Calculating the REAL Cost of Weddings

The press consistently reports that the average cost is nearly $30,000, basing that on one of several commonly cited surveys (such as surveys from the Wedding Report, wedding site the Knot and magazine publisher Condé Nast Bridal Media). Yet this figure is the mean — the sum of all wedding costs reported in the surveys, divided by the number of survey responses. Any one survey respondent who had a very expensive wedding could skew this number. A more useful representation is the median — the middle figure, when you line up all the costs in order. Let's see if it is true.


My print column this week examines how the median of these surveys — which isn’t commonly reported, but was shared with me by the surveyors — is a more useful representation, and how these wedding surveys may not be reaching people with the least-expensive weddings.

These issues have plagued the reporting of many other surveys. Darrell Huff begins his classic “How to Lie With Statistics” (which is on my recommended reading list) with a dissection of a stat, reported in the New York Sun and Time Magazine, that “The average Yalesman, Class of ‘24, makes $25,111 a year.” All of the issues plaguing the survey behind that stat — false precision, down to the nearest dollar; mean rather than median; and selection bias — also affect the average wedding-cost figures.

Yet these surveys are covered frequently, with little skepticism. For example, see this Associated Press article in June about wedding-insurance policies. Even a San Francisco Chronicle article about a new book on weddings quoted uncritically the average wedding-cost numbers. The author of that book, Rebecca Mead, questions the numbers in her book and in my print column this week.

The Gallup Organization conducted a survey of wedding costs in May, asking people to guess the average cost of a wedding these days. The mean estimate was $13,609, which Gallup pointed out was half of the Wedding Report’s estimate. Gallup’s news release, entitled “Americans Underestimate the Cost of Tying the Knot,” didn’t compare the medians of the two surveys. Frank Newport, editor-in-chief of the Gallup Poll, said medians aren’t commonly reported for these surveys. “If we had had the median available to us, it might have been appropriate to mention that, in addition to the mean,” he told me.

What do you think? Are these survey results representative of all American weddings? Is a mean or median more appropriate? Do you know exactly how much your wedding cost?

Further reading: Many of the articles citing wedding-cost figures also discuss wedding insurance, which covers cancellation or postponement. Some of these articles are collected on this Web site of one of the U.S. providers. A U.K. provider, Weddingplan, produces its own estimates that are frequently reported in the British press. (Weddingplan didn’t respond to my inquiries about their methodology.) You can see the Wedding Report’s cost-of-wedding survey here, and enter in your ZIP Code to see its local estimates here. (These are based on adjustments for cost of living and other factors, not on surveys for each locality.) The Centers for Disease Control and Prevention reports national marriage statistics in its National Vital Statistics Reports. In its review of Ms. Mead’s book, the Washington Times pointed out that wedding-cost stats could be inflated. Finally, on Slate, mathematician Jordan Ellenberg explained last week how mean and median can differ, and why that matters.