How many times a day do you turn to Google for a question? With 98% of American adults using the Internet daily, Google has also become a common part of our searches for information. So, how does this turn to Google affect our lives and communities?
SSRC scientists with the Innovative Data laboratory are asking this same question concerning searches about crime in the recent paper “Searching for Safety: Crime Prevention in the Era of Google.” Megan Stubbs-Richardson, assistant research professor, Karissa Bergene, research associate; Austin Cosby, undergraduate research assistant, and Art Cosby, SSRC director all worked on the project after discussion of how they could use Google Correlate in research.
“Studying public policy, I have an interest in how the Internet can shape or influence interactions between citizens and government. And anytime we can do a methodology that is innovative and different it’s neat for me,” said Bergene, a doctoral student in the Department of Public Policy and Administration.
Stubbs-Richardson, whose research has focused on crime, considered the theories in relation to how Google Correlate could be used, and after discussions with Bergene and Arthur Cosby saw an opening to look into crime prevention.
“People use Google for something as simple as how to zest a lemon while cooking to more serious matters like what is the best security system. We knew this data could help in many factors if we narrowed our questions,” Richardson said.
The team determined that for the first branch of this study, which was published in Crime Science in December 2018, they would look at property crime alone, although their first conversations had considered violent crime as well.
“The motivations and driving factors behind property and violent crime can vary. We knew to get the best understanding we needed to focus separately on each one at a time,” Stubbs-Richardson said.
The scientists began by looking at previous research using Internet crime searches. Other research has concentrated on violent crimes, police violence, and drug use. They then also considered the previous findings showing that property crime tends to decline in industrialized nations due to technological advances and other research that shows that “property crimes are driven by offenders’ perception of criminal opportunity” as they state in the paper.
In addition to pulling the data from Google, the scientists needed a data set that gave understanding to the amount and types of crime occurring in an area. The Uniform Crime Report (UCR) from the Federal Bureau of Investigation is where they found that source. The UCR has been published yearly by the FBI since 1930. It includes data from nearly 18,000 cities, colleges, and state, and federal law enforcement agencies. From this report, the researchers were able to define specific crimes and links within property crime.
With a background understanding, they used statistical software to aggregate the data on a date range from 2003-2014. With this data range, the researchers considered three main questions: Are higher rates for property crime associated with searches for crime prevention information?; Are increased levels of crime prevention searches associated with reductions in property crime?; and If there is a reduction, what is the magnitude of that reduction?.
These questions required the researchers to decide on what exact phrases they would consider for search inquiries. For this study, two lines of thought on crime prevention strategies were considered to hone in on for the phrases that they would study. Those two are situational crime prevention and community crime prevention.
“These emphasize the context of the place or situations in which crime is more likely to occur, and we selected certain search terms because these crime prevention approaches are known to reduce criminal opportunities and crime,” explained Stubbs-Richardson.
From these crime prevention methods, they utilized common and accepted phrases. These included terms like deadbolt, security door, alarm system, streetlights, report crime, and neighborhood crime. The next step was to merge the terms with the UCR data to look at the date ranges.
The researchers first looked at a general take on property crime by examining the average rate of property crime and its association with the selected crime prevention keywords. Then the team looked at the reduction in property crime as associated with keywords.
“Crime has been declining since the 1990s and criminologists have looked at several factors as to why that’s occurring, but there isn’t one clear answer,” said Stubbs-Richardson. “Our study included the impact of the Internet and what part that has possibly played in the reduction”
The first analysis did in-fact show a correlation between higher rates of Google searches on crime prevention and higher property crime rates. Seventy-six Pearson correlations were reported with 66 being statistically significant. Furthermore, the analysis found connections like a higher search for “car alarm system” being correlated with motor-vehicle theft or burglary rates having a strong correlation with “home alarm system” searches.
Next, the research team wanted to know if Google searches were related to a reduction in crime rates. As they stated in the article, “The correlational analysis indicated that 57.8% of crime prevention queries were significantly correlated with crime reduction.” The more searches a state had for crime preventive queries, the higher the rate of crime; however, those states experienced a higher drop in property crime over time. The team also ran ANOVAs to compare states with high, medium, and low frequency search groups. From this, Arizona showed the highest total crime reduction and the full results from that are also displayed in the publication.
“It shows that a correlation exists. Over time those areas with high crime prevention searches are exhibiting reduced crime rates. What Google data does not show us is whether an individual’s search for crime prevention items such as a deadbolt or home alarm system actually lead to a purchase. The intent of the search remains unclear. However, I think additional research in this area is needed because it may shed light on the impact Internet use can have on a community,” Bergene said.
With the first phase of the research published the team is looking forward to a second study where they will concentrate on violent crime.
“I believe the research can open new considerations for the way we look at the Internet and its effects on crime and crime trends. I hope it also shows the public what a networked resource the Internet is, whether looking at searches in our daily lives, including its impact on property or violent crime reduction,” Stubbs- Richardson said.
Stubbs-Richardson continues that when an individual searches for a product like a security system or locks, marketing systems are activated. Those searches then alert retailers who direct more online advertisements toward the individual and even their neighbors.
“The power of networks, including your neighbors, co-workers, peers, family, and access to technological resources, such as the unlimited access to information that the Internet provides, all work together to create a cumulative effect in crime reduction,” Stubbs-Richardson said. “Thus, while you may not buy that security system you’ve been window shopping for online, your next-door neighbor just might.”
The journal article is cited below:
Stubbs-Richardson, M. S., Cosby, A. K., Bergene, K. D., & Cosby, A. G. (2018). Searching for safety: crime prevention in the era of Google. Crime Science, 7(1). doi: 10.1186/s40163-018-0095-3