Dear Marko,

On behalf of myself and my collaborators, I am writing to inquire if the attached manuscript is of interest to Communication Methods and Measures.  With the continuing rise of attention to machine learning and the rapid adoption of related methods in Communication Science, we feel a rigorous understanding of the statistical problems with misclassification and error correction methods as a potential solution is urgently needed in our field. 

We aim to accomplish this in an accessible way via a real-data example using Google's perspective API, a systematic literature review, and Monte-Carlo simulations evaluating our proposed approach for error correction in comparison to others. As we could find no open-source software implementing the most effective solution we tested, we are developing an R package to make it possible for our field to routinely correct for misclassification bias in automated content analysis. 

Our study reveals that misclassification bias can easily distort statistical analysis into giving misleading results, that this threat to validity is not normally recognized in communication research applying supervised text classification, but that, perhaps surprisingly, it is possible to correct bias with data from manual content analysis. Although the statistical literature on the topic is well-developed, it is also quite technical, as I am sure you are aware.  As a result, our contribution in this article is targeted to raise social scientists' awareness and comprehension of the problem and how it can be solved.

Based on your recent work on the issue of measurement error in content analysis, you are clearly an ideal associate editor for this article. In short, we are wondering whether you think our article will be a good fit for CMM. If so, we will submit it through the system in the next days.

Sincerely, 
Nathan TeBlunthuis