In science, measurement is very important, especially in experiments, because numbers make a statement more valid and reliable. However, measurement is uncertain. There is no exact or perfect reading. This is due to the limitations of devices use in measuring objects and the limitations of the experimenter himself.
Knowing that there is no such thing as an absolute reading of measurement, there is what we call approximation. And, there are 4 Orders of Approximation.
Zeroth Order of Approximation: Order of Magnitude
This is the most practical and most convenient order of approximation. In this order we express an approximation in terms of exponents with base 10. For example, you were asked the distance between your house and your school. Following the Zeroth Order of Approximation, you would say, “About 102 meters”. This method is used by the famous physicist Enrico Fermi in answering simple approximation problems.
First Order of Approximation: Significant Figures
This method involves the use of significant figures and following its rules to minimize the number of digits in a certain reading. This includes the operations involving significant figures. The advantage of this method is to avoid confusion and crowding of numbers in calculations.
Second Order of Approximation: Best Estimates
In this order we are able to see all the values that came up in an experiment. For example, if we were asked to find the average of 5, 6, 5, 7, 4, 5, 6, 1, 7 and 10. The average will be 5.6. Given only the average (5.6) in the results, the reader of the data would not be able to know the range of the values that came up in the experiment, he would knot know that there exist values of 1 and 10. Using the Second Order of Approximation, the average would be subtracted from the highest or lowest value in the data wherein the difference would be included in the result.
For the example given above the Best Estimate would be,
1 – 5.6 = - 4.6
Best Estimate: 5.6 +/- -4.6
Third Order of Approximation: Statistical Treatment
This order of approximation is used for measurements so large that it approaches infinity thus it uses a distribution function f(x).