NCJ Number
66833
Date Published
1980
Length
19 pages
Annotation
THE INTERRUPTED TIME SERIES DESIGN COMBINED WITH A QUASI-EXPERIMENTAL METHODOLOGY IS AN APPROPRIATE STRATEGY FOR TESTING THE EFFECT OF AN UNEXPECTED EVENT ON AN UNCONTROLLED STUDY POPULATION.
Abstract
SOCIAL SCIENCE RESEARCHERS OFTEN ARE FACED WITH SUCH PROBLEMS AS MEASURING THE EFFECT OF SOME TYPE OF LEGISLATION, NATURAL DISASTER, ETC., (OVER WHICH THEY HAVE NO CONTROL) ON THE BEHAVIOR OF A POPULATION THAT IS NOT RANDOMLY SELECTED AND IS IMPOSSIBLE TO CONTROL FOR CERTAIN VARIABLES. THE CLASSICAL EXPERIMENTAL DESIGN IS NOT FEASIBLE IN THIS TYPE OF SITUATION, AND THE RESEARCHER HAS NO OR LIMITED DATA TO CONTROL FOR THE SIX COMMON THREATS TO VALIDITY--HISTORIC EVENTS OTHER THAN THE EXPERIMENTAL TREATMENT THAT MIGHT ACCOUNT FOR THE CHANGE, MATURATION OR REGULAR CHANGES WITH TIME PASSAGE, CHANGE RESULTING FROM THE PRETEST OR MEASUREMENT, SHIFT IN MEASURING INSTRUMENT INDEPENDENT OF CHANGE, AND REGRESSION. AS A RESULT, MANY RESEARCHERS WILL USE THE ONE-GROUP PRETEST-POSTTEST DESIGN IN INTERRUPTED TIME SERIES ANALYSES. TO ILLUSTRATE THE WORKINGS OF THE INTERRUPTED TIME SERIES DESIGN, THIS ARTICLE EXAMINED THE EFFECT OF THE 1955 CRACKDOWN ON SPEEDING IN THE STATE OF CONNECTICUT. OFFICIAL DATA ON TRAFFIC ACCIDENTS AND FATALITIES FOR THE STATE ARE GRAPHED. PRETREATMENT REFERS TO THOSE YEARS BEFORE 1955 WHEN THE GOVERNOR CALLED FOR STRICTER PENALTIES AND INCREASED CONVICTIONS FOR SPEEDING. THE YEAR 1955, THEN, WAS THE TREATMENT YEAR, WHILE LATER YEARS WERE POSTTREATMENT. A MULTIPLE TIME SERIES STUDY HELPS TO FURTHER CLARIFY THE EVENTS OCCURRING IN CONNECTICUT. THIS DESIGN USES DATA FOR SEVERAL STATES COVERING THE SAME PERIOD. CONNECTICUT IS THE EXPERIMENTAL UNIT. ALTHOUGH THE INTERRUPTED TIME SERIES DESIGN LEAVES MANY QUESTIONS UNANSWERED REGARDING CHANGES AND THEIR ORIGIN, ITS RESULTS CANNOT BE DISCREDITED WITHOUT BEING DISPROVEN IN SOME OTHER SCIENTIFIC STUDY. FOOTNOTES AND GRAPHS ARE PROVIDED. (AOP)