ATHK1001 Lecture Notes - Lecture 6: Blind Experiment, Numerical Analysis, Confounding
9 Aug 2018
School
Department
Course
Professor
Document Summary
Systematic distortion of data due to systematic factors. Non-systematic (random error or noise) distortions of data that are unrelated (independent) to each other and any other factors: systematic error. Clever hans (true answer confounded with handler"s breathing) Use double blind experiments in drug trials (placebo) Sources of systematic error: dealing with random error. Environmental (e. g. different sites for data collection) Psychological (e. g. desire to please by the participants) Contamination (e. g. night guess which is a placebo) Meta-analysis looking and analysing a group of studies studies. A large combined numerical analysis of the results of many. Results should converge on a conclusion: bias in data. Systematic errors due to beliefs and assumptions of researchers. Values-based bias the way in which we think is correct data. 1 confounding variables: second variable varies with the one of interest. Analytical thinking lecture 6: dealing with bias. Acknowledge all the steps you have taken.
