A note on estimating visitor spending on a per-day/night basis

There are two approaches to calculate a per-night spending average when visitor expenditures are gathered on a per-trip basis. The first and correct approach is to individually compute averages for trip spending and length of stay from the sample and then divide the two figures. The second approach, although more convenient and instinct-driven, is to compute per-night spending on a case-by-case basis and then compute the average of the per-night figures across all samples. This calculation procedure will generally produce an inflated per-night spending average, as per-day spending tends to decline with length of stay, and a biased variance.