Explanatory Concepts
In total, I developed five broad explanatory concepts which covered the concerns expressed through initial fieldwork, and which I wanted to ‘tap’ in the questionnaire in order to identify more general patterns. Though the terminology is similar, I use these concepts not according to a Hempelian ‘covering law’ methodology, but, in keeping with social science statistical research, simply to identify target phenomena (such as beliefs pertaining to archaeology, heritage, etc.) so that I can then develop particular, ‘askable’ questions related to each particular network of beliefs (see De Vaus 2001:24-5, 2002:47-52) on ‘operationalization’). These abstract, explanatory concepts include: archaeological science, heritage/identity, economic, diversion and religious/spiritual. During subsequent factor analysis of the responses to the questionnaire, the categories of economic and religious/spiritual both demonstrated a factor loading onto two respective subdivisions so that they each were broken into a set of two categories; economic-personal and economic-general; spiritual-healing and spiritual-‘nonrational belief’ (Appendix 2, Tables A2.1-2). To descend the ladder of abstraction, specific measurement objectives were developed to tap of measure each of these general concepts and serve as specific indicators (De Vaus 2002:47-52; Foddy 1993:148-52; Fowler 1995:14). From these indicators, specific questions in multiple formats were then developed to measure the explanatory concepts. To bring it back to the examples quoted above, in addition to the personal statements concerning spiritual practices I want to be able to gauge overall how important spirituality at Teotihuacan is for all respondents. To do so I can’t ask the ambiguous question “Is spirituality important?” To which I would get ambiguous answers. Rather, the explanatory concept of spirituality serves as a heuristic to develop several specific questions aimed at getting at the relative importance of spiritual beliefs and practices at the site, such as “Do you come to Teotihuacan to collect the energy of the site?” (a reason specified in interviews) (Appendix 1, Figure 3, question 1.1).
Figure 6.4: Diagram of questionnaire development: moving from concepts to specific questions
In total 37 questions were included in the final questionnaire, consisting of closed-format, open-format, dichotomous and multiple response designs (Appendix 1, Figure 3). These questions targeted the full range of response content, including behavior, beliefs, knowledge, attitudes and attributes (De Vaus 2002:95). The majority of these 37 questions ask for opinion and information pertaining to the five explanatory concepts, as well as background and demographic information. Additionally, several questions were oriented to salient, contemporary events (such as opinion regarding the Walmart) and other issues which INAH archaeologists felt to be critically important (such as the opinion regarding looting of artifacts and the administration of the site by INAH, e.g. Appendix 1, Figure 3, questions 16, 25). After running analyses on the responses to all the questions, 28 questions comprised the scale scores measuring the relative strength of the five major associations or explanatory concepts which I had identified. Figure 6.4 graphically adumbrates this process of scale score design. In order to gauge the relative strength of each major association or concept for all of the respondents as well as to permit comparison of each individual’s particular ‘mix’ of associations or the strength of any particular association relative to other association for that individual, a scale sore is necessary (De Vaus 2002:180-94; Kidder 1981:209-227). As it sounds, these scales are designed to take the responses to questions deemed to tap a concept (such as spirituality) and summate the responses to produce a numerical score. The heritage scale, for example, summates an individual’s responses to particular questions (questions 11, 12, 15, and 23, Appendix 1, Figure 3) concerning heritage and identity. An individual’s answers to these questions produces a score, allocated depending upon how favorable the answers are concerning heritage. So for question 11 (“Do you believe Teotihuacan to be an important symbol of the nation’s identity?”), a Likert Scale is presented so that the individual may rate their own attitude to the question, ranging from ‘Strongly Agree’ (a score of 4) to ‘Strongly Disagree (a score of 1), as well as ‘Don’t Know’/No Opinion’ (a ‘neutral’ score of 0). The stronger the belief that Teotihuacan forms part of national identity, the higher the score for that individual on that particular question. The results from question 11 are then combined with the results of the questions forming the heritage scale. The result is the ability to place the individual along a continuum of scores for heritage.
There are pros and cons to scaling an individual’s responses. These concerns (e.g. De Vaus 2002:192-5; Kidder 1981:199-202) focus upon: 1) allowing the individual to rate her/himself in regards to the score versus the investigator coding the answers (see below) and allocating a score; 2) ensuring that there are multiple questions or indicators measuring the same complex concept (such as heritage or spirituality); 3) taking care that the scores to all the questions being combined in a scale are equivalent (e.g. is a score of ‘4’ on one question equivalent to a score of ‘4’ on another); and, to make the scale scores meaningfully comparative, creating equal upper and lower limits for all the scales (if not, how might compare a ‘high’ heritage score of 13 relate to a ‘high’ spirituality score of 7?). Finally and most importantly, how reliable is the scale in actually measuring attitudes to a particular concept (so do questions 11, 12, 15, and 23 really measure individuals’ attitudes and values of heritage?). Provided these caveats are carefully considered and steps are taken in the research design to ameliorate their potential adverse effects, scale scores are still standard practice in social science research as they are better measures of complex ideas or concepts by providing greater reliability and greater precision using a logic of convergence of multiple sources of data (De Vaus 2002: 180-1). Moreover, they have the additional benefit of allowing both the analysis of the responses to any single question included in a scale as well as their combined analysis as a scale.
I attempted to address each of these concerns in this study. Questions which made up the scale scores included both self-assessed attitudinal responses in the form of traditional Likert scales, allowing the respondent to ‘rank’ the strength of their own opinions, as well as dichotomous questions (‘yes’/’no’), multiple dichotomous questions (choosing ‘yes’/’no’ among a defined set of responses, and ‘open format’ questions where the respondent was asked to write their own response. Using multiple format questions allowed for both the respondent to rate their own attitudes (as in Likert scales), as well as such scoring to be conducted by myself during coding. Rather than relying solely on the respondent or myself to code all of the answers, this mix of approaches to scoring was combined in the final scale scores.
Forward to Coding
Return to Questionnaire Design
Posted at Mar 23/2009 09:06AM:
C4a4mz iurqidcooqpmhttp://iurqidcooqpm.com/">iurqidcooqpm>, url=http://pfzipgugtbhr.com/pfzipgugtbhr/url, link=http://olyrvoojjiit.com/olyrvoojjiit/link, http://stpeyawkeppc.com/
Posted at Mar 24/2009 05:19AM:
zHSuw7 http://profiles.wikidot.com/profile:tramadol-321 Buy Tramadol Without Prescription - Click here for more information...
Fast worldwide shipping!30 days money back!Really prices and Free bonuses!Customer support available 24/7!No prescription required!
