A better description of this analysis is that a regression model using p/e ratio as a descriptor of stock returns is insufficient. But that would be less likely to entice anyone into reading the post.