Workshop paper -- 10. workshop Masaryk University Faculty of Social Studies Research methods in sociology 24.4.2005 Eeva Seppä Introduction In his article de Vaus (2002) writes about developing indicators for concepts. While making scientific study concepts need good indicators to be correctly measured. Author (ibid., 47) however points out, that it is difficult to find really good and accurate indicators for concepts, because usually subject that is being studied has more than one dimension, and it can be hard to find indicator that measures all its dimensions. Also if concepts are unclear, they cannot be measured or even if they are clearly defined, indicators can measure wrong things and be unadequate. Writer (ibid. 47) gives an example about difficulties in making research. He introduces an imaginery study, where assumption is that religiousness is a response to deprivation. It presumes, that unhappy people are more religious than happy people. Religiousness is measured by monthly church attendance, and deprivation by monthly income. If the result of this imaginery study is that people that earn less are also less religious, according to the author it doesn't still mean that first hypothesis has to be rejected. De Vaus (ibid., 47) points out that this kind of research concepts have to be more clearly defined to be measurable. In all the cases deprivation isn't surely only about low income, likewise religiousness isn't surely only about activeness in church attendence. This way unclearness of concepts can mislead researcher. De Vaus (ibid., 49) introduces few techniques that can be used to make creating good indicators easier, and also to avoid difficulties while making them. Firstly he writes about the importance of clarifying concepts, and explains how to make them clearer if it is needed. Secondly he writes what has to be remembered when researcher starts to create and search for indicator to one's research. (ibid., 48-53) I will present shortly the above-mentioned points of the article, and try to follow the main ideas that writer wanted to point out in this chapter of his book. Clarifying concepts According to de Vaus (2002, 48) concepts are abstract summaries of a set of behaviours and charateristics that in our opinion have something in common. Concepts don't have some kind of independent existence; they are always created by someone. Author (ibid., 48) writes that this can bring problems, because if any definition is not a set one, it can lead to what he calls "conceptual anarchy", and to useless debates about wrong and right definitions. As a solution de Vaus (ibid., 48) advises a researcher to make it very clear how the concept is defined and keep this definition in mind while drawing conclusions. While choosing a concept, author (ibid., 48) reminds that it makes most sense to use most commonly known definition for it, so it would be more easily understandable. If a concept has a number of different meanings, researcher has two options: either to decide to use only one definition, or design the research so that he has indicators for each of the different meanings of concept. (ibid., 49) In the first option it is important to justify why researcher has chosen this definition among the others. Author (ibid., 49) writes that researcher can choose an existing definition or create his own one. Even though definition is created, de Vaus (ibid., 49) writes that clarifying a chosen concept continues through the research, and there should be interaction between analysing data and clarifying the chosen concept. Author writes about nominal concept, which guides the research but doesn't tell presicely which information to collect. This way all the unexpected findings are easier to see. In the second option researcher can divide different definitions into categories. According to author (ibid., 49) in sociology definitions can be divided into two main categories. First one is inclusive and functional way of explaning. In this way of defining action is explained by it's function. For example religion provides people with meaning in life, so any set of believe that does that can be defined as religion. Another category is exclusive and substantive explanations, and according to author (ibid., 49) in the case of religion these explanations are based on content of the believe, not it's function. In this kind of explaning religion must include notion of a supernatural being. Thirdly de Vaus (ibid., 50) writes about delineating dimensions of the concepts. Concepts usually have more than one dimension, and according to author it can be helpfull to distinquish them. He writes that this may lead to creating own indicators for each dimension, or alternatively using only one dimension in the study. Distinquishing is helpful, and can lead to more sophisticated theorising. Author (ibid., 50) gives an example about multidimensional concept: deprivation can be social, economic, political, psychic or physical. When a researcher disdinguishes between them, it forces him to clarify what his theory is about, and this ensures that he developes measures relevant to that theory. Developing indicators When a researched starts to create indicators, according to de Vaus he has to "descent the ladder of abstraction" (2002, 51). This means that researcher has to move from broad to specific, from theories to more everyday issues. He writes that when researcher starts to clarify concepts and specify dimensions, he already starts to descent this ladder. Next step is to solve two broad problems: how many indicators to use and how to develop indicators. For both of these problems there doesn't exist any simple answers, but author (ibid., 51) tries to give some guidelines for them. Firstly de Vaus (ibid., 51) gives advises to the question how many indicators is needed. If researcher has decided to use many different concepts, there should also be a range of indicators for different definitions. Researcher should also think is he really interested in all dimensions of the concept, if it is multidimensional. Some of the dimensions might not be relevant for the study. Researcher should also take care that all the key concepts are thoughoutly measured using several indicators, and that indicator is complex enough to capture the scope of the concept. Pilot testing is also a good way of getting rid of unnecessary questions. Secondly author (ibid., 53) advices how to create actual indicators. He claims that for many concepts it is easy, but if concept is more abstract, it might be more difficult to find a good indicator. He introduces three main approaches how to develop initial indicators for especially questionaire research. First is that measures that are developed in previous researches can be used, and they can be very useful and function well if they are updated. Using previous indicators also gives the advantage of comparing results with other researches and their results. Secondly author (ibid., 53) reminds that for some topics, for example surveying special groups like migrants, Aborigines, young people, childless couples etc., it is useful to use less structured approach, for example unstructured interview or observing, to data collection. This way it might be easier to understand their way of thinking, and see things through their eyes. Thirdly author (ibid., 53) suggests that one option is to use so-called "informants", that are people from inside the group that is been studied. These informants could give good clues what kind of questions would be usefull while interviewing their group. Finally he summarizes, that in the end researcher has to decide what indicators to use, and how to carry it out. In this stage researcher has to know as much as possible about the population he is studying, and also know what he wants to measure. Bibliography De Vaus, David. Surveys in Social Research: Developing indicators for concepts. 2002. Melbourne. Allen & Unwin.