4 percent of the messages in Purdue's test weren't delivered.

With some 50 or so companies now marketing emergency text-messaging systems to schools and colleges in the wake of the Virginia Tech shootings last April, campus safety officials have been deluged with promises about how well these systems are able to get the word out quickly in the event of an emergency. But there has been little independent research to corroborate these claims—until now.

On September 24, researchers at Purdue University conducted an experiment to test the speed and reliability of text messaging as a form of emergency communication. They found that text messaging performed well overall, with fewer than 4 percent of messages failing to reach their intended target. Still, in a crisis situation, any failure to reach students or staff members can be critical.

Beginning at 11:25 a.m., the university sent 9,979 text messages to volunteers who signed up for the experiment. Volunteers were asked to reply as soon as they got the message. The text of the message read: “This is a TEST of the Purdue emergency notification system. In a real emergency, check www.purdue.edu for details.”

As of 6 p.m. on Sept. 24, researchers had received replies from 5,741 of the text-message recipients, with 2,915 replying within the first 10 minutes after the message was sent. Scott Ksander, Purdue’s executive director of information technology networks and security, said officials are talking with people who said they either did not reply to the text message, or never received it, to find out what happened.

Ksander said he knows of only 364 text messages that were not delivered.