When a Good Is a Bad (or a Bad Is a Good)—Analysis of Data from an Ambiguous Nonmarket Valuation Setting