Source: WASHINGTON STATE UNIVERSITY submitted to NRP
SOCIAL INTERACTION AND CONSUMER ACCEPTANCE OF GENOME EDITING IN DOMESTIC LIVESTOCK
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
AFRI COMPETITIVE GRANT
Reporting Frequency
Annual
Accession No.
1022788
Grant No.
2020-67023-31637
Cumulative Award Amt.
$445,000.00
Proposal No.
2019-08480
Multistate No.
(N/A)
Project Start Date
Sep 1, 2020
Project End Date
Aug 31, 2023
Grant Year
2020
Program Code
[A1642]- AFRI Foundational - Social Implications of Emerging Technologies
Recipient Organization
WASHINGTON STATE UNIVERSITY
240 FRENCH ADMINISTRATION BLDG
PULLMAN,WA 99164-0001
Performing Department
(N/A)
Non Technical Summary
Opposition to foods made from new technology, especially genetic engineering, has been increasing over time. Recent advances in genome-editing are widely acknowledged to have tremendous potential for beneficial results, but research is needed to understand the accompanying potential scientific, ethical, regulatory, and social consequences. Since genome editing is a relatively new phenomenon, there is an opportunity to study how social interactions affect consumer acceptance of it. Consumers obtain information about new technology in many ways, including personal interactions, social media, and traditional media. The proposed project will examine how social and traditional media, in-person interactions with experts, and watching a video affect consumer acceptance of genome editing. We consider the case of genome-edited hornless livestock that focuses on animal welfare. We will use text-mining algorithms on major social and traditional news media websites to understand their impact on public sentiment about genome editing and animal welfare. We will host public-engagement events and interviews in California and Idaho to facilitate communication between scientists, producers, an ethicist, and the public about the use of genome editing for hornless cattle. We will test whether the intervention leads to improved understanding of how genome editing for animal welfare decreases the polarization by administering pre- and post-event surveys. Finally, a one-minute video will be created for YouTube from footage of the public engagement events. Subjects can watch the video and complete an online survey, and the results of the in-person and online intervention can be compared.
Animal Health Component
80%
Research Effort Categories
Basic
20%
Applied
80%
Developmental
0%
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
6073310301034%
8033310303033%
3043310108033%
Knowledge Area
803 - Sociological and Technological Change Affecting Individuals, Families, and Communities; 607 - Consumer Economics; 304 - Animal Genome;

Subject Of Investigation
3310 - Beef cattle, live animal;

