CoChat
Sign Up
Cited by 124Advanced Causal Inference TechniquesCriminal Justice and Corrections AnalysisOptimal Experimental Design Methods

Design Sensitivity in Criminal Justice Experiments

Weisburd, David, Petrosino, Anthony, Mason, Gail·1993·Crime and Justice
DOI 10.1086/449216
Open this paper in CoChat — free account

Abstract

Interest in randomized experiments with criminal justice subjects has grown, in recognition that experiments are much better suited for identifying and isolating program effects than are quasi-experimental or nonexperimental research designs. Relatively little attention, however, has been paid to methodological issues. Using the statistical concept of power-the likelihood that a test will lead to the rejection of a hypothesis of no effect, a survey examines the design sensitivity of experiments on sanctions. Contrary to conventional wisdom advocating large sample designs, little relationship is found in practice between sample size and statistical power. Difficulty in maintaining the integrity of treatments and the homogeneity of samples or treatments employed offsets the design advantages of larger investigations.

Integrity signals

Registry

Crossref · 10.1086/449216

Cites

2 of 2 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Claims the paper makes

This paper’s full text isn’t openly available, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Studies with very small sample sizes may exhibit higher statistical power compared to those with very large sample sizes.

    “In this regard, it has been found that studies with very small sample sizes may exhibit higher statistical power compared to those with very large sample sizes (Weisburd et al., 1993).”

    From ‘Led by Intelligence': A Scoping Review on the Experimental Evaluation of Intelligence-Led Policing (2024) · cites this paper as Weisburd et al., 1993

  • Weisburd's paradox highlights a discrepancy between the expectation that increasing sample size will enhance statistical power and the reality that it often results in decreased statistical power.

    “However, from an empirical point of view, this premise must be balanced against Weisburd's paradox (Groff et al., 2015;Weisburd et al., 1993), which highlights the discrepancy between the expectation that increasing the sample size will enhance statistical power, and the reality that it often results in decreased statistical power.”

    From ‘Led by Intelligence': A Scoping Review on the Experimental Evaluation of Intelligence-Led Policing (2024) · cites this paper as Weisburd et al., 1993

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Authors
Weisburd, D., Petrosino, A., & Mason, G.
ORCID
Venue
Crime and Justice
Publisher
University of Chicago Press
Published
1993-01
Volume
17
Pages
337-379
ISSN
  • 0192-3234 (ISSN-L)
  • 2153-0416
DOI
10.1086/449216
OpenAlex
W2012992066
Cited by
124
References
2
Citations by year
2026: 3 · 2025: 5 · 2023: 5 · 2022: 3 · 2021: 7 · 2020: 5 · 2019: 6 · 2018: 7 · 2017: 5 · 2016: 2 · 2015: 2 · 2014: 6 · 2013: 7 · 2012: 4
FWCI
1.52
Type
journal article
Language
en
Open access
closed
Locations
Crime and Justice (publishedVersion)
Topics
  • Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
  • Social Sciences › Sociology and Political Science › Criminal Justice and Corrections Analysis
  • Social Sciences › Decision Sciences › Management Science and Operations Research › Optimal Experimental Design Methods
Keywords
design sensitivity analysis, randomized experiments, statistical power, sample size, sanctions, implementation fidelity, sample homogeneity, program evaluation

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review