Suppose that you have conducted a contingent valuation study to measure what adults would pay for reducing their risk of dying over a year. Half of the sample values a risk reduction of 1 in 10,000. The other half values a risk reduction of 5 in 10,000. The mean WTP for the 5 in 10,000 risk reduction is NOT 5 times the mean WTP for the 1 in 10,000 risk reduction. (Respondents do not pass a strong scope test.) Explain what implications this has for the VSL.