The sanctions most studied in macro level research which uses aggregate data are arrest, prison commitment, and police presence. The NRC report, which reviewed time series and cross section regression analyses, objected to the standard conclusions that arrest and imprisonment rates had deterrent effects on three grounds. These include the spurious relationships caused by ratio variable, the confounding of deterrence and incapacitation effects, and causal direction problems. Seventeen studies published in the 1980's dealing with the relationship between crimes and arrests, imprisonment, or police presence have had to deal with the issue of simultaneity or likely reciprocal causation problems. The author maintains that most researchers have either used lagged independent variables, attempts to explain away simultaneity, simultaneous equations, or the Granger test. 3 notes, 1 figure, and 25 references