Statistics Flashcards
What is a type 1 error?
Incorrect rejection of a true null hypothesis
False positive
Detecting an effect that is not present
What is a type 2 error?
Incorrect acceptance of a null hypothesis
False negative
Failing to detect an effect that is present
What is relative risk?
Risk in exposed vs risk in control
Relative risk of 1 = no difference
Formula for calculating standard error of mean
Standard error of mean = standard deviation / square root of sample size
What is sensitivity and how is it calculated?
Sensitivity = ability of a test to correctly identify a present disease
Sensitivity = true positives / true positive + false negative
What is specificity and how is it calculated?
Specificity = ability of a test to correctly exclude a disease
Specificity = true negatives / true negatives + false positives
What is a parametric test and when is it used?
Used to test variables within a population that are interval, such as height or age
Only applied when a normal distribution of population data is assumed
What is a non parametric test and when is it used?
Used when data is not normally distributed
Used when data is of nominal value eg eye colour
Positive likelihood ratio equation
Sensitivity / 1-specificity
Negative likelihood ratio equation
1-sensitivity / specificity
What level of evidence is that given by a panel of experts?
IV (4 = lowest level of evidence)
Best study design for assessing a diagnostic test
Cross sectional analysis
Which study designs are used for correlation tests
Spearman (non parametric)
Pearson (parametric)
What are the axis on an ROC curve?
What is an ROC curve?? Looks at performance of diagnostic tests. Greater area under curve = better diagnostic test
Axis = sensitivity and (1- specificity)
Formula for standard error of mean
Standard deviation / Square root of sample size
