prev next front |1 |2 |3 |4 |5 |6 |7 |8 |9 |10 |11 |12 |13 |14 |15 |16 |17 |18 |review
Z: the true exposure

X: the measured exposure

e: measurement error

In many epidemiological studies it is very difficult or even impossible to exactly measure the true exposure.

The usual exposure we measure is the surrogate, or observed exposure. It is the effort of epidemiologists to maximise the accuracy and minimise the error of a method of measurement to enable it measure an exposure that is as close as can be to the true exposure. Therefore, it is accepted that there is error in measurement. In fact, it is one of the major sources of bias in epidemiological studies.

Differential when exposure measurement error differs according to the disease or outcome being studied. This usually happens with recall bias related to the diseased cases reporting exposure differently from non-diseases cases.

Non differential occurs as an equal error variance between the diseased and the non-diseased. It underestimates exposure levels, and is therefore a conservative estimate of the true effect of exposure on disease.

Berkson error happens when we assign a single measured exposure to a large number of individuals who are in one category (for example a geographical category). The true exposure for the different people will vary within that group and most probably will vary randomly about the assigned value.

prev next front |1 |2 |3 |4 |5 |6 |7 |8 |9 |10 |11 |12 |13 |14 |15 |16 |17 |18 |review