Concepts, Operationalization, & Measurement Dr. Guerette
Conceptions & Concepts These are the abstract mental images (concepts) that will be turned into terms that can then be conceptualized specifically into what we mean when we use particular terms.
Conceptions & Concepts Conceptualization The process by which we specify precisely what we mean when we use particular terms. Indicators and dimensions These are the end product of the conceptualization process that provides us with measures of our concept.
Conceptions & Concepts Creating conceptual order This is the process that will lead to measurable definitions for study. The order realizes that progression of measurement from a vague sense of a term to a specific measurement
Conceptual Order Conceptualization Conceptual definition Operational definition
Operationalization This brings us one step closer to measurement by developing operational definitions, which specify what operations should be performed to measure a concept. This includes: Scoring Exhaustive & Exclusive measurement Levels of measurement
Operationalization Scoring This represents another way of thinking about measurement in that it involves actually making observations and assigning scores or values.
Measurement as Scoring Gender Impulsivity Criminal History Age 1 7 26 28 8 34 5 21 4 18 9 33 10 27 3 22 6 20 30
Operationalization Exhaustive Measurement Exhaustive qualities of the attributes of a variable allow us to classify every observation. Mutually Exclusive Measurement Mutually exclusive attributes help researchers classify every observation in terms of only one attribute.
Levels of Measurement Nominal measures - Are variables whose attributes are exhaustive and mutually exclusive and offer names or labels for characteristics. Ordinal measures - Are variables whose attributes are exhaustive and mutually exclusive and can be logically rank ordered from greater than to less than. Interval measures - Are those that have the same characteristic as nominal and ordinal but the logical distance between attributes can be expressed in meaningful standard intervals with zero having no meaning. Ratio measures – Have the same characteristics as those above except that zero has meaning.
In Class Exercise Levels of Measurement Indicate the level of measurement – nominal, ordinal, interval, or ratio – that describes each of the following variables: Length of prison sentence (in months) Attitudes toward the war in Iraq (strongly approve, approve, disapprove, strongly disapprove) Sex of marine life handlers at Sea World Type of childhood victimization (physical abuse, sexual abuse, neglect) Height (in centimeters) Score on a scholastic aptitude test Place of birth Number of years on a job (0-2 years, 3-9 years, 10 or more years) Method of drug administration (snort, smoke, freebase, ingest, intravenous, inhalation, ingest) Age at time of entering military service (under 18, over 18)
Criteria for Measurement Quality Two key standards for measurement quality Reliability - Refers to consistency. We look to be sure the measurement used will produce similar results over time. This can be achieved through the use of the Test-retest method, Inter-rater reliability, and The split-half method.
Criteria for Measurement Quality Validity - Simply put, means are you testing what you say you are testing? While more difficult to test for than reliability, ways of dealing with validity consist of Face validity – common agreement Content validity – adequate coverage of range of meanings Criterion-related validity – a comparison w/ an external measure Construct validity – based on logical relationships Multiple measures – comparison to alternative measures
Reliable but Not Valid
Valid but Not Reliable
Valid and Reliable
Composite Measures Combine individual measures to produce more valid and reliable indicators. Typologies – These are produced by the intersection of two or more variables to create a set of categories or types. An index – Where there are two distinct conceptions that are combined in such a way to produce a measure that is more parsimonious than individual variables.