Should graphs of risk or rate ratios be plotted on a log scale?

Abstract

Should graphs of risk or rate ratios be plotted on a logarithmic scale? The conventional answer to this question seems to be yes ( 1), even to the extent that some have called on the International Committee of Medical Journal Editors to ban graphs of ratio measures that do not employ a log scale ( 2). The policy of the American Journal of Epidemiology nearly achieves this ban; the Instructions to Authors state, “When plotting relative measures of effect (e.g., relative risks, relative odds), a logarithmic scale must be used unless there is a compelling reason to use an arithmetic scale” ( http://aje.oxfordjournals.org/ …