ANTC67H3 Lecture Notes - Lecture 4: Statistical Power, Repeated Measures Design, Recall Bias
Document Summary
Increase the size of the sample to increase probability of representativeness. Measuring error: occurs when there is inaccuracy in the instrument or the person operating the instrument due to chance. Methods to address: repeated measures power to detect a clinically important. Calculation: sample size calculation use standard formulae that includes. Acceptable error (chance of missing a true event) Relative size of group being compared: for case control: if disease rare, take all cases of disease and a sample of control; therefor ratio control to case is under the control of the researcher. We can increase the power of it by increasing ratio of control case. Usually maximum is reached at a ratio of 4 control to 1 case. Systematic error: occurs when results differ from the true value in a systematic manner (not due to chance); over or under estimation occurs. Larger sample size does not help: sources: Selection bias: systematic difference between individuals in study and not in study.