CMN 2101 Lecture Notes - Lecture 2: Internal Validity, Blind Experiment, Longitudinal Study

• School: University of Ottawa
• Department: Communication
• Course: CMN 2101
• Professor: Mohammed El Hashash

Document Summary

Threats to internal validity: how the research is conducted. E. g. measurement tools used - would not use elastic to measure distance (instead use ruler) Procedures, data analysis, selection bias (more girls than boys in a group: unwanted effects due to research participants. Maturation (when a participant get tired or bored) Mortality (when subject drops out of study before completion) History effects (uncontrollable variable: unwanted effects due to researcher. Researcher presence (try to limit interaction with participants) Researcher bias - do double blind to reduce. 1: association: as values in iv change, dv measures vary systematically temporal order: changes in iv must occur before dv; must apply changes to prove causation, zero alternative explanations: nothing but the treatment variable influences dv. Procedures for classical experimental design: generate hypothesis, find a sampling frame/ select sample size, random assignment (control group: keeping everything consistent// experimental group: treatment, things are manipulated, pre-test (to measure variables before treatment, treatment, post-test, analysis and findings.