How To Evaluate The Error | Education

Education

# How to evaluate the error

When measuring any values ​​may occur error, that is, the value obtained may differ from the real. Note the error, it estimates indicate the precision with which it was made, or another dimension.

### You will need:

pen, paper, measurements

## Instruction how to evaluate the error

Step 1:

First of all, it should be understood that there are two types of errors: absolute and relative. The first is the difference between the amount received and the exact value, the second - the ratio between the absolute error and the exact number. In physics, without assessing the magnitude of the error is considered to be unknown.

Step 2:

Make sure you have not made a mistake in the measurements, records from the unit, the calculations, thus eliminating the gross error. They are unacceptable.

Step 3:

Make any necessary adjustments. For example, if one is not necessary to divide the scales at the zero point, it must be considered in all subsequent calculations.

Step 4:

Make sure you know all the systematic errors. The latter may be due to malfunctioning of the device, they are generally listed in the data sheet of measuring equipment.

Step 5:

Measure the random error. This can be done by means of different formulas, for example, using the standard formula quadratic error.

Step 6:

Compare with systematic random error. If the first over the second, it should be reduced. This is achieved by repeated measurements of the same magnitude.

Step 7:

Find the true value, for which the arithmetic mean of all the calculations made.

Step 8:

Determine the confidence interval. This is done according to the formula of calculating the confidence interval using the Student's coefficient.

Step 9:

Find the absolute error in the formula: the absolute error is equal to the square root of the sum of random error in the square and systematic error in the square.

Step 10:

Find the relative error (formula given in paragraph 1).

Step 11:

Record the final result, in which x is obtained by measuring the number of plus / minus error.