The two homes with no desk

The two, side by side

Home B has the second longest median in the study and the second highest share of bare endings. Home F has the shortest median, the most sittings, the most places and the most endings. For another evidence base on people, work, or environments, see University of Nottingham.

Home B: two places and not much happening

A table and a sofa. Twenty-one sittings across five days, which is four and a bit a day, slightly under the record's median of five.

Seven endings in the week: two light, two noise, one surface, one power, one unsaid. No temperature. No wanted — nobody else was competing for the table. No seat.

Three of the five days were spent entirely at one place.

What this looks like is a dwelling where very little intervenes. The person sits at the table, works for about an hour, stops when the work stops, and does it again. Whether that is a good week or a monotonous one is not something the log can address, and the household did not say.

Home F: six places and constant movement

A bungalow with a kitchen counter, a dining table, an armchair, a back room, a second armchair and a sideboard.

Thirty-one sittings, the most in the record, with a median of thirty-five minutes, the shortest. Fourteen endings, also the most: five temperature, two each of light, noise and elsewhere, one each of wanted, surface and seat.

Never a day in one place. A median of three distinct places a day against the record's two.

The comparison that does not work

The obvious thing to do with these two is to ask which arrangement is better, and the record cannot answer it in either direction.

Home B's long stretches could be concentration or they could be a person with nowhere else to go. Home F's short ones could be a dwelling that keeps failing or a person who likes moving about a house with six places in it.

A median of sixty-one and a median of thirty-five are two facts about two dwellings. Neither is a score.

What they jointly establish

That having nowhere set up for work does not determine anything.

If a designated place were the thing that held people in position, the two dwellings without one would resemble each other, and they are at opposite extremes of the main measure in the study.

This is the most direct evidence in the record against the simplest version of the furniture argument — more direct than the six seat endings, because it concerns an arrangement rather than an object, and because it comes from the two cases that the argument would most clearly predict.

The confounds, which are large

Home F is a bungalow with six usable surfaces. Home B is a one-bed flat with two. Those are different dwellings in a way that swamps the presence or absence of a desk.

Home F also produced five of the record's eighteen temperature endings, which is a fact about a single-storey dwelling in a cold month and has nothing to do with how anybody arranged their working.

And both are one household each, for five days.

What they can carry is a negative: these two dwellings do not show the pattern that "no desk" would predict. What they cannot carry is any positive claim about what does determine how long anybody stays anywhere.

Why not simply recommend a desk anyway

Because seven of the nine homes in this record had one, and in four of those seven it was the place where the dwelling most often intervened.

A recommendation that holds in three of seven cases, fails in four, and is contradicted by the two cases that lack the thing entirely, is not a recommendation. It is a preference with a table under it.

What they did instead

Home B worked at a table and a sofa, and three of its five days never left one of them.

Home F worked at six places and never spent a day at one. Its median day used three distinct spots, against the record's two, and it produced the only two single-use places in the record that were not preceded by an ending at the dining table.

Neither household described this as a problem. Neither described it at all: the end-of-week notes from both are about other things, and Home F's mention the kitchen counter approvingly.

What a desk would have changed

Unknowable, and worth stating as unknowable because the question is the obvious one.

Home B's stretches are already the second longest in the record. There is little room for a desk to improve them and the dwelling has nowhere obvious to put one.

Home F's are the shortest, and a desk would have to compete with six existing places in a dwelling where somebody evidently likes moving between them. Whether a seventh place, differently furnished, would have held that person longer is exactly the sort of question nine observational logs cannot answer.