Bryce Ward

Who they are

Ward appears on this site for one reason: he is the end of a trace. Christine Rosen reports at page 212 of The Extinction of Experience that Americans spent about four hours a week with friends by 2019, down 37 per cent on five years earlier, and cites a newspaper column rather than the survey. The column is Ward's.

His own words are these: according to the American Time Use Survey, the amount of time the average American spent with friends was stable at six and a half hours per week between 2010 and 2013; then, in 2014, time spent with friends began to decline; and by 2019 the average American was spending only four hours per week with friends, a sharp 37 per cent decline from five years before. He adds that the figure fell to two hours forty-five minutes a week in 2021, a 58 per cent decline on 2010 to 2013. On a broader definition including neighbours, co-workers and clients, the time went from 15 hours a week a decade earlier to 12 in 2019 and 10 in 2021. The lost time, he reports, went to being alone rather than to spouses, partners or children, and the percentage declines are similar across sex, race, income, urban and rural, married and unmarried, parents and non-parents, and remote and in-person workers.

All of that is a plausible description of a real dataset by somebody competent to analyse it. What is missing is the apparatus that would let anybody check it. There is no published computation, no table, no code and no peer review, because an opinion column carries none of those. So the figure occupies an unusual and underdiscussed category. It is not a fabricated statistic. It is also not a verified one. It is an uncheckable one, and the distinction matters more than the usual true-or-false framing, because the uncheckable kind is far more common and far more respectable than the fabricated kind.

That is why this page sits next to the eight-second attention span. The goldfish figure has no research behind it at any step of its chain. Ward's figure almost certainly has a real computation behind it that no reader can see. The useful question for a reader is therefore not whether a number is true but whether it is checkable, and these two figures make the point without any need to lecture.

Two further things belong on the record and neither is an adverse judgement of him. His attribution sentence, that social media, political polarisation and new technologies all played a role in the drop, followed by the parenthetical observation that smartphone market penetration crossed 50 per cent in 2014, is an unsourced causal assertion sitting inside a descriptive analysis, and the parenthesis offers a coincidence of dates as a hint. This publication does not carry it. And a small accuracy point: the American Time Use Survey is conducted by the Census Bureau for the Bureau of Labor Statistics, which publishes it, so calling it the Census Bureau's survey is loose.

The peer-reviewed counterpart to his analysis exists. Vijay Kannan and Peter Veazie published an analysis of social isolation, social engagement and companionship in the same survey in SSM Population Health in 2023. This publication could not read it: the repository returned a reCAPTCHA challenge and a second endpoint rate-limited, so the paper is recorded as a read failure and is not cited for any figure.

Key papers

Ward is an economist working in consultancy rather than a trade author, and this publication has verified no book of his. It also has no paper of his to list in this territory, which is the point of the page rather than a gap in it. The single item below is an opinion column and is labelled as one.

Ward, B. (2022). "Americans are choosing to be alone. Here's why we should reverse that." The Washington Post, opinion section, 23 November 2022. A descriptive analysis of American Time Use Survey data on time spent with friends, published without a method section, a table, code or peer review. This is not a paper and must not be cited as one.

Key ideas

The source of the circulated figure: time with friends stable at six and a half hours a week from 2010 to 2013, four hours a week by 2019, a 37 per cent decline, and two hours forty-five minutes by 2021.

The decline is reported as broad, with similar percentage falls across sex, race, income, urban and rural populations, and across marital and parental status.

The figure is uncheckable rather than either verified or false, because the venue carries no method.

The lesson for a reader is checkable against uncheckable, which is a more useful test than true against false and catches far more of what circulates.

---