This study investigated how well students differentiate their responses' accuracies (metacognitive monitoring) and estimate their test scores beyond counting—and counting on—the number of correct responses alone. Monitoring abilities of 2832 sixth‐graders (1410 male and 1422 female native in Turkish) at an 11‐item Program for International Student Assessment (PISA)‐equivalent mathematics test were measured via response‐contingent Type‐2 signal detection theory. The students also made score estimations right before and immediately after completing the test (pre‐ and posttest estimations, respectively). Although high‐scoring students underestimated and low‐scoring ones overestimated how they would perform in the test, high‐scorers were accurate in their posttest estimations unlike the low‐scoring group, where the lattaer retained their overestimation tendencies. Having better monitoring performance, the high‐scoring group could subsequently calibrate their posttest estimations. Additional assessment methods such as measuring monitoring and score estimations seem to have the potential to reveal how mathematics students behave before, during, and after responding.