Interest in cognitive control processes is large not only in basic but also in applied research such as clinical psychology, where individual differences in conflict processing are seen as a factor underlying different types of psychopathology. Nevertheless, it remains an open question whether behavioral measures of conflict- and control-related effects reflect stable, trait-like abilities. Here, we investigated the reliability of the congruency effect and its modulation by proportion congruency in the temporal flanker task (n = 32). In line with previous findings, analyses revealed pronounced response time increases when distractor and target were incongruent compared to congruent trials (indicating reliability on the group level). Split-half (odd-even) reliability measures of this congruency effect amounted (Spearman-Brown corrected) to .80 and larger. Reliability for the proportion congruency effect, i.e., the increase of the congruency effect with increasing proportion congruency (which is a difference of a difference), however, dropped to corrected rs between .44 and .76. Given the good split-half reliability at least for the congruency effects, we also estimated retest reliability and thus stability of the measures by means of a second session after 5 months. Preliminary data of 16 participants yielded retest-reliability estimates of r>.75 for the congruency effects and of r>.63 for the proportion congruency effects. Implications for the appropriateness of the temporal flanker task to measure individual differences in cognitive control will be discussed.