Presentation 2 .pptx

bacon

1. Chapter 5: Concepts, Operationalization and Measurement © 2018 Cengage Learning. All Rights Reserved.

2. Learning Objectives Understand the role of concepts as summary devices for bringing together observations and experiences that have something in common Explain how concepts are mental images that do not exist in the real world Describe how operationalization specifies concrete empirical procedures for measuring variables Recognize that operationalization begins with study design but continues through the duration of research Explain why measurement categories must be mutually exclusive and exhaustive

3. Learning Objectives, cont. Distinguish different levels of measurement and the properties of different levels Understand precision, reliability, and validity as dimensions of measurement quality Summarize how creating specific, reliable measures may not reflect the complexity of the concepts we seek to study Understand how multiple measures of a concept can improve reliability and validity Describe composite measures and explain their advantages

4. Introduction Because measurement is difficult and imprecise, researchers try to describe the measurement process explicitly We want to move from vague ideas of what we want to study to actually being able to recognize and measure it in the real world Otherwise, we will be unable to communicate the relevance of our idea and findings to an audience

5. Conceptions and Concepts Clarifying abstract mental images is an essential first step in measurement “Crime” Conception: Mental image we have about something Concepts: Words, phrases, or symbols in language that are used to represent these mental images in communication e.g., gender, punishment, chivalry, delinquency, poverty, intelligence, racism, sexism, assault, deviance, income

6. Three Classes Direct observables: Those things or qualities we can observe directly (color, shape) Indirect observables: Require relatively more subtle, complex, or indirect observations for things that cannot be observed directly (reports, court transcripts, criminal history records) Constructs: Theoretical creations; cannot be observed directly or indirectly; similar to Concepts

7. Conceptualization Specifying precisely what we mean when we use particular terms Results in a set of indicators of what we have in mind Indicates a presence or absence of the concept we are studying Violent crime = offender uses force (or threatens to use force) against a victim

8. Indicators and Dimensions Dimension: Specifiable aspect of a concept “Crime Seriousness”: Can be subdivided into dimensions e.g., Dimension – Victim harm Indicators – Physical injury, economic loss, psychological consequences Specification leads to deeper understanding

9. Discussion Question 1 Why does the phrase “crime seriousness” require further conceptualization?

10. Confusion Over Definitions and Reality Concepts are abstract and only mental creations The terms we use to describe them do not have real and concrete meanings What is poverty? delinquency? strain? Reification: Process of regarding as real things that are not

11. Creating Conceptual Order Conceptual definition (what is SES?) Working definition specifically assigned to a term, provides focus to our observations Gives us a specific working definition so that readers will understand the concept Operational definition (how will we measure SES?) Spells out precisely how the concept will be measured

12. Discussion Question 2 What are some social science concepts that you believe might be hard to operationalize?

13. Operationalization Choices Operationalization: The process of developing operational definitions Moves us closer to measurement Requires us to determine what might work as a data-collection method

14. Measurement as “Scoring” Measurement: Assigning numbers or labels to units of analysis in order to represent the conceptual properties Make observations, and assign scores to them Difficult in CJ research because basic concepts are not perfectly definable

15. Exhaustive and Exclusive Measurement Every variable should have two important qualities: Exhaustive: You should be able to classify every observation in terms of one of the attributes composing the variable Mutually exclusive: You must be able to classify every observation in terms of one and only one attribute Example: Employment status

16. Discussion Question 3 Can you identify a variable used in crime studies that is both exhaustive and mutually exclusive?

17. Levels of Measurement Nominal: Offer names or labels for characteristics (race, gender, state of residence) Ordinal: Attributes can be logically rank-ordered (education, opinions, occupational status) Interval: Meaningful distance between attributes (temperature, IQ) Ratio: Has a true zero point (age, # of priors, sentence length, income)

18. Implications of Levels of Measurement Certain analytic techniques have Levels of Measurement requirements Ratio level can also be treated as Nominal, Ordinal, or Interval You cannot convert a lower Level of Measurement to a higher one Therefore, seek the highest Level of Measurement possible

19. Criteria for Measurement Quality The key standards for measurement quality are reliability and validity Measurements can be made with varying degrees of precision Common sense dictates that the more precise, the better However, you do not necessarily need complete precision

20. Reliability Whether a particular measurement technique, repeatedly applied to the same object, would yield the same result each time Problem: Even if the same result is retrieved, it may be incorrect every time Reliability does not insure accuracy Observer’s subjectivity might come into play

21. Dealing with Reliability Issues Test-retest method: Make the same measurement more than once—should expect same response both times Interrater reliability: Compare measurements from different raters; verify initial measurements Split-half method: Make more than one measure of any concept; see if each measures the concept differently

22. Validity The extent to which an empirical measure adequately reflects the meaning of the concept under consideration Are you really measuring what you say you are measuring? Demonstrating validity is more difficult than demonstrating reliability

23. Dealing with Validity Issues Face validity: On its face, does it seem valid? Does it jibe with our common agreements and mental images? Criterion-related validity: Compares a measure to some external criterion Construct validity: Whether your variable relates to another in the logically expected direction Content validity: Does the measure cover the range of meanings included in the concept? Multiple Measures: Alternative measures

24. Composite Measures Allows us to combine individual measures to produce more valid and reliable indicators Reasons for using Composite Measures: The researcher is often unable to develop single indicators of complex concepts We may wish to use a rather refined ordinal measure of a variable, arranging cases in several ordinal categories from very low to very high on a variable such as degree of parental supervision Indexes and scales are efficient devices for data analysis

25. Typologies “Taxonomy” Produced by the intersection of two or more variables to create a set of categories or types e.g., Typology of Delinquent/Criminal Acts (Time 1 and 2) None, Minor (theft of items worth less than $5, vandalism, fare evasion), Moderate (theft over $5, gang fighting, carrying weapons), Serious (car theft, breaking and entering, forced sex, selling drugs) Nondelinquent, Starter, Desistor, Stable, Deescalator, Escalator

26. Index of Disorder What is disorder? (Skogan, 1990) Distinguish between physical presence and social perception Physical disorder: Abandoned buildings, garbage and litter, graffiti, junk in vacant lots Social disorder: Groups of loiterers, drug use and sales, vandalism, gang activity, public drinking, street harassment Index created by averaging scores for each measure

27. Benefits of Indexes A composite index is a more valid measure than a single question Computing and averaging across all items in a category create more variation than we could obtain in any single item Two indexes are more parsimonious than nine individual variables Data analysis and interpretation can be more efficient