Field Of Science
1080 - Genetics; 3010 - Economics; 3030 - Information and communication;
Goals / Objectives
The overarching goal of the proposed project is to examine and understand how public perceptions of genome editing in domestic livestock are influenced; using social and traditional media data obtained via web scraping, public engagement events, with interviews and surveys. We will evaluate the broad social, ethical, cultural, legal, and other potential impacts of genome editing perceived by U.S. consumers. We will investigate how information acquired through different types of social interaction affects consumer acceptance of genome editing.The supporting objectives of the proposed project are:Analyze impacts of social media and traditional media. We will collect and analyze data via web scraping on social media and traditional media. We will search for the most frequently appearing words with phrases similar to "genome editing" and measure the public sentiment (positive versus negative words) regarding genome editing in domestic livestock. As part of this objective, we will identify social media influencers on genome editing and examine their impacts.Analyze impacts of public-engagement events with experts. We will organize and host public-engagement events and interviews in California and Idaho to facilitate communication between experts (see the appendix for the composition of the expert panels) and the public about the use of genome editing for hornless cattle, featuring the polled example as genetic approach to address an animal welfare problem. We will evaluate the impacts of the panel on audience members' perceptions of livestock genome editing through data collection from pre- and post-test surveys of the attendees. The events will be videotaped. A subset of the public event attendees will be interviewed, with semi-structured interview approach that allows variability in results depending on subject experience and knowledge (Crabtree and Miller 1992). Data will be used to understand attitudes toward livestock genome editing and identify connections between attitudes and motivation/rationale in food choices. Qualitative data is important because public acceptance of genome edited products does not follow a one-size fits-all approach (Friedrichs et al. 2019, 443). Analyze impacts of online video and compare across interaction types. We will create one-minute video from footage from the public engagement events. Subjects will watch the video and complete an online survey, and the results of the in-person and online intervention can be compared. We will test whether the intervention leads to improved understanding of how genome editing for animal welfare decreases the polarization by administering pre- and post-event surveys.
Project Methods
We investigate the social interaction in two ways: (1) monitoring online interactions via social networking sites and (2) collecting data from public engagement events, surveys, and interviews.We will analyze the contents of social media networks and responses. We use a cloud-based social media and networks analyzer Netlytic that generates data and summarizes in visual (e.g., network tree) and textual form (e.g., data frame with numeric and string values). In particular, Netlytic uses the social network analysis (SNA) model to visualize how individuals interact with one another through online conversations. The networking process can be divided into two parts. First, "name networks" that refers to "who mentions whom," which is identified by personal names mentioned in the body of each message, and then connecting the sender to everyone mentioned in his or her message. Second, a "chain network" is built based on information about direct interactions among online participants such as direct replies. Direct interaction thus includes message begins with a username for Twitter, and the names in the replies on Facebook and Instagram. We first specify a set of words from the literature that express either positive or negative views about agricultural technology.Then we perform sentiment analyses for each of the words.The retrieved sentiments willbe used to find patterns. We will complement Netlytic with Social Mention, which provides the frequency of sentiments and includes over 80 sites, such as Facebook, YouTube, Google+. When a search term is entered, Social Mention displays keywords, popular discussions, news stories, sentiment, media type, platform, and top users. It also provides a numerical ratio score of posts that are associated with positive to negative emotions. The sentiments can be segregated by regions (Tumasjan et al. 2010). We will use a machine learning algorithm (text mining) to conduct the sentiment analysis over a period (e.g. two weeks) and record the words mentioned with agricultural technology. The algorithm will also monitor news headlines on Google, and construct a time series of targeted words' appearance. This will facilitate comparison between traditional media and social media in mentioning of a targeted word. The second step is to identify the vocal (top users) and influential member accounts who use the words. We will categorizes them as either positive or negative, based on their posts.We will collect the set of most discussed words regarding genome editing, and the shares of technophobes and amenables. Checking the associated words with top liked posts will give the social, ethical, cultural, legal, and economic aspects of consumer perceptions. To classify a mentioned word under a category, say ethical, we use a large set of words. The set of feasible words is also extracted from the internet that are closely related to the word "ethical."We willconductpublic engagement events that will allow two-way communication between experts and members of the public. We will test whether public engagement using this animal welfare trait leads to a better understanding of how genome editing might be used to solve problems in food animal breeding programs and decrease the polarization that may potentially be associated with its use by administering pre- and post-event surveys that include questions. Each event will include a brief introduction to the topic with part of the event being devoted to discussions and two-way interactions between the panelists and the audience. This will allow for interaction with STEM professionals from the research team and enabling direction of questions directly to scientists. At the end of the event, a post survey will be taken using clickers to test for changes in knowledge and perceptions.To understand attitudes toward livestock genome editing and identify connections with motivation and rationale in food choices, we will collect qualitative data using semi-structured interviews. Ten to twelve subjects will be recruited at each public engagement event using volunteer forms that collect basic demographics for a diverse group. Questions will address knowledge and perception of biotechnology, genome editing, and U.S. animal welfare and biotechnology regulation; concern and interest in animal welfare, including any activities; concern and interest in policy approaches to livestock treatment; and motivations and rationale for eating practices.A one-minute video will be created from footage of the public engagement events. We will post the video to conduct an online survey to measure the impact of social interaction on broader audience. At least 100 users with demographics of nationally representative adult population will be randomly chosen for a benchmark survey, and will be shown the video followed by another round of questions. Unlike direct survey participants, Subjects can watch the video and complete an online survey, and the results of the in-person and online intervention can be compared.

