This study examined problem solving as a function of display type (table/graph) and numeric function (linear/nonlinear) in four problem solving domains. Twenty-two stimulus problems were developed, and line graph and table displays were constructed for each problem. Half of the displays contained linear numeric functions and the other half contained nonlinear numeric functions. Each display was accompanied by four questions which required: (1) the location of a specific value; (2) trend analysis; (3) data interpolation; and (4) a forecasting decision. Each of the 109 subjects completed six practice and 16 experimental problems. Although overall decision time between graphical and tabular displays did not differ, the subjects performed better with tabular displays when locating specific values and better with graphical displays when interpolating, forecasting, or judging data trends. Overall, subjects performed faster and more accurately for both types of display when solving problems using nonlinear functions. Tables of experimental results are appended, and 25 references are listed. (MES)