Editor's Note: Replications and Extensions
Political Analysis will begin accepting submissions for a special section of the journal reserved for replications and extensions. Pieces accepted for this section will be published in full on the Political Analysis web site, with an abstract of the piece being published in the hard copy version of the journal.
This section of the journal continues Ken Meier's replication section in the American Journal of Political Science (which has now been discontinued). The requirements for submitting to this section are similar to those published by Meier in the May, 1996 AJPS.
Four types of submissions are encouraged:
1. Simple confirmatory replications. These should simply indicate that some published study was successfully replicated. The write-up of such replications should simply indicate that an article was successfully replicated (with the article and replicator's name being put in a file of successful replications available on the web site).
2. Failure to replicate. Authors only need present analyses for results which could not be replicated (with a note that other analyses were successfully replicated). Authors should make sure they are using identical data and methods, and that failure to replicate is not due to inability to get identical results to ten decimal places. Replicators in this category \emph{contact} the original author to attempt to iron out differences. The replication should contain both the original results and the new results, along with any reasons for the failure to replicate. Submissions in this category must include enough material so that referees can evaluate the failure to replicate. Replicators in this category also must ensure that referees have access to the appropriate software.
3. Replications with extensions. These begin by clearly replicating the original findings of some article. They then consider whether the findings are robust to various methodological changes. These replications look at the robustness of findings with respect to methodological decisions and so should restrict themselves to the original author's data. Authors of extensions should highlight how results vary with choice of method. The introduction of a new method, with examples, is an article, not a replication.
4. Cross-validation studies. These replications extend the findings of some article to new data sets (e.g. countries or years). They begin by replicating the findings of the original article with the original data, but then examine how the results vary with new data sets. Cross-validation studies may also consider the impact of methodological decisions, but they must contain some data not considered by the original author.
Replication authors should deposit one copy of their replication data with the ICPSR and submit another for the Political Analysis web site.
Replication articles should be brief and to the point. No long introduction or conclusion is wanted, though the author should indicate why the replicated study is important. The replication articles should contain complete details of the data and analysis, and should be understandable by those who have not read the original article.
Simpler replications will be refereed as a group; more complex replications will be sent out to one or more referees as required.
Accepted replications will be published on the Political Analysis web site as pdf files. The articles must conform to all the style requirements of the journal. The editorial staff will assist authors with this task, but final preparation of the manuscript is the responsibility of the authors.
Those wishing to cite the replications should cite the abstract in the hard copy of Political Analysis.
Return to Information for Authors.