Source: CORNELL UNIVERSITY submitted to NRP
LINGUISTIC BIAS IN COMMUNICATION ABOUT PUBLIC CONTROVERSIES: EFFECTS OF COMMUNICATION MEDIUM
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
Reporting Frequency
Annual
Accession No.
1007222
Grant No.
(N/A)
Cumulative Award Amt.
(N/A)
Proposal No.
(N/A)
Multistate No.
(N/A)
Project Start Date
Oct 1, 2015
Project End Date
Sep 30, 2019
Grant Year
(N/A)
Program Code
[(N/A)]- (N/A)
Recipient Organization
CORNELL UNIVERSITY
(N/A)
ITHACA,NY 14853
Performing Department
Communication
Non Technical Summary
An unfortunately common result of public controversy is polarization into ingroups versus outgroups. Previous work by the PI supported by federal capacity funds showed how intergroup polarization results in linguistic bias, and several workshops on how to recognize such bias, in the specific context of the controversy over hydraulic fracturing in the Marcellus Shale, were delivered to CCE educators and the Cornell Municipal Clerks Institute. The proposed research extends the focus of that work from recognizing linguistic bias to understanding its effects on parties directly involved in controversies. Coinciding with the aims of our CaRDI collaborators, we are particularly interested in examining how the effects of biased language by intermediaries impacts the outcomes of negotiations made between community members. For example, theory from the psychology and conflict resolution suggests that the presence of biased communication may reduce the likelihood that the parties would be willing or able to find productive solutions. A second focus of the proposed research is the role of communication media on intergroup dynamics surrounding public controversies, in light of the dramatic expansion from traditional face-to-face town meetings to internationally networked social media. Today Internet Enabled Communication (IEC) allows anyone to choose a side in a controversy, no matter their degree of direct involvement. It is important to understand how features of IEC (e.g., publicness; anonymity), and under what circumstances IEC affect intergroup communication. Furthermore, our final aim is to expand our understanding of how language impacts negotiations in intergroup situations. Recent research links abstract thinking and language to integrative negotiation processes, yet this connection has not been examined within an intergroup context, such as the one developed around a public controversy. Once these various research questions have been empirically explored, we plan to create training materials and exercises for CCE instructors to use with community intermediaries. The proposed research will examine: a) how biased intergroup communication affects the target of said bias b) how IEC impacts and affects reduction of biased intergroup communication; c) how language impacts negotiation and deliberation processes; d) strategies to reduce intergroup polarization through reduction in linguistic bias.
Animal Health Component
40%
Research Effort Categories
Basic
60%
Applied
40%
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
90360993070100%
Goals / Objectives
The research centers on the linguistic intergroup bias (LIB), in which abstract words are systematically paired with positive ingroup and negative outgroup behaviors while concrete words are paired with negative ingroup and positive outgroup behaviors. This bias reinforces stereotypical perceptions of both ingroups and outgroups and is particularly insidious because of its subtlety. Research shows the production of LIB is widespread, but little is known of its effects on targets. Work I began with previous federal capacity funds suggests, for example, that LIB has negative effects on self-esteem. The proposed research will further examine LIB effects on: a) willingness to engage in problem-solving dialogue and; b) likelihood of reaching integrative (i.e., win-win) solutions to contested issues. Further, this work will examine how these effects may differ across different communication media, how language impacts negotiation outcomes in intergroup contexts, and how controlling LIB may impact perceptions of outgroup members.
Project Methods
Paradigm 1 - Linguistic Bias and Perceptions of IntermediariesAs we have argued, there is practically no direct data on how linguistic bias affects targets (Beukeboom, 2014). Therefore, the objectives of the studies within this paradigm are to examine the effects of biased language of intermediary agents within public controversies on target audiences.We propose a series of experimental studies to examine various aspects of this basic question, using the following basic design:Stimulus materials. The stimuli will be based on transcripts excerpted from community meetings over various public controversies. We seek issues where people take clear sides on the issue, and we will work with our CaRDI colleagues to identify examples that we can use as models to develop our materials (e.g., shale gas development using hydraulic fracturing; siting of wind energy turbines; decisions involving permitting casinos.). We will craft a transcript centered on a heated point during a meeting wherein a participant on one side of the controversy makes a very negative offensive comment about adherents to the other side of the controversy, requiring an intervention of the meeting moderator.Independent variables. The moderator intervention will contain our main independent variable. We will manipulate the comments to convey linguistic bias against one group or the other, plus a no-bias control condition. Additional independent variables will be included in follow-up studies (i.e., whether meeting is face-to-face vs. online; the specific controversial issue; the intermediary's role as moderator, vs. arbitrator).Dependent variables. Our