Hostname: page-component-78c5997874-ndw9j Total loading time: 0 Render date: 2024-11-08T00:30:13.809Z Has data issue: false hasContentIssue false
  • English
  • Français

Analysis of Data on the Connecticut Speeding Crackdown as a Time-Series Quasi-Experiment

Published online by Cambridge University Press:  01 July 2024

Gene V Glass
Gene V Glass*
Affiliation:
Laboratory of Educational Research, University of Colorado
Rights & Permissions [Opens in a new window]

Extract

Core share and HTML view are not available for this content. However, as you have access to this content, a full PDF is available via the ‘Save PDF’ action button.

In late 1955 in Connecticut, the number of fatalities per 100,000 population in motor vehicle accidents reached a record high for the 1950s. On December 23, 1955, Governor Abraham Ribicoff took unprecedented action to reduce traffic fatalities. Ribicoff announced that persons convicted of speeding would have their licenses suspended for thirty days at the first offense, for sixty days at the second offense, and for an indefinite period (subject to a hearing after ninety days) at the third offense. Data on traffic fatalities before and after the Connecticut crackdown on speeding can be regarded as a time-series quasi-experiment with some significance for the social sciences. When supplemented with traffic fatality data for the states of Massachusetts, Rhode Island, New York, and New Jersey, the collection of observations can be viewed as a multiple-group time-series experiment.

Type
Research Article
Information
Law & Society Review , Volume 3 , Issue 1 , August 1968 , pp. 55 - 76
Copyright
Copyright © 1968 by the Law and Society Association

Footnotes

Author's Note: The work reported in this paper was supported as Project No. 6-8329 of the U. S. Office of Education. The author is indebted to Donald T. Campbell of Northwestern University and H. Laurence Ross of the University of Denver Law School who supplied the data which are analyzed in this paper. See D. T. Campbell & H. L. Ross, The Connecticut Crackdown on Speeding, 3 Law & Soc. Rev. 33-53 (1968).

References

1. D. T. Campbell & J. C. Stanley, Experimental and Quasi-Experimental Designs for Research on Teaching, in Handbook of Research on Teaching (N. L. Gage ed. 1963).

D. T. Campbell, From Description to Experimentation: Interpreting Trends as Quasi-Experiments, in Problems in Measuring Change, ch. 12 (C. W. Harris ed. 1963).

2. Id.

3. G. E. P. Box & G. C. Tiao, A Change in Level of a Non-stationary Time-Series, 52 Biometrika 181-92 (1965).

4. G. E. P. Box & G. M. Jenkins, Some Statistical Aspects of Adaptive Optimization and Control, 24 J. Royal Statistical Soc'y B, 297-343 (1962).

5. Correlograms are sets of correlation coefficients obtained by pairing the observations in a time-series in different ways and calculating the correlation coefficients that result.

6. M. S. Bartlett, On the Theoretical Specification and Sampling Properties of Autocorrelated Time-Studies, 27 J. Royal Statistical Soc'y B (1946).

7. Id.

8. In this instance, i.e., γ = 0, the model is equivalent to the model for the “t-test” employed in elementary statistics, and the analysis is equivalent to a t-test on pre-treatment v. Posttreatment observations.

9. Inspection of the graphs is facilitated by the dotted lines which mark off the values of t (df = 106) required for significance at the .01, .05, .10 and .15 levels for a one-tailed test of the hypothesis that δ = 0. For the four control states the alternative hypothesis is that δ > 0.

10. Bear in mind that no “treatment” was actually applied in these states.

11. M. S. Bartlett, Tests of Significance in Factor Analysis, 3 Brit. J. Psychology, statistical section, 77-85 (1950).

12. W. L. Hays, Statistics for Psychologists, ch. 14 (1963).