In an attempt to understand the apparently disparate results reported in magnetic field effect studies on the same reaction system from different experimental groups, two different techniques, time-resolved methods and field-modulated ones, have been applied to systems comprising pyrene and either dicyanobenzene or dimethylaniline. It is shown that these yield different results whose origin lies less in the detection method than in the light sources which produce the radicals. The results from the time-resolved experiments, in which the effects of spin relaxation in a singlet-correlated pair are apparent, have been analyzed using a novel approach applicable to more complex reaction systems than have been studied in the past. An original experiment is introduced to observe the effects of relaxation at field strengths where the hyperfine-driven spin mixing has reached its asymptotic rate. The discrepancy between the results originates in a much shortened radical pair lifetime, resulting from a very low radical concentration under the conditions used in the modulation experiment as opposed to the time-resolved one. This lifetime effect has been confirmed by using degenerate electron-exchange reactions to affect the chemistry in the period of observation in each technique. Low-field minima in the field effect curves observed using the modulation method are shown to originate in hyperfine and effects and not to be due to J-resonances.