measures would include perceptions of the moderator's group membership (i.e., member of ingroup or outgroup or neutral) amount and direction of moderator's bias, moderator's credibility and effectiveness, the quality of the intervention the moderator offers, and willingness to have this individual moderate future controversial meetings.Samples. We will define different "publics" related to the various controversies we study We will develop additional on-campus and community-based issues partially with the help of our CaRDI collaborators. Examples of topics we have already explored with them include decisions related to issues over land use and energy, policies related to climate change, strategies to control invasive species, casino gambling, and police militarization.Paradigm 2 - Construal Level Theory and NegotiationsWe propose a 2 (high vs. low construal) by 2 (high vs. low social categorization salience) experimental study in which people will be engaged in a negotiation task. We will measure quality of the negotiation process and solution.Task. We will use a template from a standard negotiation task, or from previous studies to develop a task based on an issue that we identify in conjunction with our CaRDI collaborators, such as the debate over hydraulic fracturing. The task will be developed to allow us to score the solution quality on a continuum from integrative to distributive.Independent Variables. We will use linguistic abstraction to manipulate construal level. The manipulation will be embedded in narrative descriptions of the partners in the upcoming negotiation. The social categorization salience manipulation will also be embedded in the narratives. In the high conditions the differences in the negotiation partner's position on the issue will be described in categorical terms and with respect to the person belonging to a group of like-minded people. No mention of group membership will be made in the low categorization condition.Dependent Variables. We will measure the extent of participants' stereotyping of their partners participants' perceptions of the negotiation process (fairness, satisfactions, etc.), and the quality of the negotiation solution.Samples and Settings. We will conduct the study on a laboratory student sample and replicate it with a community-based sample.Paradigm 3 - Controlling Linguistic BiasSpecific communication goals can affect people's likelihood and ability to reduce linguistic bias. Our goal is to examine this effect in light of communication medium, specific type of controversy, type of interaction and target reactions. Our final goal is to use what we learn from these investigations to develop specific training materials, targeted for public controversy intermediaries, on how subtle forms of linguistic bias can be identified and controlled.Experimental studies within this paradigm will follow one of two basic designs. The first will attempt to replicate Douglas, et al., 2008, and focus on people's ability to control linguistic bias:Stimulus materials. We will use transcripts from heated segments of meetings about public controversies. For these studies, however, rather than supplying an intermediary response we will ask participants to respond to an offensive comment from member of one side toward the other side.Independent Variables. The main independent variables will be the group identify of the speaker making the offensive comment, thus participant will be faced with an offensive comment made by someone on the same side of the issue as themselves (i.e., ingroup) or from someone on the other side of the controversy (i.e., the outgroup), and instructions regarding the control of bias (i.e., produce unbiased response; produce pro-ingroup and anti-outgroup biased response; no instruction control). Additional independent variables to be considered in various studies include communication medium (face-to-face vs. online), the specific controversy, and the nature of the meeting (decision-making vs. information dissemination).Dependent Variables. The main dependent variables will be the degree of linguistic bias seen in the comments produced by respondents under each condition.The second design will focus on response to attempts at controlling biased communication.Stimulus materials/Procedure. Participants will be asked to consider that they are going to select a person to moderate a meeting about a controversial topic. They will be given examples of transcripts of previous meetings moderated by candidates whose content will contain the manipulation of one of our main independent variables (described below).Independent variables. The manipulation of bias in the communication from the targeted individuals will be as in previous studies, (i.e., pro- one or the other of the sides in the controversy, and unbiased version). The other variable within this study will be a manipulation of participants' expectations of how well the candidates have learned to control bias. For example, one possible manipulation would be to tell participants that one candidate is a student in a conflict resolution training program, whereas another candidate will be described as a professional will a lot of experience in controlling bias. Thus, the basic experimental design will involve expectations of high and low bias crossed with communications that actually contain bias or not. Additional variables will include the specific controversy and the nature of the interaction (i.e., information dissemination vs. decision-making).Dependent variables. The measures will include ratings of the target person's credibility, amount of bias, and likelihood of choosing the targeted person for the moderator position.In both of these designs we will conduct studies across different samples, as described in previous studies.

