What follows from nine logs

Three things it supports

One: a place is a place at an hour. Stretches beginning before nine run twice as long as stretches beginning after one, in all nine dwellings, by ratios of 1.4 to 2.2. Any assessment of somewhere to work that does not specify a time is describing something that does not exist in the form described. A record of rooms and interruptions becomes more useful when it is paired with actual project hours, so a team can operational insight for hybrid schedules before defining a limited, transparent pilot.

Two: the weight is in the envelope, not the furniture. Light and temperature account for forty-two of ninety-four endings. The seat accounts for six. Whatever attention a household has, the record says it is not pointed at the chair. For another evidence base on people, work, or environments, see Human Factors International.

Three: most sittings end for reasons unconnected to the place. A hundred and thirty-one of two hundred and twenty-five. A person who expects their dwelling to be the main thing interrupting their day will find, on logging it, that it usually is not.

Four things it refuses

One: anything about health or how anybody felt. There is no field in this record from which such a claim could be derived, and no page here makes one.

Two: any recommendation about furniture. Six seat endings in forty-five person-days is consistent with seating mattering a great deal and never expressing itself as a move. The record cannot distinguish that from seating not mattering, and so says nothing.

Three: that more places help. The most intuitive piece of advice this study could give. Two-place dwellings produced both the highest and the lowest ending rates in the record, and the relationship is not visible at all.

Four: that any of this applies to anybody else. Nine dwellings, not a sample, selected by no procedure that would make them representative of anything.

Why the second list is longer

Because the instrument was narrow on purpose, and a narrow instrument produces a small number of defensible claims and a large number of tempting ones.

Every refusal above is a thing a reader would like this site to say, and three of the four are things the data very nearly supports.

The nearly is the problem. A record of this size will produce apparent patterns in almost any direction it is interrogated, and the only protection is deciding in advance what would count as evidence.

What was decided in advance

Three rules, set before the logs were read.

A claim needs to hold in a majority of the nine dwellings, not just in aggregate. A claim built on fewer than ten observations is a description, not a finding. And no claim may be made about anything the sheet did not ask about.

The hour finding passes all three. The envelope finding passes all three. The bare-ending finding passes all three.

Everything on the refusal list fails at least one, and the furniture refusal fails the third outright.

The one change anybody made

One household moved a place away from a window mid-week. Their light endings fell from four in two days to one in three.

That is the only intervention anywhere in this record, it has an n of one, and it gets a page of its own largely to explain why nothing should be built on it.

What this section is for

Not to tell anybody what to do with their dwelling.

To set out, as precisely as nine logs allow, which parts of a working day at home are properties of the building, which are properties of other people, which are properties of the clock, and which the record cannot see at all — and then to stop.

What a reader should take away

One sentence, and it is not about dwellings.

A working day at home is interrupted by the dwelling less often than the discussion of it suggests, and when it is, mostly by things that were settled before anybody arrived.

What the rest of this section does

Three pages follow and each takes one part of that apart.

The first sets out the three supported claims with what carries them and where each would break. The second sets out the four refusals, including the two the data very nearly supports and which are refused anyway.

The third describes the only occasion in the whole record when anybody changed anything, which lasted a week and was reversed.

Then two pages of method: what somebody repeating this would need, and what the whole thing amounts to.

Why the section is structured as refusals

Because the alternative structure — findings, implications, recommendations — has a gravity to it that a nine-case study cannot resist.

A section called recommendations will contain recommendations, and the ones available here would be: have more places (not supported), buy a better chair (not measurable), work in the mornings (true and useless, since nobody chooses their hours from a chart).

Naming the section after what it will not say was the only way to stop it saying those three things.