Percent Error (Percent
Deviation, Relative Error)
Accuracy
When scientists need to compare the
results of two different measurements, the absolute
difference between the values is of very little use. The
magnitude of error of being off by
10 cm depends on whether you are measuring the length of a piece of paper or the
distance
from New Orleans to Houston. To express the magnitude of the error (or
deviation) between
two measurements
scientists invariably use percent error .
If you are comparing your value to an
accepted value, you first subtract the two values so
that the difference you get is a positive number. This is
called taking the absolute value
of the difference. Then you divide
this result (the difference) by the accepted value to
get a fraction, and then multiply
by 100% to get the percent error.
So, % error = |
your result - accepted value | x 100 %
accepted value
Several points should be noted when using this
equation to obtain a percent error.
1) When you subtract note how many
significant figures remain after the subtraction, and
express your final answer to no more
than that number of digits.
2) If neither of the two values
being compared is an "accepted value", then use either
number in the denominator to get the fraction. If one value is
more reliable than the
other, choose it for the denominator.
3) Treat the % symbol as a unit. The
fraction is dimensionless because units
in the values
will cancel.
4) Notice that the error is a positive number if
the experimental value is too high, and is a
negative number if the experimental value is too low.
Example: A student measures the
volume of a 2.50 liter container to be 2.38 liters.
What is the percent error in the student's measurement?
Ans. %
error = (2.50 liters - 2.38
liters) x 100%
2.50 liters
= (.12 liters)
x 100%
2.50 liters
= .048 x 100%
= 4.8% error
(Note only two sig figs left in the answer after the
subtraction)
Precision
Frequently in science, an accepted or true value
is not known. The accuracy of a measurement cannot be reported if an accepted
value is unavailable. Precision is a measure of how reproducible
experimental measurements are. Precision is reported as Deviation or
Difference of values.
The Absolute Deviation, or Absolute Difference,
of each measurement is the difference of each measurement from the mean or
average:
Absolute deviation = Measured value - Mean
The Average Deviation, or Average Difference, is
the average of all of the absolute deviations. The Percent Deviation, or
Percent Difference, is the average deviation reported as a percentage:
Percent Deviation = Average
Deviation x 100%
Mean