Progress 09/01/20 to 08/31/23

Outputs
Target Audience:Our target audiences are scientists and other researchers, agricultural industry participants along the entire food supply chain, and policy makers. Changes/Problems:Because of the Covid pandemic, we could only hold one in-person public engagement event. Other events were held online. What opportunities for training and professional development has the project provided?Several graduate students were trained on this grant. Joseph Navelski was trained as a PhD student in Economics at Washington State University. He defended his dissertation in May 2023.Srijan Budhathoki, an MS student in Agricultural Economics at the University of Idaho, was trained in data analysis.Cameron Kester, an MFA Student at Washington State University, created graphics for the surveys.Xueying Ma, a PhD student at Washington State University, was trained. She defended her Ph.D. in October 2023. How have the results been disseminated to communities of interest?Thus far, the results have been presented to the academic community. Findings were presentedat an international conference and at a dissertation defence. Several journal articles are being revised and are under peer review. Next steps are to report to policy makers and industry participants. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? In terms of the overarching goal, as a result of this grant project, we better understand how public perceptions of genome editing in domestic livestock are influenced. We obtained data through webscrapping, social media academic interfaces, public engagement events, interviews, and surveys. In terms of the specific supporting objective #1, we used Twitter's Academic Research's application programming interface (API) to identify social media influencers and their impacts on the sentiment of social media users.We develop a model to infer the ideologies and influence of agents using social media. To do this, our model uses only connections and network structure, without requiring detailed socio-demographic characteristics of users. The model generates a score of ideological position for experts and their relative influence. Estimates show that followers opposed to genome editing have more influence than those in favor, e.g., anti-GEL followers own 69% of the social influence in a typical conversation. Our empirical results can help inform investment, marketing, and policy-making decisions within the livestock industry. Also under supporting objective #1, we utilized Twitter text data and machine learning to analyze U.S. user sentiment toward genome editing in livestock. Analyzing 384,473 tweets from January 2010 to January 2019, we found a mostly neutral but slightly negative sentiment. Terms such as"animal welfare," "organic," and "biotechnology" are associated with positive sentiments, while "dehorning" and "genetically modified" are viewed negatively. The sentiment varies by state, with South Dakota, Mississippi, and Kentucky showing the highest positive sentiment, contrasted with Maine, Tennessee, Wyoming, and California. A bootstrap analysis reveals significant differences between specific search terms, such as "dehorning" and "genome editing," 42.3% of the time. The results imply a near neutral sentiment towards genome editing in livestock, differing perceptions of associated words, and varying state-level perceptions. Data was also collected from traditional media with webscraping and text analysis to understand sentiment.Policymakers can use this information to understand how genome-edited livestock is viewed across the United States. With respect to supporting objective #2, we held public-engagement events and interviews in California, Washington, and Idaho. We evaluated the survey data and interviews. From the survey data, most participants possessed limited knowledge about genome editing. The impact of the public engagement events was positive on consumers' sentiments towards genome editing in livestock. The participants were also interviewed to obtain qualitative data. The data was analyzed and compared. With respect to supporting objective #3, we created a video. We conducted an online survey, where one quarter of the respondents received no additional information, one quarter watched the video, one quarter were presented with actual social media posts and one quarter read anexerpt from an actual newspaper article.The information treatments have differential effects on choices, with the science video having the largest positive impact. The newspaper article, which discussed potential risks, had the largest negative effect.

Publications

  • Type: Conference Papers and Presentations Status: Published Year Published: 2023 Citation: McCluskey, J.J., 2023. Gene-Edited Cows and Animal Welfare, Simon Brand Memorial Address presented at the African Association of Agricultural Economics and the Agricultural Economics Association of South Africa Conference in Durban, South Africa, September 20.
  • Type: Theses/Dissertations Status: Published Year Published: 2023 Citation: Navelski, Joseph. APPLIED MICROECONOMIC AND STATISTICAL METHODS USING SOCIAL MEDIA DATA, PhD dissertation, Washington State University.
  • Type: Journal Articles Status: Under Review Year Published: 2023 Citation: McCluskey, J.J., X. Ma, R.K. Gallardo, and R. Yang, Is Ignorance Bliss? Milk from Gene-Edited Cows and Animal Welfare Considerations."
  • Type: Journal Articles Status: Under Review Year Published: 2023 Citation: Navelski, J., S. Badruddoza, J.J. McCluskey. "The Expert Effect On Network Formation: An Application To Genome Editing Opinions On Twitter."
  • Type: Journal Articles Status: Under Review Year Published: 2023 Citation: Navelski, J., S. Badruddoza, J.J. McCluskey. Gaining Inference in a Machine Learning Natural Language Processing Sentiment Analysis: Genome Editing in Domestic Livestock Using Twitter Data.


