Valuing tap water quality using stated preference methods: Does the number of choice alternatives matter