Measuring social characteristics

Random errors >> Measuring social characteristics >> Scales of measure
The supporters of quantitative approach take the greatest pleasure in measuring things. However, not everyone knows how to do it properly. For you, it is important not to fall into the category of such botchers.
Stacks Image 362
Dembitskyi S. Measuring social characteristics [Electronic resource]. - Access mode: http://www.soc-research.info/quantitative_eng/4.html
The premise of all operations with quantitative expressions of properties of social objects and processes lies in the primary measurement of qualitative attributes or their quantification. Any measurement starts with the search of the simplest qualitative characteristics, whose relations could be expressed on a numerical scale.
Measurement is a procedure where the object to be measured is compared to a certain standard and receives numeric expression on a certain scale.
In principle, any properties of social objects, qualitative and quantitative, are subject to measurement. With quantitative ones (such as age in complete years, earnings in monetary units, education in years of study, etc.) the situation is relatively simple. They already have common standards of measurement: one year for age or educational level, one UAH for earnings, one person for the size of a family, etc.
Qualitative characteristics (social class, people’s opinions, etc.) don’t have established measurement standards. They have to be designed in accordance with the nature of the object under study and the hypothesis of the study. Therefore, measurement procedures of qualitative characteristics could be said to be inventions, which, however, are made up in compliance with the well-established principles and techniques.
But despite the existence of different approaches to the measurement, it is possible to single out general requirements for measuring instruments. In the Soviet tradition it is validity, stability, precision and accuracy, and in the western tradition – reliability and validity. As the western tradition is discussed in a different chapter, the following chart describes only the requirements for validity, stability, precision and accuracy:
Quality criterion
The essence of the criterion
Ways of verification
Validity Confidence that the planned property is registered, and not any other. Logical reasoning as strict self-criticism of the invented scale: a test for a known group on a small sample; the use of a different independent (additional) measurement; judging by the group representing the micro model of the sample population of the planned study; constructing composite index.
Stability Uniqueness, repeatability of measurement results with minimal error. Repeated measurement calculating the absolute error (the measure of full stability) or the relative error (the index of the relative scale stability); the use of several people in the measurement of the same property to verify the legibility of scale indices.
Precision and accuracy Sufficient scale sensitivity as the optimal fractionality giving stable measurement without systematic errors. Eliminating obvious errors: data skewness to one pole, scale points missing; testing several scales of different fractionality with the calculation of the absolute error of each of them and comparing the relative errors; minimizing the relative error with high stability increases precision and accuracy.
You can read about the technical aspects of ensuring these requirements in the book by V.A. Yadov. However, it is preferable to focus on specialized sources dedicated to the construction of measuring instruments in a particular area.
If we talk about the most authoritative books on measuring instruments in social studies, two books come to mind: “Asking Questions” by Seymour Sudman and Norman M. Bradburn, and “Psychological Testing” by Anne Anastasi.
Note:
Thurstone technique is based on the selection of statements describing the phenomenon under study with their further ranking on the 11-point scale. The scale registers how favorably or unfavorably a statement describes the phenomenon under study. Then, based on the specific statistical analysis, the best statements are selected, and they are included in the measuring procedure.
Guttman scale is based on the selection of statements in whose dichotomous assessment (assessment that suggests only two answers, in this case – yes or no) agreement with the statement of a higher rank implies agreement with the statements of a lower rank.
  • default_titleЯдов В.А. Стратегия социологического исследования. Описание, объяснение, понимание социальной реальности. – М.: Омега-Л, 2007. – 567 с.
  • default_titleСадмен С., Брэдберн Н. Как правильно задавать вопросы. - М.: Фонд "Общественное мнение", 2002. - 382 с.
  • default_titleАнастази А., Урбина С. Психологическое тестирование. - СПб.: Питер, 2009. - 688 с.
  • Show More
free counters
Яндекс.Метрика