DATA SHOW A HIGH CORRELATION BETWEEN COLORADO'S UNEMPLOYMENT RATE AND THE NUMBER OF NEW COURT COMMITMENTS. THUS, UNEMPLOYMENT RATES, STATE POPULATION FIGURES AND SEASONAL VARIATIONS IN COMMITMENTS AFFECT THE TOTAL PRISON POPULATION. THEY WERE COMBINED WITH PAROLE REVOCATION RATES AND PERCENTAGES OF DETERMINATE SENTENCES TO PREDICT THE PRISON POPULATION THROUGH 1980. HOWEVER, DATA COMPILED IN 1976 AND 1977 INDICATED THAT COMMITMENT PROJECTIONS WERE 6.9 PERCENT IN 1976 AND 14 PERCENT IN 1977. THESE INORDINATE ERRORS WERE ASCRIBED TO UNDERESTIMATING THE UNEMPLOYMENT RATE IN THE MODEL AND ASSUMING NO DIVERSION WOULD OCCUR. ALSO A HIGH CORRELATION BETWEEN UNEMPLOYMENT AND COMMITMENT IN THE MODEL WAS FOUND TO BE ERRONEOUS, THE ACTUAL RELATIONSHIP BEING INSIGNIFICANT. WITH THE LOSS OF THIS RELATIONSHIP, THE UNEMPLOYMENT RATE, OR DIVERSION RATE, ON THE ACTUAL NUMBER OF COMMITMENTS CANNOT BE QUANTIFIED DUE TO THE AMOUNT OF ERROR ASSOCIATED WITH THE MODEL. FURTHER, THE MODEL DID NOT ACCOUNT FOR NEW PENAL LEGISLATION, CHANGES IN RATES OF PAROLE REVOCATION, AND OTHER MATTERS. A SECOND PROJECTION ATTEMPT, AIMED AT 1977-1981 POPULATIONS, RESULTED IN A 3.8 PERCENT UNDERPROJECTION FOR 1977, DUE TO ERRORS IN ESTIMATING THE UNEMPLOYMENT RATE, AVERAGE PRISON STAY, AND UNFORESEEN EVENTS SUCH AS A CHANGE IN ANY ASSUMPTION USED TO MAKE A PROJECTION, WHICH WILL MAKE BOTH SHORT-TERM AND LONG-TERM PROJECTIONS DIFFICULT. GRAPHS AND TABLES ARE INCLUDED.