Average or median, the more fundamental problem is that this number is just wrong. Wedding industry statistics are spin from wedding-industry-affiliated companies, who have a tremendous incentive to convince you that it's normal to spend this kind of money on your wedding (as an average U.S. household--make sure you spend more if you are above-average!).
The source here, valuepenguin.com (which is trying to sell you wedding insurance), cites three data sources. Once source is no longer online. The other two sources (TheKnot and Costofwedding.com) claim to have conducted surveys. But The Knot just surveys people on its own website (not representative). Anyone want to venture a guess as to whether a representative number of courthouse-ceremony-and-lunch-buffet-afterwards weddings are counted in these figures?[1]
Meanwhile, Cost of Wedding is "owned and operated by The Wedding Report, Inc." and reports the results of a survey The Wedding Report did. And The Wedding Report has this gem in its methodology description:
We take a bottom up approach to "Average Wedding Cost." The total "Average Cost" is calculated using "Weighted Demand Average (WDA)," which is; average spent times demand equals the weighted demand average. Sum of weighted demand average for all items equals the average wedding cost. Using WDA instead of a summed average gives a better calculation, because it takes into account all items that couples may or may not purchase.[2]
As far as I can tell, this means the average cost is calculated by first determining average costs for each item you might possibly purchase, and then adding them all up as if every couple purchases every item! Umm... no. That's not how you do math.
[1] "Respondents were recruited throughout the year from TheKnot.com membership." https://xogroupinc.com/press-releases/the-knot-2017-real-wed...
[2] https://www.theweddingreport.com/index.cfm/action/home/view/...