In this study we performed an initial investigation and evaluation of altmetrics and their relationship with public policy citation of research papers. We examined methods for using altmetrics and other data to predict whether a research paper is cited in public policy and applied receiver operating characteristic curve on various feature groups in order to evaluate their potential usefulness. From the methods we tested, classifying based on tweet count provided the best results, achieving an area under the ROC curve of 0.91. KEYWORDSAltmetrics, Social Media, Public Policy BACKGROUND 1.IntroductionThe growth of social media in the academic community has enabled scholars to develop new methods to evaluate the impact of research. Historically, evaluations of the impact of research have been limited to the reception by the scholarly community. However, with the advent of altmetrics we are able to track the social impact of research [8][2] . For example, Ding et al. [3] explored the use of social media tagging as it relates to scholarly works.Policy documents have a vital role in generating demand for scientific innovation [4]. Haunschild and Bornmann [5] study the relation between Web of Science fields and the researches' use in public policy and found that less than 2% of every category is cited in public policy. Thelwall, and Kousha [6] explored the relationship between citations in patents and technological impact and found that the number of patents citing a resource indicates the technological capacity or relevance of that resource. Winterfeldt [9] presented a framework to bridge the gap between science and decision making in the policy sphere.To better understand the possibilities of altmetrics, we conducted an exploratory study to determine the potential of social media JCDL2017, June 2017, Toronto CA 2017. This is a preprint of an extended abstract to be presented in JCDL '17. data for creating valuable models to describe the use of research articles in public policy. CollectionThe primary source of data for our analysis was a database dump from altmetric.com [1]. The dataset, which is from June 4th 2016, consists of 5.2 million articles. We separated articles into 2 classes: papers cited in public policy documents and papers not cited in such documents. Policy documents, as used in this paper and by altmetric.com, currently includes mostly policy published by medical organizations. The dataset comprised 89,350 research articles referenced in policy documents and 5,097,207 articles not referenced in that context.We drew on the altmetric.com dataset for meta-info about each research paper, specifically journal, publisher, and Scopus subject information, as well as social media activity. To augment our dataset, we collected the citation counts for articles of interest. Initially, we collected citation counts for all the policy documents and 120,000 of the non-policy documents from the Thomson Reuters Web of Science. In addition, we collected journal impact factors for 9,000 journals. Feature Select...