Using Information Cascades to Understand How I Form an Opinion on a Movie
Source: https://pubsonline.informs.org/doi/abs/10.1287/mnsc.2014.2082
My friends and I are avid Marvel fans and always make plans to go to see the newest movie together. After walking out of the theater, we gawk over the intense explosions and plot twists that are common to these movies. More often than not, the friend group is in agreement on their approval for the movie. Rarely, a movie will be so atrocious that we all agree upon our disappointment. Even more rarely, we will disagree on our opinions on the movie and argue over our general impression of the movie. While we may disagree on specific plot points, we rarely disagree on whether we like or dislike the movie.
In the paper Do I Follow My Friends or the Crowd? Information Cascades in Online Movie Ratings published in Management Science, the researchers attempt to understand how prior movie ratings by the crowd and friends alike may differently affect a subsequent user’s rating. The approach of the researchers to examine the process of online user rating generation relies on informational cascades theory, which describes situations where people observe others’ actions and make a choice that is often independent of their own private signals. To address the research questions, the study uses data from software agents that collect data from several public websites for all movies released in theaters in the United States in 2007.
In the paper, the researchers outline how an online user and movie watcher form an opinion about a movie. Users make inferences from a variety of sources of information that may be available at a given time. These sources are the high and low signals in the Cascade Model. In real life, these sources include advertisements, box office numbers, other user reviews, and eventually their own experience watching the movie. After taking in this information and their signals, the user forms an opinion and decides to rate the movie. While the rating should be primarily influenced by the user’s experience from watching the movie, this is not always the case. As we saw in class, if two users in a row get the same outcome, my personal signal is irrelevant, and an information cascade begins. Information cascades where large numbers of online users feel the same way about a movie creates a herding effect where users form the same opinion. In these situations, users will often make their ratings independent of their own experience and rely on the opinions of others that came before them.
Through their empirical strategy, the researchers find that a user’s rating method tends to be different for different types of movies. For niche movies, users largely rely on their own experience. For popular movies, users often herd with the crowd. Regardless of movie type, users trust their friend’s opinions more closely than other online users. As friend interaction increases, users are less likely to herd with online crowds. To avoid information cascade and hering behavior, the researchers suggest that social media sites take steps to avoid bias and get honest opinions. Going forward, the next movie I watch with my friends, I will be more cognizant to establish my own opinion and express it quickly.