September 19, 2022

School bullying, cyberbullying, and remote learning

Andrew Bacher-Hicks and Joshua Goodman discuss the effects of the pandemic on bullying.

Source: melis

The pandemic has taken a toll on the education system. School enrollment has decreased, teacher turnover has ticked up, and students have experienced substantial learning losses

But there is at least one silver lining, according to a paper in the American Economic Review: Insights

Authors Andrew Bacher-Hicks, Joshua Goodman, Jennifer Greif Green, and Melissa K. Holt found that school bullying and cyberbullying dropped by more than 30 percent as schools shifted to remote learning in the spring of 2020.

Their results are based on real-time tracking of internet searches, which they show contain useful information about actual bullying behavior.

Bacher-Hicks and Goodman recently spoke with Tyler Smith about their approach to studying bullying trends and the degree to which bullying originates in schools and then migrates online.

The edited highlights of that conversation are below, and the full interview can be heard using the podcast player.



Tyler Smith: What motivated you to look into what's happened to bullying over the pandemic?

Andrew Bacher-Hicks: The motivation for this focus is that around March of 2020, when schools closed for in-person instruction and shifted to remote and virtual instruction, we were hearing a lot of concern among parents and school leaders that spending more time online would result in higher levels of cyberbullying. And this is a reasonable concern because the pre-pandemic research that was conducted on this showed that there is a correlation between students spending more time online and increased risk of cyberbullying exposure. What we wanted to do is use data from Google Trends in real time to explore whether or not there were changes to both in-person bullying and cyberbullying as schools shifted from in-person learning to remote learning. 

Joshua Goodman: I'll also add that one thing that struck me at the start of the pandemic and the shift to remote learning was that for many of us who are parents and many of us who know children, it seemed like a very difficult shift. But then I remember reading at least a couple of news articles where parents would be interviewed about how their children were doing in remote learning. And many parents talked about the struggles. But once in a while you'd hear a story about a parent saying that, actually, my kid used to have a lot of anxiety about going to school. But they've blossomed in remote learning. They feel safer there. That struck me as an intriguing potential story about the fact that not all kids' experiences of school are identical and that there might be multiple stories going on in this massive shift in the mode of schooling that we had in the spring of 2020.

Smith: I think what's unique to your approach is this Google search data. Can you just explain what that is and how it's different from the way that previous studies looked at bullying trends?

Bacher-Hicks: Traditionally bullying research relies on surveys which have a number of limitations. For example, surveys require a lot of planning and a lot of resources. You have to recruit people; they have to respond. The biggest survey, at least of which I'm aware of in the United States, comes from the Youth Risk Behavior Survey conducted by the CDC, which is done every two years. The most recent data that's available from this survey are still from 2019, so this survey doesn't provide much information about bullying levels during COVID. So instead, we turned to Google Trends, which can provide a much more real-time sense of what's happening. Google Trends has been used in a wide range of studies to predict things like consumer behavior, voting patterns, and disease spread. But it had not previously been used to document a lot in education, and it had not been used to document bullying. We decided to try to explore this unique data source as a way to gather real-time information about what families and students were searching for as schools closed right around March of 2020. So the biggest benefit to being able to use Google Trends data is that it can answer these questions about bullying and other questions in real time.

Smith: Let's just get into what it was you found. Can you lay your results out? 

Goodman: I think that the paper is roughly divided into two parts. One is a part trying to make the case that pre-pandemic Google Trends data on searches for terms like “bullying” and “cyberbullying” really does seem to reflect real-world behavior—real-world bullying and cyberbullying behaviors. And then the second half of the paper is saying that if you now believe that we've made the case that Google Trends really does contain information about actual behaviors, let's see what's happening with bullying and cyberbullying when the pandemic hits. What we see in that second half is that online searches for terms related to bullying and cyberbullying plummet dramatically in a way that's quite unprecedented. If you look at the last five years of data on online searches, the headline takeaway is that the amount of bullying and cyberbullying that seems to be going on dropped roughly 30 percent, which is a pretty big number, particularly given that there has been prior work on bullying and cyberbullying suggesting that the negative impacts of these behaviors are pretty awful. Any kind of drop in them, we think, is good news, but this was a particularly large drop.

Smith: Were you able to drill into whether or not this drop in bullying was different for different types of students, in terms of socioeconomic status or grade level?

Bacher-Hicks: That speaks to one of the limitations of using the Google internet data. While it does provide real-time insights, it doesn't provide a lot of great information on who is doing the searches. For example, it doesn’t say what the nature of the precipitating incident that may have led to the search was. It limits some of the nuanced conclusions you can make based on the data from Google Trends. That said, we were able to examine what happened in the fall of 2020. It was the first new school year after the onset of COVID, and some schools began offering in-person instruction while others remained remote. What we were able to do is examine how searches for both in-person bullying and cyberbullying changed in schools that offer in-person instruction versus those that offered remote instruction at the beginning of this new school year. What we found is that both forms of bullying dropped no matter what type of instruction was offered. Bullying dropped if you offered in-person and if you offered remote. But the drops were far larger in areas that maintained remote instruction. So this provided some nice additional evidence that the shift away from in-person schooling is consistent with a reduction in bullying, both in-person and online.

I think it turns out that what we're finding is consistent with the sort of theories that cyberbullying is not something that necessarily arises online. At least some of it starts in-person and then shifts to online modalities.

Joshua Goodman

Smith: I think it's intuitive to think that there would be less bullying in-person if kids are remote, but it's this cyberbullying drop that certainly stands out. Why would cyberbullying drop as well?

Goodman: I think you're absolutely right that the result that bullying dropped when schools shut down is arguably not at all surprising. How can you actually bully someone in-person if you are not physically in a room with them or at least near them? I think the cyberbullying result is arguably the more surprising one, and I think what our findings are consistent with is a theory that most, or at least a substantial fraction, of cyberbullying is actually in-person bullying that moves online, or in other words, it has in-person origin. You can imagine two theories. One is that bullying and cyberbullying are compliments, and the other is that they're substitutes. I think it turns out that what we're finding is consistent with the sort of theories that cyberbullying is not something that necessarily arises online. At least some of it starts in-person and then shifts to online modalities. When you remove whatever those in-person contexts and triggers are, the online piece of it then also fails to materialize. I think that teaches us something about where at least some portion of cyberbullying comes from.

Smith: Do you think that these results offer any broader lessons for either schools or parents about handling bullying?

Bacher-Hicks: When schools returned for in-person learning in fall of 2020, the fact that we still saw drops in both in-person bullying and cyberbullying even when schools returned to in-person instruction suggests that there was something different going on that year in the fall of 2020, post-COVID, as opposed to the pre-COVID in-person instruction. One hypothesis is that there were a lot of additional structures that were put into place during schooling beginning in the fall 2020 year. Things like lunch and recess had a lot more structure to them than they did in the past. We know from previous research that a lot of bullying that happens originates in-person and then moves online—perhaps it occurs during these unstructured times. I think this provides some evidence for that theory and suggests that providing more structure to these times that have traditionally been quite unstructured could be a way to potentially reduce bullying moving forward.

The COVID-19 Pandemic Disrupted Both School Bullying and Cyberbullying” appears in the September 2022 issue of the American Economic Review: Insights. Music in the audio is by Podington Bear.