Become an
Expert in Spirometry

Dependent and independent variable, measurement errors and biasErrors and bias in dependent variable

It is common for rulers to have no graduation on the extreme left and right. Let us assume that you were very nonchalant and would forget to take into account the say 8 mm ungraduated zone to the left (i.e. would take a reading as if the origin was on the left edge of the ruler). In that case the data would not only have reading errors as in the previous example, but also all readings would be 8 mm too small. We now have two sources of error: a random (non-systematic) error and a systematic error (bias). The points still scatter about a straight line, which when fitted by eye still has a slope of about 2.54, but which intersects the Y-axis at -0.8 cm. The empirical relationship between cm (Y) and inch (X) is now:

Y = -0.8 + 2.54·X

Top of page | | | ©Philip H. Quanjer