124-126

Operationalization must be guided by the precision required.
Clarity of needs is essential; favor precision over vagueness to avoid inaccuracies.
Reliability: consistent measurements do not guarantee accuracy.
- Example: A faulty speedometer may show a consistent reading (e.g., 65 mph) while the actual speed is different (e.g., 70 mph).
- Reliable measurement does not equate to accurate measurement, particularly in criminal justice research.

Criminal justice research often encounters reliability issues, particularly with crime reported by police records.
A reliable measurement must yield the same results when applied repeatedly to the same target.
Example of measurement reliability problems: burglary definitions varied among patrol officers and detectives.
Detectives provided more consistent results due to a clearer understanding of the official definition of burglary.

Distinction between precision and accuracy:
- Precision: describes how well a measurement captures detail.
- Accuracy: reflects the truthfulness of the information.
- Example: Describing a birthplace as "Stowe, Vermont" is precise but may not be accurate if the person was born in "Boston".

Involves repeating measurements to check for consistency over time.
Example: West and Farrington (1977) examined responses from interviews at ages 16 and 18, finding high consistency for questions about school attendance.
Memory lapses can cause discrepancies in answers over long gaps between tests.

Refers to consistency among different observers or measure administrators.
Reliability issues can arise due to subjective judgment by individual observers.
Example: Interviewers may influence responses based on their demeanor or biases.

Involves dividing test items into two sets and comparing results to assess reliability.
Used in studies to measure complex social concepts like prejudice and intelligence.
Example: Woodward and Fergusson (2000) demonstrated a reliability value of 0.93 using split-half methodology.

Judging Reliability: Evaluating how reliably judges assess defendants' risk of flight before trial.
- Study showed that judges' evaluations were not reliable predictors of failure to appear in court.
Forensic Reliability: DNA evidence collection must avoid contamination and analyst bias; errors can lead to wrongful convictions.
Observer Reliability: Unique perceptions of a single observer can skew data representation.
- Important when aggregating categories or coding information (e.g., political organizations).

Validity is crucial for ensuring that measurements accurately reflect what they intend to measure.
In criminal justice, even high reliability still requires assessment to ensure the measure accurately represents the constructs of interest.