on writ of certiorari to the supreme court of
[January 13, 2004]
Justice Stevens, with whom Justice Souter and Justice Ginsburg join, concurring in part and dissenting in part.
There is a valid and important distinction between seizing a person to determine whether she has committed a crime and seizing a person to ask whether she has any information about an unknown person who committed a crime a week earlier. I therefore join Parts I and II of the Court's opinion explaining why our decision in Indianapolis v. Edmond,531 U. S. 32 (2000), is not controlling in this case. However, I find the issue discussed in Part III of the opinion closer than the Court does and believe it would be wise to remand the case to the Illinois state courts to address that issue in the first instance.
In contrast to pedestrians, who are free to keep walking when they encounter police officers handing out flyers or seeking information, motorists who confront a roadblock are required to stop, and to remain stopped for as long as the officers choose to detain them. Such a seizure may seem relatively innocuous to some, but annoying to others who are forced to wait for several minutes when the line of cars is lengthened--for example, by a surge of vehicles leaving a factory at the end of a shift. Still other drivers may find an unpublicized roadblock at midnight on a Saturday somewhat alarming.
On the other side of the equation, the likelihood that questioning a random sample of drivers will yield useful information about a hit-and-run accident that occurred a week earlier is speculative at best. To be sure, the sample in this case was not entirely random: The record reveals that the police knew that the victim had finished work at the Post Office shortly before the fatal accident, and hoped that other employees of the Post Office or the nearby industrial park might work on similar schedules and, thus, have been driving the same route at the same time the previous week. That is a plausible theory, but there is no evidence in the record that the police did anything to confirm that the nearby businesses in fact had shift changes at or near midnight on Saturdays, or that they had reason to believe that a roadblock would be more effective than, say, placing flyers on the employees' cars.
In short, the outcome of the multifactor test prescribed in Brown v. Texas,443 U. S. 47 (1979), is by no means clear on the facts of this case. Because the Illinois Appellate Court and the State Supreme Court held that the Lombard roadblock was per se unconstitutional under Indianapolis v. Edmond, neither court attempted to apply the Brown test. "We ordinarily do not decide in the first instance issues not resolved below." Pierce County v. Guillen,537 U. S. 129, 148, n. 10 (2003). We should be especially reluctant to abandon our role as a court of review in a case in which the constitutional inquiry requires analysis of local conditions and practices more familiar to judges closer to the scene. I would therefore remand the case to the Illinois courts to undertake the initial analysis of the issue that the Court resolves in Part III of its opinion. To that extent, I respectfully dissent.