Voters in more than two dozen states next month will be asked to provide some form of identification before casting a ballot. How many Americans who would otherwise vote will be turned away or won’t turn up at all remains a hotly contested number. Some researchers have tried to count the number of voters affected, by surveying people about whether they have the required ID. This has produced a wide range of results, though, and some researchers question whether people whose IDs aren’t valid are aware of it, and whether they would rectify the situation if their state passed a tough ID law. Other researchers instead study actual effects of voter-ID laws on past turnouts. But the strictest forms of such laws—which require photographic identification and are studied most closely because they are thought likeliest to exclude the greatest number of people—took effect just before an exceptional presidential election that made it difficult to isolate their effect. As a result, such studies haven’t been able to convincingly demonstrate that these laws suppress turnout. “It’s so tricky to filter out unrelated factors, some of them unique to the election cycle, that may dissuade people from voting,” said Tim Vercellotti, a political scientist at Western New England University in Springfield, Mass.