Download Experimental Designs - Applied Research Methods - Lecture Notes | ESP 178 and more Exams Environmental Science in PDF only on Docsity! 1 ESP 178 Applied Research Methods 2/5 Lecture: Experimental Designs Key Concepts from Chapter 7: Experiments: greater confidence in the validity of causal conclusions than is possible in other research designs – rate well on all criteria for internal validity (see Chapter 6). Key features of experiments: 1. Experimental and comparison groups (to establish association) Group Definition Example Experimental or treatment group Group of subjects who receive “treatment” Families living on cul-de-sacs Comparison or control group Group of subjects to whom experimental group can be compared – exposed to different treatment (or no treatment) or has different value on the independent variable Families not living on cul-de-sacs 2. Pretest and Posttest measures (to establish time order) Test type Definition Example Pretest Measures dependent variable prior to treatment in both groups; not always done. Survey of street play before families move to cul-de-sac Posttest Measurement of dependent variable in both groups after experimental group has received treatment; always done in true experiments. Survey of street play after families move to cul-de-sac 3. Randomization (to establish nonspuriousness) Approach Definition Example Random assignment Start with one group, randomly assign individuals to experimental and control groups; removes bias but differences based on chance; useful for internal validity but not generalizability. Take a sample of families, randomly assign them to living on a cul-de-sac or not living on a cul-de-sac Matching Match individuals in treatment group with individuals in comparison group by pairing individuals based on selected characteristics (e.g. age, gender, etc.); not as good as random assignment Find a sample of families moving to cul-de-sacs; find families not moving to cul-de-sacs that match on number of children, age of children, income Types of Experiments Type Key Features Independent Variable Example Experimental True experiment Two comparison groups; variation in independent variable before assessment of change in dependent variable; random assignment to comparison groups Treatment or intervention, controlled by researcher Sample of families selected and surveyed about street play; randomly assigned to treatment and control groups; treatment group families move to cul-de-sacs; both groups surveyed again about street play Quasi- Experimental Nonequivalent control group design Treatment group and control group not randomly assigned; before and after measurement Treatment or intervention, may be controlled by researcher Sample of families in neighborhood surveyed about street play; city puts in barrier on selected streets, creating cul-de-sacs; same sample surveyed again (panel design) and street play levels compared for those who live on new cul-de-sacs and those who don’t; can use different sample (repeated cross- section design)
