Toward a benefit-cost theory or survey participation: Evidence, further tests, and implications