Analyzing the Quality of Counseling Conversations: the Tell-Tale Signs of High-quality Counseling

Analyzing the Quality of Counseling Conversations: the Tell-Tale Signs of High-quality Counseling Ver´onica P´erez-Rosas1, Xuetong Sun1, Christy Li1, Yuchen Wang1 Kenneth Resnicow2 and Rada Mihalcea1 1Computer Science and Engineering, 2 School of Public Health University of Michigan {vrncapr, xuetong, christyq, lucasyc,kresnic,mihalcea}@umich.edu Abstract Behavioral and mental health are pressing issues worldwide. Counseling is emerging as a core treatment for a variety of mental and behavioral health disorders. Seeking to improve the understanding of counseling practice, researchers have started to explore Natural Language Processing approaches to analyze the nature of counseling interactions by studying aspects such as mirroring, empathy, and reflective listening. A challenging aspect of this task is the lack of psychotherapy corpora. In this paper, we introduce a new dataset of high-quality and low-quality counseling conversations collected from public web sources. We present a detailed description of the dataset collection process, including preprocessing, transcription, and the annotation of two counseling micro-skills: reflective listening and questions. We show that the obtained dataset can be used to build text-based classifiers able to predict the overall quality of a counseling conversation and provide insights into the linguistic differences between low-quality and high-quality counseling. Keywords: conversation analysis, counseling, mental health 1. Introduction Mental and behavioral disorders, such as substance abuse, are top on the list of the most costly and prevalent condi- tions worldwide.1 Particularly in the US, a recent survey on public health reported that in 2014 3.3% of all adults had co-occurring mental illness and substance abuse disor- ders.2 As behavioral counseling has been been shown to be an ef- fective treatment method for these conditions, the number of people seeking counseling services is increasing (Chava, 2014). Despite its potential benefits, such as combating addiction and providing broader disease prevention and management, the mechanisms behind successful behavioral counseling have not been fully elucidated (Moyers et al., 2009). Specific counseling skills have shown to increase the like- lihood of positive health outcomes (Gaume et al., 2009; Vader et al., 2010). Regardless of the counseling method, counselors follow general principles, such as supporting au- tonomy, expressing empathy, centering on the patient and engaging patients using specific skills such as reflective lis- tening (Charles et al., 1997; Harting et al., 2004). In con- trast, using a more directing style – characterized by coun- selors providing instruction and advice, and patients obey- ing, adhering and complying (Miller and Rollnick, 2013) – is usually avoided. The guidelines described above can be used to differen- tiate between low and high quality counseling. Thus, in a broad classification, psychotherapy conversations where counselors follow preferred practices can be considered as high-quality (or guideline adhering) counseling, whereas 1Word Health Report 2001, whr/2001/media_centre/en/http://www.who.int/ 2The State of Mental Health in America https://www. samhsa.gov/disorders/substance-use those conversations where they do not can be regarded as low-quality counseling (or guideline non-adhering). Following this idea, our paper analyzes counseling conver- sations with the final goal of distinguishing between low and high quality counseling. In particular, we focus our analysis on counseling conducted using Motivational Inter- viewing (MI), …

Content truncated to 500 words