Introduction
Measurement can be defined as “a set of operations having the object of determining the value of a quantity”. Most sciences, particularly chemistry and physics, involve doing experiments in which the output is a quantity or value aligned with a certain variable (eg mass, yield, reaction rate, velocity etc). However, measuring these through experiment introduces a problem: how certain are the figures that we obtain? Also, are the measurements accurate, or precise? Is every number that we get significant in terms of the actual figure? To understand these questions more clearly, it is necessary to revisit some definitions.
Significant figures
As mentioned above, all experimental results have a degree of uncertainty associated with them. The precision of any value is expressed by the number of significant digits written when the value is quoted. The more significant digits quoted implies greater precision, with the number of significant figures giving useful information. For example, it is practical to quote the length of a human hair to the nearest fraction of a millimeter, but not for the length of a swimming pool!
For a given value, how many of these digits are significant? Let’s take the following value, 9.851 as an example. The number of significant figures in this value is equal to the number of them that are known with some degree of confidence (eg 9, 8, and 5). The last digit (1) is less certain. The level of certainty depends is derived from the experimentalist’s knowledge and experience, or
determined from the standard deviation of values obtained from repeated experiments.
In general, rules for counting significant figures are as follows:
1. Non-zero digits are always significant
2. Any zeros between two significant digits are significant
3. A final zero or trailing zeros in the decimal portion ONLY are significant
a. Zeros that do nothing but set the decimal point are not significant (230,000 has 2 significant figures, or can be written as 2.3 x 105 so the zeros are not expressed)
b. Trailing zeros that are given but after the decimal point are significant (eg 4.00 has 3 significant figures)
How many significant figures appear in calculations? When answers are combined in calculations, the accuracy of the final answer cannot be any more than the least accurate answer! For addition and subtraction, the answer cannot have more decimal places than the least accurate value used. For example, if we calculate 2.345 + 7.68, the answer obtained (10.025) must be expressed only to 2 decimal places (as 7.68 has only 2 decimal places). For this, the value must be rounded up to the nearest integer number (10.03). The value here (10.05) has 4 significant figures.
In cases of multiplication and division, again the result cannot be more accurate than the least accurate value used. In this case, the answer cannot contain more significant figures than the least accurate value. For example, take 3.2 x 5.81. The value of this on the calculator is 18.592. However, this can only be expressed to 2 significant figures (as 3.2 has only 2 significant figures). The final value then must be rounded up to the nearest whole number. If rounded to 18.6 this contains 3 significant figures, which is also more precise than the original value used (3.2). Therefore the final answer must be rounded to the nearest whole number having 2 significant figures (19).
Precision and Accuracy
In science, it is not appropriate to conduct an experiment only once as this gives no information as to the accuracy, precision and reliability of the result. Only through repetition and analysis of the results obtained do these become clearer
and enable us to say whether results have meaning or not. All experiments have the potential for error, giving rise to uncertainties in the results obtained. The accuracy of a result is a measure of its correctness of a measurement; in effect, the distance of the measured value from the true value. In the majority of cases, we cannot know the true value of a measurement, unless precise standards exist for validation. Therefore, measurements need to be repeated and the mean ( x , or average) value of a number of experiments obtained. This is derived using equation 1, where x is calculated from summing all the results and dividing by the number of data sets