Conceptualization, Operationalization, and Measurement Measuring • « Satisfaction with College » – How will you measure this concept? Measurement: -Careful, deliberate observation of the real world for the purpose of describing objects and events in terms of attributes composing a variable. Attributes are qualities or charateristics of something Variables are logical sets of attributes. Measuring • « Party affiliation » – How will you measure this concept? • Choices – Look at voter lists of who is registered as a Republican or a Democrat – Ask people if they are with the Republicans or Democrats ‘Party affiliation’ by itself does not exist in the real world. These are words that we have associated with certain facts that exist in the real world. In other words « party affiliation » is a concept, when we defiine the concept as « belonging to either the Republican party or the Democratic party » we are doing CONCEPTUALIZATION Conceptualization • We notice many things as we go through life – We hear people say bad things about minorities – We hear people say that women are inferior to men – We also read that women make less than men for doing the same work etc. • These things exist in the real world. In order to describe them we invent a word that describes such behavior by people against people, and we call that word « prejudice » This is the process of conceptualization. Converting the mental images of what you observe in the real world into a word or concept and then describing that concept carefully in words. Conceptualization • How you conceive something as « prejudice » or as « party affiliation » or as « religiosity » or as « development » might be different to how I conceive it, so how can we research a thing that we both conceive differently? – If we compare our menal images in the mental file cabinet then maybe we can agree that if x, y and z are present then we can agree that « prejudice » etc. exists – The process of coming to an agreement on what terms mean and then formally writing it down is called conceptualization and the end result is a CONCEPT. Meaning • « In the big picture, language and communication work only to the extent that you and I have considerable overlap in the kind of enteries we have in our mental file sheets . The similarites we have represent that agreement exists in our society…cultural factors shade our understanding of concepts » – Earl Babbie 2012:127 Foundations of ‘Scientific’ Language • Note: In order not to talk past each other there has to be shared understanding that is dependent on cultural context and lived reality-Language is rooted in societal structure- The language of science tries to minimize this structural rooting of its language to attain a higher level of ‘objectivity.’ – Conceptualization: is therefore the foundation of a scientific language, without which everything else is meaningless! Literally… Measuring • In order to measure « scientifically » remember what we said in the beginning lecture: – You must have a thing to study that is real and not imaginary. In order to measure the « real » you rely on your senses while being careful and deliberate so as to guard against pitfalls, cultural and logical. Measuring • « ABRAHAM KAPLAN » – Three classes of things that sociologists measure: 1. Direct Observables (age, sex, height) 2. Indirect Observables (the force exerted, or the answer given) 3. Constructs (that measure either 1 or 2 since they have 1 or 2 as their ‘indicators’) Constructs are based on mutually agreed mental conceptions of some object or behavior, the conceptions must be based on observables to be scientific. Constructs do not have intrinsic meaning The word « sex » means nothing in the real world, but by associating it with some characteristics that are present in some people and missing in others, (the INDICATORS of « sex »)we designate « sex » based on observable characteristics to be either male or female. Measuring • Sometimes you need to clarify the « dimensions » of concepts: – Compassion: feeling versus action dimension – Religiosity: literary versus ritualistic dimension etc. Interchangebility of indicators If we both disagree on what indicates <compassion> but when we both use our own conceptualization of it to study objects or behavior and come up with the same results then we can talk of all dimensions of the concept leading to similar directions of behavior, e.g. women being more compassionate to men on various dimensions of compassion etc. Defining Concepts • Real Definitions, that signify an « essential nature » of things are problematic scientifically speaking, so we rely on 1. Nominal Defintions: arbitrary based on consensus on how the concept will be used, it focuses our concerns for the purpose of the study. 