4.3
© 2014 by Pearson Higher Education, Inc Upper Saddle River, New Jersey 07458 • All Rights Reserved
Class Name,
Instructor Name
Date, Semester
Hagan, Research Methods in Criminal Justice and Criminology, 9/e
Chapter 9
Validity, Reliability, and Triangulated Strategies
*
*
Validity, Reliability, and Triangulated Strategies
9
Error is another term for invalidity.
Sources of potential error or invalidity are always present, even in some of the best research.
Error in Research
*
*
Validity, Reliability, and Triangulated Strategies
9
Validity
Accuracy of measurement
Does the instrument in fact measure that which it purports to measure?
Reliability
Consistency and/or stability of a measuring instrument
Validity, Reliability, and Triangulated Strategies
9
- Little professional esteem in replication
- Lack of complexity in technique
- Design faults in original study
- Unfavorable climate
- Interjurisdictional disputes
- Tradition
Reasons for Lack of Validation Studies in Criminal Justice
*
Validity, Reliability, and Triangulated Strategies
9
Construct validity
Content validity
Face validity
Convergent-discriminant validity
Pragmatic validity
Types of Validity
*
© 2014 by Pearson Higher Education, Inc Upper Saddle River, New Jersey 07458 • All Rights Reserved
9
*
*
Validity, Reliability, and Triangulated Strategies
9
Face Validity
- The simplest measurement
- Asks “Does the measuring instrument appear, at face value, to be measuring what I am attempting to measure?”
Types of Validity
*
Validity, Reliability, and Triangulated Strategies
9
Content Validity
Examines each item—the content of an instrument—to judge whether each element measures the concept in question
- Similar to face validity, content validity is also judgmental and usually nonempirical in nature.
- However, an item analysis may be employed to eliminate nondiscriminatory items.
Types of Validity
*
Validity, Reliability, and Triangulated Strategies
9
Construct Validity
- Sometimes called concept validity
- Asks whether the instrument in question does, in fact, measure what it has been designated to measure
*
Types of Validity
Validity, Reliability, and Triangulated Strategies
9
Pragmatic Validity
Asks “Does it work?”
Two types:
- Concurrent Validity: whether the measure enhances the ability to gauge present characteristics of the item in question
- Predictive Validity: concerns the ability to accurately forecast future events or conditions
*
Types of Validity
Validity, Reliability, and Triangulated Strategies
9
Convergent-Discriminant Validation (also called Triangulation)
Involves the use of multiple measures to measure multiple traits
- Convergence: different measures of the same concept should yield similar results
- Discrimination: the same measure of different concepts should yield different results
*
Types of Validity
*
© 2014 by Pearson Higher Education, Inc Upper Saddle River, New Jersey 07458 • All Rights Reserved
9
*
*
*
Validity, Reliability, and Triangulated Strategies
9
Reliability
Demonstrated through stable and consistent
replication of findings on repeated measurement
Test-Retest
A means of determining reliability in which the same
instrument is administered twice to the same population
Multiple Forms
Involves the administration of alternate
forms of the instrument to the same group
Split-Half Reliability
A form is administered one time to one group,
then split in half and inter-correlated
*
Validity, Reliability, and Triangulated Strategies
9
Drug Use Forecasting (DUF)
A research program sponsored by the National Institute of Justice (NIJ)
Asked volunteers from a population of arrestees in various cities to provide urine specimens that are then tested for drugs
Adam (Arrestee Drug Abuse Monitoring Program)
*
*
© 2014 by Pearson Higher Education, Inc Upper Saddle River, New Jersey 07458 • All Rights Reserved
The question of error unreliability and invalidity is central to criminal justice research.
Validity refers to the accuracy of findings, whereas reliability is concerned with consistency and predictability of research.
In its assessment of the state of Criminal Justice Research and Development, the National Advisory Committee on Criminal Justice Standards and Goals partly explains why there has been a lack of validation studies in criminal justice.
The principal measures of validity are face validity, content validity, construct validity, pragmatic validity (which includes concurrent and predictive validity), and convergent-discriminant validity (triangulation).
CHAPTER SUMMARY
*