Over the last century, nonnative earthworms have invaded forests of the Great Lakes region of North America. Although a growing body of scientific research has documented short-term changes associated with invasive earthworms, there is little research describing the effects of invasive earthworms over multiple decades. To investigate the long-term effects of invasive earthworms on forests, sites sampled in the past need to be classified as wormed or unwormed when originally sampled. However, this is often difficult to accomplish because field methods for sampling earthworms have only recently been developed, and the few historical permanent sites available for resampling largely do not have past information about earthworm presence or absence. Although historic sites lack data on invasive earthworm presence, many of these sites do have information about soil horizon thickness. Therefore, soil horizons can potentially be used as an indicator of the presence or absence of invasive earthworms. In this paper we developed a logistic regression generalized linear model to classify 40 sugar maple-basswood sites in the Chippewa National Forest of Northern Minnesota as wormed or unwormed (i.e., presence or relative absence of earthworms, respectively). A model using the thickness of the O horizon as a predictor variable correctly classified 93% of sites resampled in 2017 as wormed or unwormed. This result suggests we can predict which sugar maple-basswood stands in the Chippewa National Forest were wormed in the past. By comparing historic conditions to those present today, we can then analyze the long-term effects of invasive earthworms.