Why not? It's a REAL person solving a REAL problem for himself. Why is that "invalid data"?

You prefer studies with 100 people that generate just noise because people are so different?