THE FBI's statistical "Crime Clock" that ticks off the figures -- "One violent crime [in 1979] every 27 seconds, one murder every 24 minutes, one forcible rape every 7 minutes, one robbery every 68 seconds" -- sounds ominous indeed. But has there really been a dramatic recent increase in violent crimes committed in the United States if we go by the numbers?

The answer will depend on which of two conflicting sets of crime statistics you choose, one more alarmist than the other. The FBI's annual Uniform Crime Report, which collects statistics from the nation's police departments, recorded an overall 11 percent growth in violent crimes in 1979 over the previous year, and its 1980 statistics report an overall 13 percent jump last year, the highest rate of increase since 1968.

Across the street on Pennsylvania Avenue, however, the Justice Department's Bureau of Crime Statistics followed release of the FBI figures by unveiling its latest National Crime Survey of "households . . . touched by crime during 1980." Based on interviews with 132,000 people, the so-called "victimization" survey found all crimes, including violent ones, remaining "relatively stable since 1974 [when the survey began], with just under one-third of all households in the nation affected by crime."

The apparent disparity between the FBI's portrait of a growing violent crime wave in the country and the National Crime Survey's failure to reflect any dramatic increase reflects the limitations of both instruments for measuring crime. The FBI's Uniform Crime Reports have always been forced to cope with the inadequate and uneven reporting procedures of local police or sheriffs, in situations where either political pressures or changes in record-keeping can result in specific crimes being either under- or over-reported. The starkest example of this among violent crimes has involved instances of forcible rape, in previous decades an under-reported crime, which increased (according to the Uniform Crime Report's figures) over 100 percent during the 1970s -- at a time when police began dealing with rape victims more sympathetically -- compared with a 33 percent increase in robberies and a 34 percent increase in murders during the decade.

The National Crime Survey's statistics have their own problems. By its own evidence, for example, two out of three violent crimes go unreported to the police (and how many more to the NCS's own interviewers?). The truth of an observation in a 1976 study by a National Academy of Sciences' panel on surveying crime -- that "victimization surveys are concerned not only with crime, but also with societal reaction to it" -- deserves careful reflection. The fact that some Americans report violent crimes while many more do not suggests broad disagreement within the society on such questions as what constitutes permissible as opposed to criminal behavior, should crime victims trust the police and the courts, and whose interests are served -- and whose hurt -- by the criminal justice system.

The limitations contained in both major sets of crime statistics are examined in a provocative op-ed article today by Brian Forst, in which he uses the homicide index to make a more general point about increases in crime. Undeniably, Mr. Forst is correct in arguing that the murder rate deserves closer scrutiny in overall surveys of crime, despite the self-evident fact that "homicide victims make poor survey respondents." Still, it could be argued that the figures on murder -- though distinctive -- contain a special impurity of their own as an indicator of overall violent crime because of the close links between so many offenders and victims. In 1979, for example, 52 percent of murder victims knew their assailants and, in 20 percent of the cases, one relative murdered another.

At the same time, comparison of the 1969 and 1979 statistics on homicide (which differ in some categories used) suggests in some rough, overall way that the number of killings performed by total strangers may have risen significantly over the past decade. Perhaps it is sufficient to recall at this point that, whether or not violent crimes have been increasing greatly since 1978, few statistical surveys find any decrease. The crucial question does not concern the precise percentage of any increase but the public's clear perception that an enormous increase in violent crimes has occurred. The fact that the ordinary American now believes a random, vicious criminal incident will disrupt unexpectedly the moral fabric of his or her life deserves the most careful examination -- both of the belief and its consequences -- but on another day.