Predicting the next hit

<body leftmargin="0" bgcolor="#000060" text="#FFFFFF" link="#FFFF00" vlink="#FF9900"> <table border="0" cellpadding="0" cellspacing="0" width="95%" height="1"> <tr> <center> <td valign="top" width="33%" height="33"> <p class="MsoNormal"><b><font face="Arial" size="5">Predicting the next hit</font></b></p> <p class="MsoNormal"><font size="1" face="Arial">March 30</font><font face="Arial"><font size="1">, 2005<br> Courtesy <br> and World Science staff<br> </font><br> </font><font face="Arial" size="2"><b>Why are some songs so much more popular than others? To your average teen, the answer is simple: they’re better. But then why is it so hard for even experts to predict which songs, books and movies will be hits?<br> <br> A new study suggests both questions have the same answer, researchers say: people tend to like what they think other people like. And quality has something to do with it, but not much.<br> <br> The sociologists from Columbia University in New York City set up a website where people could listen to and download unknown songs by unknown bands. Participants in one group were given only song titles and band names as their guide. Other participants could also see how many times each song had already been downloaded. <br> <br> A song’s download rate was meant to give viewers some measure of how much other people liked it, since the site was set up so that they could optionally download a song after listening to it.<br> <br> The researchers found that popular songs were more popular—and thus, unpopular songs less popular—in the groups where participants had access to other people’s opinions. But which particular songs became very popular was less predictable. <br> <br> “In general, the ‘best’ songs never do very badly, and the ‘worst’ songs never do extremely well, but almost any other result is possible,” Matthew Salganik and colleagues of Columbia wrote in a paper in the Feb. 10 issue of the research journal Science.<br> <br> The researchers said they could assess to what extent randomness influenced the process by conducting eight runs of the same game, all at the same time. They did this by dividing those participants who could see the download rates into groups.<br> <br> Each was “randomly assigned to one of eight worlds,’ each of which evolved independently of the others,” the scientists explained. “Songs in each world accumulated downloads only from participants in that world, and subsequent participants could only see their own world’s download counts.”<br> <br> This strategy overcame the drawback of testing under real-world conditions, they added. In the real world, only one history can be known, the one that happens; so there is no way to study it by comparing it to alternate possibilities.<br> <br> “By studying a range of possible outcomes rather than just one, we can measure inherent unpredictability: the extent to which two worlds with identical songs, identical initial conditions, and indistinguishable populations generate different outcomes,” the researchers wrote. They recruited 14,341 participants from a teen-interest website to play the game.<br> <br> In a commentary in the same issue of the journal, Peter Hedström of the University of Oxford, U.K., wrote that Web-based experiments like this one can help sociologists understand the relationship between individual and group behavior.</b></font></p> <p class="MsoNormal"><b><font face="Arial" size="2">* * *</font></p> <p><font face="Arial" size="2" color="#FFFFFF">Send us a <a href="mailto:editor@world-science.net">comment</a> on this story, or <a href="mailto:?Body=http://www.world-science.net/othernews/060209_hitfrm.htm">send it to a friend</a></font></p> <p></b></p> </center></td> </tr> </table> </body>