NCJ Number
91063
Journal
Journal of Police Science and Administration Volume: 11 Issue: 3 Dated: (September 1983) Pages: 282-289
Date Published
1983
Length
8 pages
Annotation
This study examines the quality and quantity of statistical procedures in eight of the leading criminal justice journals.
Abstract
The period selected for analysis was 1976-80. A total of 2,390 articles were reviewed, with 53 percent containing no quantitative data. Each journal article was coded for the presence of quantitative data, the characteristics of the data, and the type of statistical or mathematical procedure used. The prevalent type of data used by the researchers was nominal data, with ordinal data found in a small percentage of the quantitative articles, while only 16 percent used interval-ratio data. The trend in the use of more powerful data (interval or ratio) appears to be increasing, however. Only 20 percent of the articles with quantitative data used some form of probability sampling. In the other quantitative articles, data were obtained on a nonrandom, accidental, or judgmental basis. About 50 percent of the quantitative articles relied on available subjects. Although the sample and population sizes varied considerably, the greater percentage of quantitative articles used limited population sizes. Descriptive analysis was prevalent in all eight of the journals. In 22 percent of the articles analyzed, one or more statistical procedures were observed. The procedures are categorized as simple, descriptive parametric, inferential parametric, nonparametric, mathematical, scaling, and miscellaneous procedures. The absence of experimental design and more advanced statistical procedures may be due to the lack of statistical training for criminal justice researchers or the obstacles posed by the complex criminal justice system. Still, given the problems associated with criminal justice data, researchers could make better use of statistical procedures. Tabular data and 29 references are provided.