • 1. Correlational Research:
    • Question: Does time spent on social media correlate with teen mental health?
    • Findings: In $78\%$ of $97$ studies evaluated, the answer was yes. The correlation was most dramatic when isolating social media exposure among girls (e.g., British data in Figure 2.8).
  • 2. Longitudinal Research:
    • Question: Does social media use at Time 1 predict mental health outcomes at Time 2?
    • Findings: Among $57$ longitudinal studies analyzed, $37$ confirmed that initial social media exposure predicted subsequent mental health declines.
  • 3. Experimental Research:
    • Question: Is there a direct cause-and-effect connection between screen time and mental health?
    • Study Design: Approximately $1,700$ participants were instructed to deactivate their Facebook accounts for $4$ weeks, while a control group maintained standard habits.
    • Results: Compared to the control group, participants in the deactivation group spent more time watching TV and socializing with friends and family, and reported lower depression, greater happiness, and higher life satisfaction.
    • Synthesis: Out of $44$ experiments reviewed, $80\%$ demonstrated a causal effect of social media use on negative mental health outcomes.
  • 4. Quasi-Experimental Research Design:
    • Definition: A research design used to evaluate an intervention or exposure without the use of random assignment.
    • Study Design: Evaluated mental health outcomes following the campus rollout of Facebook or community arrival of high-speed internet.
    • Results: In $80\%$ of quasi-experiments, mental health declined as social life moved online, disproportionately impacting girls.
• Conclusion and Policy Implications:
  • Extended social media exposure is a major cause of youth mental health decline (contributing to "the anxious generation").
  • Proposed intervention: Creating phone-free school days to enhance educational outcomes and reduce rates of anxiety, depression, and self-harm.
  • Caveat: Research is ongoing, and some investigators continue to question the exact scale of the social media effect.

Experimental Research Design and Control Methods

• Purpose of Experiments:
  • To isolate cause and effect, experiments manipulate independent variables while controlling or holding constant other extraneous variables.
• Essential Components of Experimental Design:
  • Experimental Group: The condition in which participants receive the treatment (e.g., reduced screen time or an active drug).
  • Control Group: The contrasting condition in which participants do not receive the experimental treatment, serving as a baseline.
  • Random Assignment:
    • Definition: Assigning participants to experimental and control conditions by chance (e.g., using a random number generator or coin flip).
    • Function: Equalizes the groups by minimizing pre-existing differences (such as age, intelligence, attitudes, or physical traits).
    • Illustration: If $1/3$ of volunteers in a study can wiggle their ears, random assignment ensures approximately $1/3$ of the participants in both the control and experimental groups will be ear-wrigglers.
• Evaluating Remedies and Preventing Bias:
  • Misleading Testimonies: Humans naturally seek remedies when ill or down, often misattributing natural recovery to an inactive treatment (e.g., taking zinc for a cold that naturally subsides, or historical bloodletting in the 1700s where survival was attributed to the treatment and death was attributed to disease advancement).
  • Placebo: An inert, pseudo-treatment condition (such as a sugar pill containing no active drug). Derived from Latin meaning "I shall please".
  • Placebo Effect: Experimental results driven purely by participant expectations. Administering an inert substance can produce real psychological and physiological symptom reductions in pain, depression, and anxiety.
    • Documented Examples: Athletes running faster on fake performance drugs (McClung & Collins, 2007); decaf coffee drinkers experiencing vigor/alertness when believing they consumed caffeine (Dawkins et al., 2011; Atlas & Wager, 2012; Kirsch, 2010).
  • Blind Procedure: Participants are kept uninformed about which treatment (if any) they are receiving.
  • Double-Blind Procedure:
    • Definition: An experimental procedure in which both the participants and the research staff administering the treatment and collecting data are uninformed ("blind") regarding group assignments.
    • Application: Commonly used in drug evaluation studies to isolate a treatment's actual physiological effect from participant and staff expectations.

Variables, Operational Definitions, and Field Experiments

• Defining Experimental Variables:
  • Independent Variable (IV): The experimental factor that is manipulated; the variable whose causal effect is being studied.
  • Dependent Variable (DV): The outcome variable that is measured; the variable that may shift depending on manipulations of the independent variable.
  • Confounding Variable: An extraneous factor other than the independent variable that might influence the experiment's results. Controlled via random assignment.
  • Operational Definition: Precise, explicit statements specifying the exact procedures used to manipulate the independent variable or measure the dependent variable. Crucial for enabling exact replication by other researchers.
• Case Study: Viagra Clinical Trial:
  • Sample & Procedure: $329$ men with erectile disorder randomly assigned in a double-blind procedure to receive active Viagra or an identical-looking placebo pill across $21$ clinical trials.
  • Results: At peak doses, $69\%$ of Viagra-assisted intercourse attempts were successful, compared to $22\%$ for men given the placebo (Goldstein et al., 1998; replicated by Fink et al., 2002).
  • Variables:
    • Independent Variable: Drug assignment (Viagra vs. placebo).
    • Dependent Variable: Sexual performance (percentage of successful intercourse attempts).
• Case Study: Discrimination in Rental Housing (2006):
  • Method: Identically worded email inquiries were sent to $1,115$ Los Angeles area landlords, varying only the sender's name to imply different ethnic backgrounds.
  • Results (Positive reply/invitation rates):
    • Patrick McDowell: $89\%$
    • Said Al-Rahman: $66\%$
    • Jell / Jabil Jackson: $56\%$
  • Variables:
    • Independent Variable: Implied ethnicity of applicant names.
    • Dependent Variable: Percentage of positive landlord responses/invitations.
  • Societal Application: Highlighted by President Barack Obama on June 26, 2015 (eulogy for Senator Clementa Pinckney) regarding subtle institutional biases in hiring and housing.
• Evaluating Social Programs:
  • When resources for new social programs (e.g., early education, anti-smoking campaigns, school sex education) are limited, researchers can run experiments using a lottery to randomly assign regions/individuals to receive the program vs. a control group (Passell, 1993).

