Stats Lecture 11

What is a t-test used for?

to assess differences between two means; compare single sample mean against known population mean

Basic rule of parameter estimation

the higher the observations (N) of sample, the more reflective of overall population

Independent t test

independent samples (2 different groups, 1 test)

Dependent t test

correlated samples; same group tested twice; much more conservative compared to independent t test

Assumptions for the t-test

population from which sample was drawn is normal, random samples, homogeneity of variance (samples have smaller variances, variance of one group should not be more than 2x larger than the other)

t ratio

signal to noise ratio

Signal

difference between means

Noise

standard error of mean difference

Relative risk

a measure of the likelihood of an event occurring in one group compared to another, often used in statistical analysis of epidemiological studies

Relative risk equation

[A/(A+B)] / [C/(C+D)]

Absolute risk reduction

the difference in risk between two groups, indicating how much the risk is reduced due to an intervention

Absolute risk reduction equation (ARR)

[A/(A+B)] - [C/(C+D)]

Sensitivity

tells us how well a positive test detects disease; true positive rate

Sensitivity equation

number with disease AND (+) test / number with disease

Specificity

tells us how well a negative test detects non-disease; true negative rate

Specificity equation

number without disease AND (-) test / number without disease

Positive predictive value

the proportion of all people with positive tests who have the disease

Positive predictive value equation

people with (+) test AND disease / all people with positive test

Negative predictive value

the proportion of all people with negative tests who do not have the disease

Negative predictive value equation

people with (-) test AND no disease / all people with negative test