Behavioral methods in infancy: pitfalls of single measures