Retrieval Practice & Methodological Comparisons

• Retrieval Practice Questions & Direct Solutions:
  • RP-11: What measures do researchers use to prevent the placebo effect from confusing their results?
    • Answer: Researchers utilize double-blind procedures and assign control groups an inert placebo treatment.
  • RP-12: By using random assignment, researchers are able to control for ______ ______.
    • Answer: Confounding variables.
  • RP-13 Matchings:
    • Double-blind procedure: Controls for the placebo effect; neither researchers nor participants know who receives the real treatment.
    • Random sampling: Helps researchers generalize from a small set of survey responses to a larger population.
    • Random assignment: Helps minimize pre-existing differences between experimental and control groups.
  • RP-14: Why, when testing a new drug to control blood pressure, would we learn more about its effectiveness from giving it to half the participants in a group of 1,000 and a placebo to the other half, rather than giving it to all 1,000 participants?
    • Answer: Giving it to only half allows researchers to compare outcomes against a control group baseline, isolating the active medication's true effects from placebo effects or natural recovery.

Summary of Main Psychology Research Methods (Table 2.2)

• Descriptive Method:
  • Basic Purpose: To observe and record behavior.
  • How Conducted: Case studies, naturalistic observations, or surveys.
  • Manipulated Variables: None.
  • Weaknesses: No variable control; atypical single cases can yield misleading conclusions.
• Correlational Method:
  • Basic Purpose: To detect naturally occurring relationships; to assess how well one variable predicts another.
  • How Conducted: Collect data on two or more variables.
  • Manipulated Variables: None.
  • Weaknesses: Cannot specify or establish cause and effect.
• Experimental Method:
  • Basic Purpose: To explore cause and effect.
  • How Conducted: Manipulate one or more variables; utilize random assignment.
  • Manipulated Variables: Independent variable(s).
  • Weaknesses: Sometimes not feasible; results may not generalize to real-world contexts; unethical to manipulate certain variables.

Scope of Research, Laboratory Realities, and Ethics

• Limits and Possibilities of Psychological Testing:
  • Untestable Questions: Science cannot directly test non-empirical or theoretical questions like "Does free will exist?", "Are people born evil?", or "Is there an afterlife?".
  • Testable Questions: Psychologists can experimentally measure whether beliefs in free will, aggressive traits, or afterlife beliefs affect human thoughts, emotions, and behaviors (Et al., 2003; Sheriff et al., 2014; Webster et al., 2014).
  • Ethical Boundaries: Studies cannot assign children to abusive vs. loving homes to study development. Laboratory proxies for aggression include measuring a participant's willingness to deliver supposed intense noise blasts to a stranger.
• Laboratory Experiments vs. Real-World Behavior:
  • Purpose of Lab Settings: A laboratory experiment provides a simplified reality designed to simulate and isolate core psychological forces (analogous to how a wind tunnel tests aerodynamic forces on airplane designs).
  • Focus on Principles: The goal is not to recreate exact daily behavior, but to test theoretical principles (Mook, 1983). The resulting general principles explain complex everyday behaviors (e.g., noise blast experiments generalise to physical aggression; dark room light-blink experiments generalize to night flying; Mitchell, 2012).
• Psychology's Animal Research Ethics:
  • Rationale: Psychologists study nonhuman animals out of direct fascination with species differences and to understand shared human biological processes.
  • Medical Contributions: Animal testing led to insulin for diabetes, vaccines for polio and rabies, and organ transplant procedures.
  • Model Systems: Simple nervous systems reveal fundamental mechanisms (e.g., sea slugs illuminate neural learning mechanisms; honeybees reveal stress-coping mechanisms; Dinges et al., 2017).
  • Comparative Biology: Rats possess biological similarities to humans (humorously noted by Dave Barry, 2002: "Rats are very similar to humans except that they are not stupid enough to purchase lottery tickets").
  • Ethical Dilemma: The animal protection movement challenges animal research. Roger Ulrich highlighted the philosophical paradox: Work cannot be defended scientifically based on human-animal similarities while simultaneously defended morally based on human-animal differences.