Language Style Matching as a Measure of Librarian/Patron Engagement in Email Reference Transactions


Published: January 23, 2020 by Ann Agee 

When two people are deep in conversation, they unconsciously mimic each other. Both will cross their arms, pick up their coffee cups, or touch their hair as they talk. Research shows that this mimicry is a signal of the high level of engagement between the conversationalists (Scheflen, 1964). Using a technique called language style matching (LSM), social psychologists discovered that similar synchronization appears in written correspondence (Niederhoffer & Pennebaker, 2002). Correspondents with a high level of engagement use similar words, down to the level of pronouns, articles, and prepositions. It is these function words that are used to calculate an LSM score. High LSM scores have been shown to be indicative of a sense of perceived support.

For my recent sabbatical project, my goal was to see if LSM could potentially be used as a metric for determining how well librarians succeed in engaging with and creating a supportive environment for their virtual patrons, the crucial first step in any successful reference transaction (Agee, 2019). To do this, I applied LSM to 1,139 anonymized email threads gathered between myself in my role as liaison librarian to the School of Information and the School’s graduate students. For comparison, I also applied LSM to 111 email threads from the Dr. Martin L. King, Jr. Library’s general email reference service, LibAnswers.

To measure the level of LSM, I used Linguistic Inquiry and Word Count (LIWC) software to count and classify the function words in the emails. Because LSM measures engagement, both caring conversations and angry fights can generate a high LSM score. To make sure the email conversations analyzed did not fall into the “angry” category, I also used LIWC to calculate the number of negative emotion words used in the emails. As a secondary measure, I measured a LIWC-generated variable called emotional tone, which provides a metric for the upbeat or downbeat tenor of a piece of writing. These two measures together demonstrated that the high-engagement exchanges analyzed tended toward a neutral style, as might be expected in business correspondence.

After analyzing the two data sets, the results showed that email reference provided through the LibAnswers general reference service demonstrated a moderate to high level of LSM 67% of the time and email reference provided by the liaison librarian demonstrated a moderate to high level of LSM 84% of the time. One interpretation of these findings is that the expertise of liaison librarians lends itself to an increased ability to create a supportive online environment for their students, a level of specialization not possible in a general reference service. However, both data sets demonstrated that librarians were successful in creating a supportive environment for patrons the majority of the time. Additional research needs to be done, but this exploratory study shows the potential of LSM in providing a more holistic assessment of virtual reference services, giving libraries the ability to measure how well librarians are meeting their patrons’ affective as well as informational needs.


Agee, A. (2019). Language style matching as a measure of librarian/patron engagement in email reference transactions. Journal of Academic Librarianship, 45(6). doi:10.1016/j.acalib.2019.102069

Niederhoffer, K. G., & Pennebaker, J. W. (2002). Linguistic style matching in social interaction.Journal of Language and Social Psychology, 21(4), 337–360.

Pennebaker, J. W., Boyd, R. L., Jordan, K., & Blackburn, K. (2015). The development and psychometric properties of LIWC2015. Austin, TX: University of Texas at Austin.

Scheflen, A.E. (1964). The significance of posture in communication systems. Psychiatry, 27(4), 316–331.


Post new comment