Price research
Price dispersion begins with comparability.
A minimum, maximum, or average can be calculated from almost any list of numbers. The difficult part is deciding whether those numbers describe the same thing.
Define the price
Decide whether the analysis uses item price, delivered price, or another documented total. Shipping must not become zero merely because it is absent. Taxes, membership benefits, coupons, and location-dependent costs should be kept separate unless they are observed consistently and apply to the comparison.
Construct the comparable set
Match product model, variation, quantity, condition, bundle contents, and listing type. Record why an observation was excluded. A cheap accessory, damaged item, deposit, installment amount, or single unit from a multipack can otherwise become a misleading lower bound.
Use more than the average
The median can be less sensitive to extremes than the arithmetic mean, but neither explains the distribution alone. Report the range, middle spread, observation count, and visible clusters where the dataset supports them. With a small sample, use cautious language and show the individual comparable observations when possible.
Investigate outliers
An outlier may be an error, a mismatched product, a promotion, a distressed offer, or a real premium. Do not delete it only because it is inconvenient. Check its identity and terms, then either retain it with explanation or exclude it under a documented rule.
Time changes the interpretation
A marketplace price is an observation at a moment. Mixing old and current offers can create artificial dispersion. Record capture time and avoid describing a historical range as current without a new check. Seasonal events and short promotions require especially careful labels.
Conclusion
Useful price research explains the object being priced, the total-cost definition, the inclusion rules, the observation time, and the limitations. Statistics summarize that work; they do not replace it.