CoChat

Evidence and Inference in the Comparative Case Study

Dion, Douglas · 1998 · Comparative Politics

The DOI matches its Crossref record; 3 of the 3 references it cites were found in a registry or an index.

Abstract

Comparative case studies often rely on a practice known as selecting on the dependent variable. This technique involves choosing some phenomenon of political interest, gathering data on occurrences of the phenomenon, then determining what characteristics the occurrences have in common. A methodological strategy with a long history, selecting on the dependent variable has found numerous applications in political science, including the study of economic growth, social revolutions, and international conflict.' Selecting on the dependent variable has provided an important method for gathering information when there is relatively little data, though it is not limited to such situations. Its use has recently come under rather sharp attack. If observations are selected based on the value of the dependent variable, then estimates obtained by ordinary least squares will be biased. This bias can not be corrected by the introduction of control variables. Gathering more data will not solve the problem; the bias remains even if the number of observations goes to infinity. In short, inferences drawn from cases selected on the dependent variable are highly suspect.2 Impressive treatments of the problem from the position of comparative case study work have been presented by Barbara Geddes and by Gary King, Robert Keohane, and Sidney Verba.3 They have not only alerted the discipline to an important problem but have suggested concrete ways to deal with problems of selection bias.4 I would like to offer a qualified defense of analyses of a small number of cases selected on the dependent variable. As is well-known, selecting on the dependent variable is perfectly admissible if one is evaluating necessary (as opposed to sufficient) conditions.5 However, with a few exceptions, this argument has gone unnoticed in the debate on case selection.6 By concentrating on the logic of testing necessary conditions, this paper attempts to show that the methodological implications of selecting on the dependent variable are much less threatening to the validity of small-n comparative case study work than has often been claimed.

Overview

Integrity signals

Registry

Crossref · 10.2307/422284

Cites

3 of 3 references verified0 citation errors · 0 not checkable

Statistics

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

Verification

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?

Citations

Cited by

Items 101–200 of 263

Metadata

Authors
Dion, D.
Corresponding author
Dion, Douglas
Venue
Comparative Politics
Publisher
JSTOR
Published
1998-01
Volume
30
Issue
2
Pages
127
ISSN
0010-4159 (ISSN-L)
OpenAlex
W2017659746
Cited by
261
References
3
Citations by year
2026: 5 · 2025: 4 · 2024: 2 · 2023: 10 · 2022: 2 · 2021: 1 · 2020: 7 · 2019: 9 · 2018: 7 · 2017: 11 · 2016: 7 · 2015: 12 · 2014: 10 · 2013: 12 · 2012: 13
FWCI
20.37
Type
journal article
Language
en
Open access
closed
Locations
Comparative Politics (publishedVersion)
Topics
  • Social Sciences › Sociology and Political Science › Qualitative Comparative Analysis Research
  • Social Sciences › Political Science and International Relations › Political Science Research and Education
  • Social Sciences › Sociology and Political Science › Political Conflict and Governance
Keywords
selection bias, comparative case study, necessary conditions, sufficient conditions, small-N research, ordinary least squares bias, case selection, validity of inference

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

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.