Progress 09/01/21 to 08/31/22

Outputs
Target Audience:We reached consumers and researchers during this reporting period. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Our PhD student in Economics at Washington State Universitywill graduate this year. We also worked with a MS student at the University of Idaho received training, and we worked with a post-doctoral student in animal science. How have the results been disseminated to communities of interest?We have conducted an in-person public engagement event in Pullman, WA. Our video on genome editing, which is accessable to the public has been posted online. What do you plan to do during the next reporting period to accomplish the goals?This year we will conduct our online survey. We willanalyze the data from traditional media and the online surveys. We will publish papers, a dissertation, and make presentations.

Impacts
What was accomplished under these goals? On goal 1:PhD candidateJoe Navelski completed the analysis of Twitter accounts. Heused scraped data from both top pro- and anti-GMO to estimate sentiments both for and against genome editing. He developed a new method to estimate the social network in a learning model. Agents build a network by comparing the experts they follow. He used a latent variable spatial following model to explain why agents follow these experts. The model is used to estimate the underlying individual parameters that explain following decisions in a network, including the followers' and experts' ideological positions. We use these estimates to derive each agent's level of influence in the social network. We applied this method to the top (in terms of influential followers) experts in the field of genome editing in domestic livestock (GEDL) on Twitter. We find that the anti-GEDL followers own 69% of the social influence in any conversation. Implications are that any conversations about GEDL on Twitter will be heavily influenced by anti-GEDL followers, making it difficult for pro GEDL opinions to be accepted. In terms of traditional media, we collected data on GEDL from National Public Radio and several newspapers (including the New York Times and Wall Street Journal) from Lexus Nexus. On goal 2:Our philospher Patricia Glazebrook conducted follow-up interviews after the zoom public engagement events. We held an in-person public engagement event and conducted pre- and post-surveys. We created a video based on the public engagement events. On goal 3:We created the video and completed a draft of the online survey.

Publications

  • Type: Theses/Dissertations Status: Under Review Year Published: 2023 Citation: Navelski, Joseph, Applied Microeconomic and Statistical Methods using Social Media Data, PhD dissertation, School of Economic Sciences, Washington State University.


Progress 09/01/20 to 08/31/21

Outputs
Target Audience:Our target audiences included scientists, industry participants, and consumers. Changes/Problems:The COVID-19 pandemic made our in-person events impossible. As a substitute, we conducted live virtual public engagement events. What opportunities for training and professional development has the project provided?A doctoral student and a master's student are being trained with this project. An undergraduate female student also has started to work on the project. How have the results been disseminated to communities of interest?We have interacted with industry and researchers during the public events. So far, we have only made one presentation of our results to other researchers. We will more to report on this in the next report. What do you plan to do during the next reporting period to accomplish the goals?Now that the pandemic is winding down, we have an in-person public engagement event planned. We also expect to analyze all the data we collected. We will also complete the video and all objectives in the grant.

Impacts
What was accomplished under these goals? Under the goal of analyzing the impacts of social and traditional media, we have completed an analysis of social media and working on the traditional media part. On the second goal, we held two public-engagement events with experts. Unfortunately due to the pandemic, they had to be held virtually rather than in person. Still, we held the events with one hosted at UC Davis and the second hosted at the University of Idaho. We collected data at both events and we are currently analyzing the data. In addition, the qualitative data has also been collected. In the third goal, We are currently working to create a video, so that we can compare interactions across multiple formats.

Publications

  • Type: Conference Papers and Presentations Status: Published Year Published: 2021 Citation: Navelski, J. "Gaining Inference in a Machine Learning NLP Sentiment Analysis: An Application to the Topic of Genome Editing in Domestic Livestock using Twitter Data," presented at the School of Economic Sciences Seminar Series, Washington State University.