DATA USED IN THE ANALYSIS ARE COMPRISED OF COMPUTER RECORDS OF ALL INMATES RELEASED FROM THE FLORIDA DEPARTMENT OF OFFENDER REHABILITATION DURING FISCAL YEARS 1974 THROUGH 1975 AND 1975 THROUGH 1976. THE STUDY IS PART OF A RESEARCH EFFORT DIRECTED TOWARD ESTIMATION OF FUTURE INMATE POPULATION UTILIZING THE THREE PRIMARY FACTORS OF ADMISSION, TIME SERVED, AND RELEASES. THE NULL HYPOTHESIS OF THE STUDY IS THAT THERE IS NO ASSOCIATION BETWEEN THE INDEPENDENT VARIABLES OF OFFENSE AND LENGTH OF SENTENCE AND THE DEPENDENT VARIABLE OF TIME SERVED. BECAUSE THE STUDY DEALS WITH THE INTERACTION OF THREE FACTORS, MANY ALTERNATIVE HYPOTHESES WERE POSSIBLE AND WERE CONSIDERED. AFTER EXAMINING THE PROBLEM, IT WAS DETERMINED THAT THE BEST ANALYTIC TECHNIQUE TO BE APPLIED WAS THAT OF MULTIPLE REGRESSION. SINCE ONE OF THE VARIABLES TO BE EXAMINED, OFFENSE, WAS NOMINAL IN NATURE, A DUMMY VARIABLE STRATEGY WAS USED TO PLACE THIS VARIABLE IN THE REGRESSION EQUATION. ANALYSIS REVEALED THAT BOTH OFFENSE AND LENGTH OF SENTENCE WERE SIGNIFICANTLY ASSOCIATED WITH TIME SERVED. IT WAS FURTHER FOUND THAT BOTH LENGTH OF SENTENCE AND OFFENSE RELATED IN SUCH A WAY THAT, WHILE BOTH WERE OF SIGNIFICANT HELP IN PREDICTING TIME SERVED, THE SUM TOTAL OF THEIR PREDICTIVE ABILITY WAS LITTLE BETTER THAN LENGTH OF SENTENCE TAKEN ALONE. LENGTH OF SENTENCE, THEREFORE, WAS THE BEST PREDICTOR OF TIME SERVED. FOOTNOTES ARE PROVIDED IN THE STUDY. (LWM)