In 1995, Craigslist took a very old fashion service, the classified ad, online. Twenty years later, people in around 700 cities and 70 countries have used the site to advertise everything from second-hand furniture to 300 stuffed penguins. Craigslist changed how people swapped, shopped, found a partner and, according to a new study from the University of Minnesota, how they contracted HIV.
The study indicates that the site may have contributed to a 15.9 per cent increase in HIV infections across 33 US states.
‘I actually think that the creators of Craigslist had no intent of harming society. They came in with good intentions,’ says Jason Chan, assistant professor of Information and Decision Sciences at the Carlson School of Management. ‘At the same time, they did not anticipate that users could use the features in an unexpected way with unintended consequences.’
The study suggests that when Craigslist entered the market in 33 US states from 1998 to 2008, HIV cases jumped. At the national level, the increase associated with Craigslist translated to 6,000 additional HIV cases, costing the US healthcare system $62million–$65million to treat, according to the research.
Craigslist was chosen for the study because it was the first and largest online classified ad site. ‘By virtue of that, Craigslist became the ‘go-to place’ known to casual sex seekers in our study period,’ says Chan. ‘Given that Craigslist did not make an explicit effort to enter into specific locations based on the prevailing levels of HIV, site entry is likely to be random with respect with the HIV trends, which justifies the validity of the experiment,’ he suggests.
The researchers also compared the predicted ‘natural’ HIV infection rate based on multiple demographic factors with the actual rate where Craigslist was introduced.
‘The main strategy we used is the difference-in-difference econometric technique, commonly used in analysing natural experiments,’ says Chan. The difference-in-difference method inherently contrasts the HIV levels in places with and without Craigslist (using treated and control groups), before and after the site is introduced, he adds.
‘By taking these differences, the method accounts for prevailing HIV trends before Craigslist was present and accounts for the natural growth of HIV over time via trends in control groups. This method will eliminate a large proportion of external factors that may influence infection rates,’ says Chan.
Craigslist has a smaller footprint in the UK personals market than in the US, according to the Terrence Higgins Trust. Popular mobile apps that combine a user’s GPS-identified location with social networks make Craigslist personal ads look clumsy in comparison.
‘It is really important that the safer sex message stays strong, and that people know how to protect themselves and their partners. Apps themselves have a role to play in delivering this message, and we currently work with a number of them on this,’ says Dr Shaun Griffin, director of external affairs at the Terrence Higgins Trust.
‘Given the opportunity to analyse data from other sites and apps, I would be interested to compare how these different sites could influence the transmission of sexually transmitted diseases,’ says Chan.
Web personals influence more than just HIV infection rates. ‘We also tested the for a link to syphilis in our study and the result came back positive,’ says Chan. Understanding how social networks, apps and dating sites influence infection rates could be useful for future health policy, he adds.