Numbers!
Observation Study: observing individuals and measuring variables of interest without attempting to influence the responses Correlational Research: examining the relationship between two variables Caution: Correlation is not causation. You must conduct an experiment to determine cause Developmental Designs: assessing changes over a period of time Longitudinal: observing the same group over an extended period of time Cross-sectional: using multiple groups of different ages over a short period of time Survey Research: administering questions to a selected sample of the population Recall Bias: respondents cannot accurately recall past events Wording Bias: wording of the question is confusing, misleading, or loaded
Define the population Determine appropriate sample size Choose a sampling design Choose an appropriate research design
Probability Sampling: all members of the population have an equal chance of being chosen Simple Random Sample (SRS): Stratified Random Sample: population is split into groups (strata) and an SRS is taken within each group (stratum) Cluster Sample: population is split into groups (clusters), a random sample of clusters is taken, and all individuals from the chosen clusters are selected to be in the sample Nonprobability Sampling: subjects are chosen systematically Convenience Sampling: choosing individuals who are the easiest to reach Quota Sampling: certain traits are chosen, quotas are set for those traits based on their proportion in the population, and subjects are selected based on those quotas
Sample must be large enough to represent the population sufficiently Rule of Thumb: 10% of the population LOOK IN BOOK Larger samples give more accurate results than smaller samples Only for probability samples
Undercoverage: some groups of the population are left out when selecting the sample Nonresponse: an individual in the sample cannot be contacted or refuses to participate Response: behavior of the respondent or of the interviewer results in an untrue or misleading answer Under/overrepresentation: a group is disproportionately represented in the sample Probability sampling minimizes sampling bias Acknowledge that bias may be present
Control: effort to minimize the effects of unwanted variables Without control, there is no way to tell whether the outcome of an experiment was due to the variable in question or an underlying variable Controlling Confounding Variables: randomization, control groups, replication Experimental Design: subjects are randomly selected, randomly assigned to groups, and groups are randomly assigned to a treatment Quasi-Experimental Design: uses naturally occurring groups as opposed to randomization Ex Post Facto Design: groups are assigned based on a pre-existing trait
One-shot Experimental Case Study: a single group is studied after a treatment is assumed to have taken effect The outcome is compared to expected results if the treatment had not taken place One Group Pretest-posttest Design: a single case is observed before and after treatment All changes are presumed to be caused by the treatment Static Group Comparison: one group is given a treatment and compared to a group that is not given a treatment Differences are presumed to be caused by the treatment Difficult to determine if results are due to the treatment or another variable
Pretest-posttest Control Group Design: One group is given a treatment and a second group is not. Both groups a given a pre- and post-test Solomon Four Group Design: Two groups are given a pretest (one control group) and two groups are not given a pretest (one control group). All groups are given a posttest Posttest Only Control Group Design: One group is given a treatment and another group is not. Both receive a posttest Within-Subject Design: Every participant is subjected to every treatment and the control
Nonrandomized Control Group Pretest-posttest Design: same as classic Pretestposttest Control Group Design, except subjects are not randomly assigned to groups Simple Time-Series Design: periodic measurements are taken from a defined group both before and after administering a treatment Control Group Time-Series Design: same as Simple Time-Series Design, except with the addition of a nonequivalent control group Reversal Time-Series Design: repeatedly introducing and withdrawing a treatment Alternating Treatments Design: administering one treatment for a period of time, then administering a different treatment for the same period of time Multiple Baselines Design: Establishing a base line for multiple individuals, settings, or behaviors, and administering a treatment to all individuals, settings, or behaviors
Simple Ex Post Facto Design: One group has a particular trait or characteristic and one group (the control group) does not. Researchers are measuring how the characteristic affects a dependent variable Two Factor Experimental Design: data is collected for all possible combinations of of two factors
Combining data from independent studies Can be qualitative and quantitative data Develop one, strong conclusion
LOOK IN BOOK
Studying side effects of a new drug Examining the relationship between amount of junk food eaten and time spent exercising per day Studying how the Great Depression affected farmers Measuring the correlation between rainfall and mosquito population A focus group to determine the effectiveness of a new commercial A survey (scale of 1-5) to determine the effectiveness of a new commercial Does AHS serve good food? Describe this classroom.