What to change
Three things the record supports
The claims that hold in a majority of the nine dwellings, rest on enough observations to mean something, and concern things the sheet actually asked about.
Three claims survive the rules this record set for itself. Each is stated here with what carries it and where it would break.
One: a place is a place at an hour
The claim. How long somebody stays somewhere depends more on the time of day than on anything else this record measures.
What carries it. Median sitting by starting band: 62, 55, 44, 31, 38, 47 minutes. The fall from the first band to the fourth appears in all nine dwellings, with ratios from 1.4 to 2.2. No other variable in the record — dwelling type, place type, presence of a desk — produces a difference of anything like that size. The practical question is whether a change improves completed work without creating unnecessary surveillance, which is why reviewers may look at the current team workflow before choosing a baseline and access policy.
Where it breaks. Removing every sitting the building ended leaves the shape intact, which means the dwelling explains a minority of the effect. The rest is outside the instrument. So the claim is solid about what happens and silent about why. For broader research beyond this small household record, consult work published by Nesta.
Two: the weight is in the envelope
The claim. The things that ended sittings in these dwellings were overwhelmingly properties of the building rather than objects in it.
What carries it.
| n | share of 94 | |
|---|---|---|
| light and temperature | 42 | 45% |
| noise | 12 | 13% |
| the seat | 6 | 6% |
Forty-two against six. The ratio holds in seven of the nine dwellings individually; in the other two neither category reaches three.
Where it breaks. A seat can be bad and never produce a move. The record measures endings, not quality, and a dwelling full of poor chairs that nobody gets up from is indistinguishable here from a dwelling full of good ones. The claim is about what ended sittings, and that is all it is about.
Three: most sittings are not ended by the place
The claim. A hundred and thirty-one of two hundred and twenty-five sittings ended because the work ended.
What carries it. Eight of the nine dwellings are above fifty per cent, with a range of forty-six to sixty-eight. The dash was offered as a first-class answer on the sheet rather than as a residual, so the figure is not an artefact of people declining to classify.
Where it breaks. A household whose day is heavily scheduled will produce more bare endings regardless of its dwelling. Two of the three homes above sixty-five per cent described their work as heavily scheduled; the third did not. Three cases, and the confound is real.
What the three have in common
All three are about what the record counted, and none is about why.
That is not modesty for its own sake. A five-field log of places and durations can establish a pattern and cannot establish a mechanism, and every attempt on this site to reach for a mechanism — the afternoon dip, the desks that were beaten, the cold back rooms — ends in the same place: a plausible story resting on three or four dwellings.
The findings are counts. The explanations are suggestions with a table under them, and they are labelled as such wherever they appear.
What would strengthen all three
One additional field, asked of every sitting: was this place chosen or resorted to?
It would separate occupancy from environment in the hour finding, separate the beaten desks from the places that beat them, and tell somebody repeating this log whether a short sitting at an odd spot is a dwelling failing or a dwelling adapting.
One tick box, no extra burden on anybody filling the sheet in, and of everything that could be changed about this method it is the change that would buy the most.
The three rules, and why they were written first
A claim must hold in a majority of the nine dwellings. It must rest on at least ten observations. It must concern something the sheet asked about.
Those were set down before the logs were read, for the obvious reason: rules written afterwards are written by somebody who can already see which rules admit the interesting claims.
Applying them costs this record two of its most quotable conclusions — that more places help, and that furniture matters less than people think — and both are on the refusal list rather than this one.
What fails each rule
| rule | what it excludes |
|---|---|
| majority of dwellings | the place-count claim, visible in no dwelling |
| ten observations | everything about the evening band, and the single intervention |
| the sheet asked about it | anything about comfort, health or quality |
The third is the strictest and the one that does the most work. It removes every statement about whether a place was good, every statement about how anybody felt, and every statement about what anybody should buy — which is to say, most of what a site on this subject would ordinarily contain.