You are on page 1of 1

Mathematics of Evaluating Accuracy and Precision

In a number of your laboratory experiments you will be asked to evaluate your data for accuracy
and/or precision. The following discussion and examples will be helpful in understanding the mathematical
treatment of errors.

Evaluating Accuracy:

If the True (accepted) value for an experimental quantity is known, then you will be expected to
calculate the percent error for your lab report.

Percent Error = (experimental value acccepted value) x 100


accepted value

Note that your experimental value may be the arithmetic average of a set of experimental data, or may be a
single value. Also, the sign of the percent error can be positive or negative. This direction of error is as
valuable to assess as the magnitude of the error.

Evaluating Precision:

In most real laboratory experiments, the True value of the result is not known. In this type of
experiment the most probable value is obtained by assuming that positive and negative errors occur with
equal frequency and tend to cancel each other out. Thus the most probable value is given by the arithmetic
mean (average) of the measured values.

The mean value (X) or arithmetic average may be calculated as follows;

X = X1 + X2 + X3 + Xn
n
where: X = the mean value (average)
X1, X2, Xn = individual data points
n = total number of data points

Once you have obtained the mean value you will need to determine the precision of your data to
communicate to others the reliability of your measurements and results. The precision of your results is
usually stated in terms of the sample standard deviation (S). When the precision of the data is good, the
standard deviation is small. To determine S you must first calculate the deviation (di). Deviation (di) is the
difference between the measured value and the calculated mean ( X ).

di = Xi X

When the total number of experimental (N) measurements is small the standard deviation is designated by S,
and is determined by:

d d d ....d (X i X )2 2
d2
2 2 2 2
S 1 2 3 n
=
( N 1) N 1

N-1

Page 26

You might also like