Performance in task switching experiments is worse when the current stimulus is associated with different responses in the two tasks (i.e., incongruent condition) than when it is associated with the same response (i.e., congruent condition). This congruency effect reflects some sort of application of the irrelevant task's stimulus-response translation rules. Manipulating the recency and the proportion of congruent and incongruent trials results in a modulation of the congruency effect (i.e., Congruency Sequence Effect, CSE, and Proportion Congruency Effect, PCE, respectively), suggesting attentional adjustment of processing weights. Here, we investigated the impact of task switching practice on the congruency effect and the modulation thereof by (a) re-analyzing the data of a task switching experiment involving six consecutive sessions and (b) conducting a novel four-session experiment in which the proportions of congruent and incongruent trials were manipulated. Although practice appeared to reduce the reaction times overall and the task switch costs (i.e., slower reaction times after task switches than after task repetitions) to an asymptotic level, the congruency effect as well as its modulations remained remarkably constant. These findings thus do not provide evidence that conflict effects between tasks and attentional adjustment are affected by task switching practice.