2. Operational Definitions: based on how the concept is actually measured, e.g. SES- income and wealth – The process through which we specify what a concept means by definition is called SPECIFICATION, a form of conceptual funnel passing through which your focus gets narrowed for the purpose of your study – Great researchers continually refine their concepts for more objective application and clarity Conceptualizing Anomie « ANOMIE »: Without Law. Feelings of meaninglessness. Durkheim, 1897- Suicide Merton, 1938- Social Structure and Anomie Leo Srole (1956) scale- Anomia- the feelings of meaninglessness in individuals: measured through agreedisagree questionnaire so that people could subjetively agree or disagree with 5 statments. An example of the process that describes how a concept is measured, i.e OPERATIONALIZATION Measuring • « unemployment » • « disability » – How you define the concept will determine the discription of the unemployed or the disabled, BUT might not make explanatory results very different if the definition is logical- as we saw in the « interchangeability of indicators » segment a few slides back. . Operationalization • « Choices available for operationalization » 1. Range of Variation: e.g. income what is the highest category you want. Should you measure the full range of variation or not? Should be based on distribution in the population you are studying. 2. Precision of variation: is it necessary to have three categories between the highest and lowest or less? Operationalization 3. « Levels of measurement » : 1. 2. 3. 4. Nominal: merely differnt attributes don’t imply a logical rank ordering eg Sex, Race Ordinal: different attributes that can be rank ordered using some dimension Class, Education etc. Interval: different attribute that can be rank ordered and have equal distance between adjacent attributes, e.g. 1-32* is the same as 33-64, age ranking 17-18 same distance as that between 32-33 Ratio: different attributes, can be rank ordered and also based on true zero value- e.g. gives greater detail than interval level data by being able to measure the exact percentage difference between two intervals however apart because there is a true ZERO. 4. Single or multiple indicators: – GPA- college performance- one course or all? Operationalization Research Question: Are women more compassionate than men?, how can you go about researching this question? Note: Qualitative research is more flexible in open ended conceptualization and operationalization but end results are equally « open » to interpretation, there is often no falisfication that is inherent in hypothesis testing with predetermined concepts as in most quantitative analyses. Measurement Quality – Precision Has to do with detail and depth. Precise measures are generally superior to imprecise ones but precision is nnot always necessary: Age: 43 or « in her 40s », which one is more precise? Also, precision does not mean accuracy -Accuracy: The correctness of the measure. -born in Illinois is less precise than born in Carbondale, Il, but suppose the person was born in Chicago, Il, then the less precise measure will be the more accurate one. Reliability and Validity • Reliability: – A particular technique applied to the same object, does it produce the same result? e.g. measuring morale at the factory Qualitative approach: looking at behavioral cues Quantitative approach: looking at grievances filed in objective terms. Which is more reliable and why? Is reliable more accurate? Is reliable more precise? Techniques for Checking Reliability of measures • 1. Test-Retest Method: – Comparing answers to the same questionnaire administered at two different times – Checking your qualitative observations in the same setup at two different times then comparing your notes • 2. Split-Half method • Split your measures into half and then test them seperately to see if the results between the halves are similar. E.g. ten items that measure prejudice, split them into a group of two seperate 5 item measures, then compare results between two. Reliability and Validity • 3. Using Established Measures: – Measures that have been used in previous studies (does not guarantee that they are either accurate or precise) 4. Reliability of Research Worker In qualitative do independant coding to compare with your or someone else’s coding In quantitative call-back respondents or revisit a sample of records to ensure correct recording. Validity • Validity: – The extent to which a particular measure accurately reflects the underlying concept. – Validity means that we are measuring what we say we are measuring, e.g. measuring height for « how tall is x? » rather than weight Types of Validity i) Face Validity: common understanding /images of the concept, the measure should be logically consistent with such understanding and images. Reliability and Validity ii) Criterion-related validity (or predictive validity) – The degree to which a measure can have predictive success, related to some external criterion. III) Construct Validity: - The logical/theoretical relationship between variables. How is this variable supposed to be related to other variables, e.g. poverty and health, socioeconomic status and crime etc. IV) Content Validity: Does the measure the complete range of the concept, its various dimensions e.g. militarization, how will you measure it? .