Sources of Measurement Error - Essay Example

Business Research Methods Essay - Measurement Error This essay explains the four major sources of error common to measurement, with two examples given in each case to support each error types. It is necessary first however, to explain the basic ideas behind the concept of measurement before addressing these in detail. We can only make sense of and use the idea of measurement, if we accept certain premises relating to the world and to the particular phenomena or phenomenon under investigation.

This involves commonly assigning values and shared meanings to concepts and ideas, often, in the form of abstract models, from which, through observing and collecting real world information, known also as variables, we acquire and extract meaningful and often valuable data. Essentially, this combining of theoretical with empirical, permits as well as justifies the use of measurement, and stringent observance of these rules, allows for better and more accurate methods for describing behaviors or in making predictions.

These take on their specific meanings through the following and now widely used scales of measurement, namely, nominal ordinal, interval, and ratio. Classifications, which by using a system of numbers, find common consent among scientists. Here then, consensus must come over how these numbers represent different agreed-to values, in order to sensibly group, sort, or differentiate between data. These numbers are ordered: that is, 'one number is greater than, less than, or equal to another number'.

i Differences connecting numbers create what is called distance between them, allowing for a scale of measurement, for example, by agreement that zero stands as the origin number, we might simply assign numbers to the above quotation. Thus, it is possible to state clearly and unambiguously that the number value one is less than the number value two, and the number value three is more than the number value two. One factor of abstract models, is their use in designing and performing studies that measure real-world variables under (as far as possible) controlled conditions.

As complete control of the environment and conditions in which these studies are conducted is not possible, by implication, error must prove an inevitably. This can and does occur frequently, a systematic resulting of bias, the rest as randomly, and occurring therefore, far more erratically.There are four major sources of error, which may taint the results of a study. These are, the respondent, the situation, the measurer (for example, the interviewer) and the instrument(s) used to record, measure or calculate the results of a study.

With the first, two examples might emerge over the respondents opinion differences on variables themselves usually considered 'stable characteristics'ii, such as social class, or ethnic grouping, the other, the result of their - temporarily or otherwise - experiencing boredom or fatigue during the study. The second, the situation, carries too, a variety of potential contaminants, or unintended unrelated factors, consisting of anything placing a strain on the environment in which the study is conducted.

Two possible examples being the effects of room temperature on measurer and participants, again, carrying the potential for fatiguing or provoking feelings of boredom in both. The third, the measurer's potential for distorting the responses they get. This might emerge in the former's unconsciously rewording questions, or, to give another example, their careless or incorrect use of the fourth source of error, the instrument(s) of measurement. The latter, may corrupt the study in proving difficult or complex to use, or even outdated, possibly unable to investigate and therefore exploit the full potential of a research study.

Rather than detracting from studies however, accounting for these major sources of error measurement, enhances and adds validity to continuing research. Word count: 599