HLTB15H3 Lecture Notes - Lecture 3: Confounding, External Validity, Dependent And Independent Variables
Document Summary
Concept of critical analysis question 8 (lec. 2) at rst 6 minutes of lecture. Pico=> outcome is just what is measured not the actual outcome/results. Investment purposes: knowing what to invest in and what to distribute among the population. Want to know which treatments that is helpful/bene cial. Policies or legislations on a broader view. Validity of comparisons made within the study. Necessary to make validity inferences for the source population of study subjects. Threats to internal validity may or may not be fatal. Depending on type and severity of bias. Random error: probability that observed result is due to chance. Equally likely to bias the study in either direction. Remains after systematic error (bias) is eliminated. Random errors is the red line on the graph, a bigger groups = less random error. Systematic error is the blue line based on how the study was designed, no mater how big the sample size, systematic error can still be present.
