Because evangelicalism has no headquarters, no confession and no membership roll, defining it is a scholarly problem rather than an administrative one, and one definition has dominated the field since 1989.
The quadrilateral
David Bebbington proposed four marks. Conversionism, that lives must be changed. Activism, that the gospel must be expressed in effort. Biblicism, a particular regard for the Bible. Crucicentrism, a stress on the sacrifice of Christ. The four are deliberately about emphasis rather than doctrine, which is what lets them cover Anglicans, Baptists, Pentecostals and independents at once.
Where it is challenged
Three objections recur. The definition was built from British evidence and travels imperfectly to the United States, where evangelical identity is entangled with a particular political history the four marks say nothing about. It describes a temperament rather than a boundary, so it cannot say who is outside. And survey researchers who operationalise it find that self-identification and belief measurement select very different populations, which means the published numbers are not comparing the same people.
Why the definition matters practically
Almost every claim about how many evangelicals there are, and about how they vote, rests on which of these measures a pollster chose. The gap is not small. Studies using belief-based screens routinely return a substantially smaller and more racially diverse population than studies using the label, and the two are reported in the same sentence as though they were one figure.