MEASUREMENT & SOURCES OF ERRORS
In research context “measurement is assigning numbers to empirical events according to a set of rules”. Measurement is a three step process,
- Selecting observable empirical events.
- Using numbers or symbols to represent aspects of events.
- Applying a rule to develop a connection between observation and symbols.
Sources of Measurement differences
the research study should be conducted and design and control in such a way that measurement of variables should give precise and unequivocal and undoubted in meanings, to achieve such a level of preciseness, we should try to recognize the sources of potential errors and should try to neutralize these and should try to deal with these errors. Mostly these potential errors are a result from a bias, while some of these errors occur randomly. The four major sources of measurement error are listed as under, which may affect the results of study.
- The respondent as an error source
- The situation as an error source
- The measurer as an error source
- The instrument as an error source
The respondent as an error source
The respondent is a person that is giving answer or response to the question asked or about a particular variable. Different characteristics affect the answer or response provided by respondent, among these are social; status, job rank, religion or ethnic group to which he/she belongs to, residence location, education status and many more. Many of these are taken under consideration while finalizing the design of research study to be conduct.
The respondent can be an error source, if he/she is reluctant to express their true experiences, negative feelings, about the variable that affect the results of research. Next to it lack of knowledge about the object or variable and reluctant to admit it yield research results to be biased.
Respondent can be an error source if he/she is suffering from fatigue, boredom, anxiety, or any other problem like that, then these personal feelings of respondent can affect the results; and these conditions limit the ability of respondent to answer accurately and with full attention.
The Situation as an error source
This area includes countless factors that pay a serious impact on results, weather it’s an interview or a survey based research study. Presence of a third person may affect the results in such a way that he/she may disturb the responses by interacting, or by distracting the real area of consideration, respondent may not feel easy and relax in presence of third person, as well he/she can be reluctant to express their true emotions in front of a third person, and therefore yields the results to be ambiguous.
The measurer as an error source
The interviewer is the person, who is asking questions or opinion about a particular variable. The interviewer can affect the results or can be a cause of error by repeating the same question again and again in different words, by using paraphrasing, by placing questions in a wrong manner/order.
Personal factors like wrong pronunciation, appearance, gestures and postures, conscious and unconscious prompts with smiles, etc may cause respondent to be confuse and therefore affect the results and cause errors.
Carelessness in processing replies, and counting the wrong response and failure of recording the full response will definite causes of major errors in measuring results. Other causes of errors in measurement can be wrong statistical formulae; wrong entries of value in tables, and wrong coding, faulty data analysis, and wrong relationship development among data variables can cause serious errors.
The Instrument as an error source
An instrument is a device that is use for measuring a variable. For example: thermometer (instrument) is use to measure the temperature (variable). A defective instrument can cause major errors.
- It can cause major confusion.
- Technical complex terms and syntax is not understandable to respondent.
- Unclear question, doubt full meanings, mechanical defects (inadequate space for writing answers), lack of choices, misprinting etc can cause a long range of major errors.
- Instrument can be a cause of error because an instrument can’t explore all the related terms and areas of topic under study.
- All instruments have deficiency to explore universe of even a single variable understudy and its inter related variables, and therefore affects the result.
- Variables like leadership, organization of labor, company management strategy, relationship among managers and subordinates, working conditions, motivation, and many more qualitative variables can’t be measure by an instrument.