PART 1
Linear regression and calibration curves
Up to now we have tested differences between samples (numeric or categorical variables)
In this lecture we will investigate the association between variables.
Illustrate linear regression and correlation using calibration curves as an example.
How to calculate a line of best fitting line through a cloud of data points. → Least squares method.
A calibration curve (or line) is an example of applied linear regression.
Relation between body weight and body height. There seems to be a relation.
So where to draw the line. Is there an objective way to obtain a line
of best fit (with corresponding function).
Most common used applications of (linear) regression:
Is there an association between two variables?
(In this case body weight and height.)
- What does this association look like (linear, curved)
- Is there some biological explanation for this association?
Can we predict the value of one variable using the value of another variable?
(Calibration curves)
Are there two different methods designed to detect the same variable equivalent
(produce the same values)?
Linear regression and calibration curves
Up to now we have tested differences between samples (numeric or categorical variables)
In this lecture we will investigate the association between variables.
Illustrate linear regression and correlation using calibration curves as an example.
How to calculate a line of best fitting line through a cloud of data points. → Least squares method.
A calibration curve (or line) is an example of applied linear regression.
Relation between body weight and body height. There seems to be a relation.
So where to draw the line. Is there an objective way to obtain a line
of best fit (with corresponding function).
Most common used applications of (linear) regression:
Is there an association between two variables?
(In this case body weight and height.)
- What does this association look like (linear, curved)
- Is there some biological explanation for this association?
Can we predict the value of one variable using the value of another variable?
(Calibration curves)
Are there two different methods designed to detect the same variable equivalent
(produce the same values)?
