This paper discusses the claim made in Altonji and Pierret (1997) and Lange (2005) that a high speed of employer learning indicates a low value of job market signaling. The claim is first discussed intuitively in light of Spence’s original model and then evaluated in a simple extension of a model developed in Altonji and Pierret (1997). The analysis provided indicates that, if employer learning is incomplete, a high speed of employer learning is not necessarily indicative of a low value of job market signaling.