EBM - Statistics and Bias Flashcards
Selection bias
Study population does not reflect a representative sample
Recall bias
Individuals with an adverse outcome are more likely to “remember” more data, especially if a lot of time has passed
Measurement bias
Improper, inadequate, or ambiguous recording of factors
Hawthorne effect
Participants aware they are part of a study are more likely to change their behavior
Observer bias
A researcher’s belief about the effect of a treatment may be more likely to document certain outcomes
Confounding bias
A characteristic not in the causal pathway being studied can distort the effect of an exposure on the outcome.
Effect modification
The effect of an exposure differs depending on a third variable known as the modifier
Lead-time bias
Early detection can appear as a gain in survival though the course of disease has not changed
Length-time bias
Overestimation of survival duration due to the relative excess of cases detected that are slowly progressing
Quantitative (numerical) variables
Continuous and discrete
Quantitative (categorical) variables
Ordinal, nominal, and binary
Power
The probability of detecting a difference; 1 - beta
Type I error (alpha)
Rejecting the null hypothesis when it is actually true
Type II error (beta)
Accepting the null hypothesis when it is actually false
Parametric tests
Typically require a normal distribution and assume equal variance between groups
Examples: t-tests, ANOVA, Pearson’s correlation, linear regression
t-test
Compares the means of 2 independent groups
Paired t-test
Compares the means of the same group before and after intervention
ANOVA (Analysis of Variance)
Compares the means or 3 or more groups
Pearson correlation coefficient
Measures the strength of a linear relationship between 2 continuous variables
Linear regression
Measures the relationship between a dependant variable and at least 1 independent variable
Nonparametric tests
Do not assume normality and can handle qualitative data
Examples: chi-square, Fisher’s exact test, Mann-Whitney U test, Wilcoxon signed-rank test, Kruskal-Wallis test
Chi-square
Tests the association between 2 categorical variables
