Tabla de Contenidos
Every measurement process has associated errors, and it is important to know that errors are part of the measurement process. In science and technology, the error or uncertainty of a measurement is called experimental error or observation error.
There are two kinds of errors or uncertainties : random and systematic errors. Random errors occur unpredictably in each measurement, while systematic errors have the same incidence in each determination that is made. Both types of errors are intrinsic to any measurement process, but random errors can be treated statistically and yield determinations that cluster around the true value, while systematic errors can sometimes be minimized by calibrating the measuring equipment, but it is It is important to take them into account since if they are not corrected they can cause incorrect measurements that affect the conclusions of the study being carried out.
random errors
If several measurements of the same magnitude are made, it will be seen that the values you obtain are grouped around a certain value; therefore, the random error mainly affects the accuracy of the measurement . Random errors usually affect the last significant digit of a measurement.
The main reasons for random errors are associated with instrument limitations, environmental factors, and slight variations in the measurement procedure. Let’s see some examples:
- When weighing on a scale, the item to be weighed is positioned differently each time the measurement is made.
- When taking a volume reading on a flask, you can read the value from a different angle each time you look at the graduated scale.
- The measurement of the mass of a sample on an analytical balance can differ if it is affected by air currents .
- The measurement of a person’s height is affected by changes in posture.
- The measurement of the wind speed depends on the height and the moment in which the measurement is made; several readings must be taken and the values obtained averaged to obtain a representative measurement, since gusts and changes in wind direction modify each specific determination.
- Readings should be estimated when they fall between marks on a scale or when the thickness of a measurement mark is taken into account.
Because random errors always occur and cannot be predicted, it is important to include in the measurement procedure taking several data readings, and then averaging them to have an accurate determination of the true value of the parameter and at the same time know what it is. the variability of the measurements.
systematic errors
Systematic errors are predictable and always have the same incidence. Typical causes of systematic errors include observation errors, imperfect instrument calibration, and the incidence of environmental factors. Let’s see some examples:
- Forgetting to tare or zero the scale. This produces mass measurements that are always off the actual value by the same amount (coincident with the tare in this case). An error caused by not zeroing an instrument before use is called an offset error.
- Do not read the meniscus on a graduated scale at eye level for a volume measurement. This will always result in an incorrect reading. The observed value will underestimate or overestimate the correct measurement, depending on whether the reading is taken above or below the mark.
- Measuring the length with a metal ruler will give a different result depending on the ambient temperature, due to the thermal expansion of the material.
- A calibrated thermometer can give accurate readings within a certain temperature range , but may become inaccurate at higher or lower temperatures, since all calibration is valid within a certain range of variation of the parameter.
- The measured distance is different using a new tape measure compared to the determination made with an old, stretched tape measure. In this type of situation the errors are proportional to the measurement and are called scale factor errors.
- Measurements of electronic equipment tend to vary over time due to variations in component temperatures. They are said to be susceptible to drift. Measurements obtained with other types of instruments can also be affected as the device heats up.
Once the cause is identified, the incidence of systematic errors can be reduced to some extent, and can be minimized by routinely calibrating the equipment, for example by including controls in experiments, bringing the instruments to the operating temperature at which it was performed. calibration before taking readings, or by comparing measurements with standards.
Although random errors can be minimized by increasing the number of determinations and averaging the results, there will always be measurement uncertainty associated with intrinsic variability of the parameter or measurement procedure. In the case of systematic errors, the best way to minimize them is to become familiar with the limitations of the instruments, have experiences in their correct use and establish measurement procedures, and strictly adhere to them.
Fountain
David A. Freedman. Statistical Models: Theory and Practice . Cambridge University Press, 2005.
