A robust left digit effect arises in number line estimation such that numerals with different leftmost digits but nearly identical magnitudes (e.g., 298 and 302) are placed farther apart on a number line than is warranted. Past efforts to motivate individuals to perform more accurately on the number line estimation task using trial-by-trial feedback and summary accuracy feedback have not led to a reduction in the left digit effect. In the present study, we asked whether it is possible to reduce or eliminate the left digit effect in number line estimation through an instructional intervention in which one is explicitly taught about the left digit effect. All participants (N = 143) completed two blocks (60 trials per block) of a self-paced 0-1000 number line estimation task and were randomly assigned to either an instructional or a control condition. While there was a large left digit effect in each block in each condition, the intervention did not lead to a reduction in the left digit effect in the instructional condition relative to the control condition despite participants in the instructional condition indicating greater effort, reporting higher confidence ratings, and dedicating more time to the task than those in the control condition. These findings illustrate the robustness of the left digit effect in number line estimation and demonstrate that the bias cannot be easily corrected, even when one is made aware of it.