This study was to investigate whether the application of different diagnostic criteria affected the frequency of convergence insufficiency (CI). Methods: Eighty one subjects with mean age of 22.54 years (20 to 27 years) were evaluated. Binocularity tests after refraction were performed as the following tests: near point of convergence (NPC) with an accommodative target, phoria using von Graefe method, positive fusional vergence (PFV) with a phoropter. Subjects with CI were diagnosed when exophoria (exo) was greater at near than at distance (exo≥4 Δ, or >6 Δ), fusional vergence was PFV≤11 Δ for blur, PFV≤15 Δ for break, Sheard's or Percival's criterion, and NPC was NPC≥6 cm, ≥7.5 cm or >10 cm. Results: Frequency of CI with one diagnostic criterion was ranged from 6.2% to 77.8%, and was overestimated or underestimated according to criteria. It was reduced to the range of 6.2% to 43.2% with diagnostic criteria more than two, especially to the range of 24.7% to 28.4% with lower variability in diagnostic criteria including phoria and Sheard's criterion. There were high relationship between total score of signs and phoria score (r = 0.772, p<0.001), and measured phoria and Sheard's criterion (r = -0.654, p<0.001), but NPC had a high variability and a weak or no significant relationship with other diagnostic criteria. Results suggested exo≥4 Δ, Sheard's criterion and NPC≥7.5 cm for diagnostic criteria of signs and sequence for CI. Conclusions: Frequency of CI is likely to be over-and underestimated with diagnostic criteria. Cutoff values and procedures for phoria, Sheard's criterion and NPC as clinical signs should be suggested definitely in diagnosis associated with CI.