More inference from “researcher” Research by Correlation o The Nature of Correlation Statistical relation between two or more variables No independent variable is manipulated o Correlation and Causation Problem of directionality (e.g. breakups and depression; substance use and impulsivity) Correlation does not mean causation (e.g. smoking and drinking) o Nature of Correlation and Strength of Association Range from -1.0 - +1.0 Negative vs. positive correlation o Why use correlational studies? Can't randomly assign individuals to groups and can’t manipulate IV (e.g. smoking; psychological disorders) o Epidemiological research Study incidence , prevalence , and course of disorders - looking for clues about the disorder Incidence - number of new cases during a specified time Prevalence - number of people with a disorder at any given time Distribution - more or less common in certain populations What factors are associated with frequency of disorder (e.g. gender, socioeconomic status (SES), certain behaviors) Research by Experiment o Nature of experimental research Manipulation of independent variables (e.g. therapy or not; meds or not; levels of exercise) Random assignment Attempt to establish causal relations Premium on internal validity o Group experimental designs Nature and purpose of control groups Necessary to show that IV is responsible for observed changes Should be nearly identical to treatment group Placebo and double-blind controls
Placebo group (ensure that treatment effect isn't due to expectation that one will improve) Easy to do with medications; less so with psychological treatment Double blind - both researchers and participants are unaware of their group assignment Comparative Treatments Designs Type of group design - often next step after showing that treatment is better than placebo Compare different forms of treatment in similar persons (e.g. psychotherapy vs. medication vs. combination) Addresses treatment outcome (did change occur?) Dismantling studies (break study into parts and remove or focus only on certain aspects). Necessary to figure out the “active” components of the treatment Single-Case Experimental Designs o Nature of single subject design “Systematic study of individuals under a variety of conditions” Rigorous study of single cases; manipulations of experimental conditions and time Repeated measurement (rather than just once before and once after treatment - like larger studies) Premium on internal validity o Types of Single-Subject Design Withdrawal designs Baseline (depression) Treatment (e.g. Zoloft); assess depression Withdrawal (stop medication): assess depression Assets: better sense if treatment caused changes Liabilities: remove a treatment that might be helpful; risk relapse; impossible to “withdraw” most psychological treatments (once learned, can’t force patient to unlearn them) Multiple Baseline Designs Don't start and stop treatment Instead, start intervention at different times across settings (home vs. school) or behaviors (hitting; talking back; doing homework)
You've reached the end of your free preview.
Want to read all 24 pages?