Progress 10/01/15 to 09/30/19

Outputs
Target Audience:Academic audiences through conference presentations. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?A graduate student had the opportunity to supervise undergraduate research assistants and to present results of one study at the International Communication Association Conference. Undergraduate students learned advanced data analyses techniques. Three undergraduate students presented results of research at a professional conference (National Communication Association); they won the best student paper award in the group communication division. How have the results been disseminated to communities of interest?Through presentations at the following conferences: Interdisciplinary Network of Group Research, International Communication Association; National Communication Association. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Studies completed: 1. We conducted a study of people's responses to biased communication about candidates in the 2016 U.S. Presidential election, in which we did show that systematic and subtle bias in word choice to describe candidates had the predicted effects on people's emotional responses. The results of this study were presented at the meetings of the International Communication Association - the leading academic conference in the discipline of Communication. 2. We completed a study of linguistic indicators of gender bias and stereotypes in negotiations. Results of this study were presented at 3 academic conferences (National Communication Association; International Communication Association; Interdisciplinary Network of Group Research) 3. Using the same methodology, I completed data collection of a study of people's responses to subtle bias in news stories on both sides of the abortion debate. We also found that the subtle linguistic bias affected people's responses to these news stories. These data are still under analysis. 4. We completed pilot testing of a methodology and stimuli to examine expression of linguistic bias in a teamwork setting. The stimuli tested will be used to examine the operation of linguistic bias during interactive collaboration. Materials developed: 1. Video expressions of distinct emotions that can be used in training workshops. 2. Written examples of linguistic bias in text that can be use in training workshops. 3. Templates for developing written stimuli for manipulating linguistic bias in text.

Publications

  • Type: Conference Papers and Presentations Status: Accepted Year Published: 2019 Citation: Wu, Y., Chan, C., Ding, C., McLeod, P. L. & Williams, M. (2019) Effects of avatar gender on negotiation in a virtual environment. Presentation at National Communication Annual Conference, Baltimore, MD. Winner of Best Student Paper Award, Group Communication Division


Progress 10/01/17 to 09/30/18

Outputs
Target Audience:Presentations at two conferences reached academic audiences interested in social identity effects in interactions. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The project provided training in experimental design and data analysis for 1 graduate student and 10 undergraduate students How have the results been disseminated to communities of interest?We have presented some of the work at academic conferences. What do you plan to do during the next reporting period to accomplish the goals?We will conduct at least two experimental studies, and will submit them for academic conferences; following those submissions, we will work toward journal submissions.

Impacts
What was accomplished under these goals? 1) We developed and pilot tested a new set of stimuli for an experiment on nonverbal and linguistic markers of hostile emotions. 2) We completed data collection on Study 1, of multiple studies on intragroup dynamics associated with the effects of negative emotion expression in group interaction; data collection for Study 2 is in progress as of this writing. We expect to report results of these studies at academic conferences in 2020.

