Preventing the Spread of False Information
Information, regardless of its accuracy, spreads rapidly through social media, reaching and influencing millions of readers. In special instances, stories achieve viral status, where a large number of people receive the material within days, if not hours. Unfortunately, oftentimes information is incorrect, yet people accept it as true.
Chris Lepre ’15 and Rachel Friedman ’15 are developing a program that will analyze the spread of misinformation in our technologically-driven world. They will produce a Facebook program that accesses people’s post history, and create an algorithm to perform sentiment analysis involving three viral Facebook events.
Sentiment analysis measures the frequency of target words, and related terms, to summarize opinions on an article, blog post or any other writing. Lepre explained that movie ratings are a classic example of this, where a computer can interpret an author’s reaction to a movie based on their word choice.
Once the team has obtained posts from participants, they will apply the algorithm to the writings. Positive and negative sentiments will be attached to keywords in a model document to train the algorithm, which will then be identified in the Facebook posts. Lepre and Friedman, now registered Facebook developers, started designing this data-collecting program, and they will see which users redistributed false information.
Stuart Hirshfield, the Stephen Harper Kirner Professor of Computer Science, helped the students select two measures to observe those who spread misinformation. The first is the “need for cognition,” or “willingness of an individual to engage in cognitive processes,” according to Lepre. Essentially, it’s the ability to use one’s background knowledge and double-check sources. The second measure is the propensity to trust what others have previously posted.
Friedman admitted the project presents challenges regularly. They have made multiple breakthroughs, but much of the coding has been a process of trial and error. This research is giving them crucial experience in detecting and troubleshooting errors.
Both students are experienced with Python and other programming languages, but unfamiliar with web development. This opportunity will allow them to strengthen and hone their skills toward a goal that they constructed. Lepre has found “web development intimidating, but also interesting,” and is now considering it as a career path.
While categorized under the Computer Science Department, this research could be applicable to psychology, communications, anthropology and other subjects. Advertising agencies are able to utilize this analysis and cater marketing to individuals, depending on their posting history. Lepre hopes their work goes toward “more fruitful efforts to make unreliable information less prevalent.”
Lepre is a graduate of The Wheatley School (N.Y.). Friedman is a graduate of Saratoga Springs High School (N.Y.).