Biostats:

Regression 

Regression types

• Linear regression

  • Used when outcome is a continuous variable

• Logistic regression

  • Used when outcome is a nominal (binary – yes/no) variable

• Multivariate regression

  • When multiple predictors are included in a regression model

    • The effect of each predictor is now “adjusted for” all the other variables included in the regression model

    • Can be used to estimate the effect known confounders may have on the study outcome to ”control for...” those influences in the final analysis 

• Cox proportional hazards regression

  • Used when the outcome is survival probability/time-to-event

• For all types except linear, predictors can be nominal, ordinal,
or continuous

Define “correlation coefficient” and “coefficient of determination”

Correlation (linear regression)

• Measures the strength of association between two variables (independent and dependent)

  • POSITIVE: change in the independent variable causes change in the dependent variable that is in the same direction

  • NEGATIVE: change in the independent variable causes a change in the dependent variable that is in the opposite direction

Interpret r for a given plot

Calculate and interpret r2

question

question