Publications

  • Type: Conference Papers and Presentations Status: Accepted Year Published: 2018 Citation: Wu, Y., Chan, C., Ding, C., McLeod, P. L. & Williams, M. (2018) Effects of avatar gender on negotiation strategy and outcomes in an immersive virtual environment. Presentation at the International Communication Association Annual Conference, Prague Czech Republic.
  • Type: Conference Papers and Presentations Status: Accepted Year Published: 2018 Citation: McLeod, P. L. & Williams, M. (2018) Effects of sex-role stereotypes on a negotiation task in an immersive virtual environment. Presentation at the Interdisciplinary Network of Group Research Conference, Bethesda, MD
  • Type: Journal Articles Status: Published Year Published: 2018 Citation: Song, H., Schuldt, J. P., McLeod, P. L., Crain, R. L. & Dickinson, J. L. (2018) Spillover effects of group norm violations on impression formation in a green social network. Group Processes & Intergroup Relations http://journals.sagepub.com/eprint/kXEwdqKnpkrncgY5Dzxa/full


Progress 10/01/16 to 09/30/17

Outputs
Target Audience:We have shared results of current studies with academic audiences through several conference presentations. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?This project provided opportunities for development of professional academic skills for two graduate students and several undergraduate students. The graduate students conducted each one of the studies supported by NIFA funds, wrote a paper which they then presented at major professional conferences. Undergraduate students learned research skills associated with data analysis, and also learned to organized and present research findings, which enabled them to participate in a university-wide showcase of undergraduate research. How have the results been disseminated to communities of interest?We have presented these results thus far at two major professional conferences. What do you plan to do during the next reporting period to accomplish the goals?We plan to continue collecting data relative to the studies described above, and to use the emotion stimuli to conduct further studies examining people's responses to nonverbal cues associated with expressions of bias. We will continue to work with our CCE partners to develop programs for CCE educators based on the findings from these studies.

Impacts
What was accomplished under these goals? In one study (preliminary results presented at the 2017 National Communication Association) we found that people's political orientations predisposed them to seeing bias in news reports about their favored political candidate, even when no bias existed, and determined by objective measures. This finding is relevant to the goals of understanding factors that would affect people's willingness to engage in problem-solving dialogue with people who hold opposing viewpoints (i.e., outgroup members). A second study demonstrated some of the impacts on group decision-making outcomes when people are exposed to threatening situations; our evidence suggests that decision may become either more extreme or more cautious, depending on a group's starting point. This again is related to the goal of understanding how dialogue is affected when people are interacting with others who hold opposing, and possibly threatening viewpoints (i.e., outgroup members).

Publications

  • Type: Conference Papers and Presentations Status: Published Year Published: 2017 Citation: Kruzan, K & McLeod, P. L. (2017) Effects of existential anxiety on group polarization. Paper presented at the 2017 Meeting of the Interdisciplinary Network of Group Research, St. Louis, MO.
  • Type: Conference Papers and Presentations Status: Accepted Year Published: 2017 Citation: Kazerroni, K. & McLeod, P. L. Linguistic Bias effects in news accounts of the 2016 presidential election: Experimental Evidence. Paper presented at the 2017 Meeting of the National Communication Association, Dallas, TX.
  • Type: Conference Papers and Presentations Status: Accepted Year Published: 2017 Citation: Wu, Y., Brathwaite, K., Chan, C., & Ng. S. Intergroup and language bias in the 2016 U.S. presidential election. Presented at the Cornell University Undergraduate Research Board Spring Fair


Progress 10/01/15 to 09/30/16

Outputs
Target Audience:We have conducted a study with a national sample regarding perceptions of biased communication during the 2016 U.S. presidential election. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The project has provided training of a graduate student and undergraduate students in data collection and analysis. How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals?We arein the process of analyzing the data from the study of the election, and will submit a paper from this for conference presentation and journal publication by the end of the 2017 calendar year.

Impacts
What was accomplished under these goals? We completed a study of perceptions of bias in the 2016 Presidential election, using a national sample. We are conducting an ongoing study of effects of biased communication among college students.

Publications