Research Methods in Developmental Psychology Michael Hoerger Observation Hypothesis generation Laboratory Observation: Parent-child interactions, marriages, intrusive interviews, attachment style Naturalistic observation: bullying, ADHD Case Study Used to gain detailed information on a single or small number of cases, commonly used in medicine and clinical psychology: rare events, new events, complex events Correlation Correlation r = Strength of relationship between two variables (-1 to +1) What is a “big” correlation? Reliability: r = .90 IQ tests: r = .50 to .90 Personality research: r = .30 Life/death: r = .01 Problem: Correlation ≠ Causation due to 3rd variable problem and directionality problem Solution: Methods and argument Cross Lagged Panel Design (or “Cross lag panel” or “Cross panel lag”) Look at correlation between two variables over time Does X correlated with changes in Y? Smoking at Time 1 causes increased mile time at Time 2 Age 20 Smoking Cigarettes r = -.05 r = .10 Time to run a mile Age 40 Smoking Cigarettes r = .10 r = .30 Time to run a mile Look at correlation between two variables over time Does X correlated with changes in Y? Maternal depression at Time 1 causes increased behavior problems at Time 2 Maternal Depression Maternal Depression Maternal Depression Child Behavior Problems Child Behavior Problems Child Behavior Problems THIS DRUG HAS HELPED TO TREAT: HAY FEVER, ASTHMA ATTACKS, ANXIETY, PAIN, ULCERS, ENURESIS, WARTS, ARTHRITIS, MALIGNANT TUMORS, DIABETES, NARCOTIC WITHDRAWAL, INSOMNIA, COLDS, AND INATTENTIVENESS Experiment “id!” Independent variable: the manipulation; different conditions or groups Alcohol vs. placebo; CBT vs. waitlist Dependent variables: depends on the independent variable; the outcome variable Age at death; depression; liver functioning Problem: Participants must be similar across IV groups Solution: Random assignment Survey Interviews, questionnaires, tests Used for correlational studies or as outcome (DV) measures in experimental studies Highly efficient Can be anonymous Problems: Wording, Response bias (e.g. social desirability) Solutions: Design with care Online Research Most surveys and some experiments can be run on the web (e.g. priming studies) Benefits: most efficient, useful for screening large samples Risks: Lower experimental control, random responding, technical problems, nonrepresentative sampling, ethics http://funpsych.com example Physical Measures Physiological: changes in functioning Galvanic skin response (sweating), pupil dilation, heart rate Physical: walking speed, eye movement, speed of responding, height, weight Neurological: neurotransmitter levels, brain structure Benefits: reliability of measurement Risks: expensive, often fail to provide new information, low correspondence Cross-Sectional Research Groups differ by age Compare children to teens to young adults to older adults Differences are presumed to be the result of age Older people are slower due to aging BUT differences may simply be due to contextual factors, such as the era each group was born in OR older people are slower due to differences in nutrition growing up Longitudinal Research Follow one group over time to what changes with age Problem: expensive, bias due to dropout Cross-Sequential Research Combines cross-sectional and longitudinal research Michael Hoerger To cite this lecture: Hoerger, M. (2007, January 10). Research Methods in Developmental Psychology. Presented at a PSY 220 lecture at Central Michigan University.
© Copyright 2019