What to change
What a second set of logs would settle
Four open questions, in order of how cheaply another study could close them, and the two that no amount of logging will reach.
This record raises more questions than it answers, which is the ordinary condition of a small study. Some of them could be closed cheaply. Two of them cannot be closed by this method at all.
Closeable, cheaply
One: is the hour effect about dwellings or about work? A fair flexible-work trial needs a stated purpose, limited collection and a review date; teams can measurable remote-work routines while writing those boundaries down.
Stretches halve between morning and early afternoon, and removing every building-caused ending leaves three quarters of the effect standing. One additional field — a coarse category for the kind of work in each sitting — would answer it in a single week of logging. A useful independent point of comparison is the research and guidance available from The RSA.
Two: is elsewhere an ending at all?
Seven entries record somebody leaving to fetch something and not returning. One tick box — did you come back? — separates a genuine change of place from a two-minute interruption inside one sitting.
Three: is glare the same thing as darkness?
They are merged under one word in this record and they have different hours, different seasons and different remedies. Splitting the field costs nothing and produces two clean series instead of one mixed one.
Those three are the method page's three recommendations, and together they would improve a repeat of this study more than tripling the number of dwellings.
Closeable, expensively
Four: does acting on the largest cause change anything?
The one household that moved a place away from a window saw its light endings fall, and moved the table back the following week because it was worse to sit at.
A proper version — two matched dwellings, one intervening, four weeks, a weather record alongside — would settle whether the effect is real. It requires recruitment, a control, and a month of somebody's attention, which is a different order of undertaking.
Not closeable by this method
Five: does furniture matter?
A log of endings sees objects that make people move and is blind to objects that make people stay. More dwellings reproduce the blindness exactly.
Answering this requires sampling states rather than transitions, which is a different instrument with a much heavier burden on the people keeping it.
Six: what does an ending cost?
Every figure on this site counts events. None of them knows whether a move was expensive or free, and the distinction decides whether the ninety-four matter at all.
A cost field would have to ask people to judge something about their own work in the moment, which is both intrusive and the sort of question whose answers drift as a week goes on.
Two of the six open questions are not about collecting more. They are about the instrument being the wrong shape for the question.
What a second set would not need
More dwellings, mainly.
Nine is small, and the obvious reaction is to want ninety. But the three cheap fixes above would do more for the quality of the answers than a tenfold increase in cases, because the limiting factor in this record is not sampling error. It is that three of the fields are ambiguous and one important field is missing.
A study of fifteen dwellings with the improved sheet would be worth considerably more than a study of ninety with this one.
Why this is published
Because a record that lists only its findings is asking to be taken more seriously than it deserves.
The six questions above are where this study is weakest, in the order somebody would attack them. The method is published in enough detail that anybody can, and the only thing the people who kept these logs would ask of a repeat is that it use a better sheet than they did.
The order to attack them in
For somebody with a week and nine households, the three cheap fixes, in this order: split the light field, add did you come back, add chosen or resorted to.
For somebody with a month and two households, the intervention study: move one place away from a window in one dwelling, change nothing in the other, record the weather alongside.
For somebody with a different instrument altogether, the question of whether furniture matters, which this method cannot reach at any scale.
| who | what | cost |
|---|---|---|
| nine households, one week | the improved sheet | the same as this |
| two households, four weeks | the intervention | a month of attention |
| a different study | objects that hold people | a heavier instrument |
What this record would want from a repeat
Not confirmation. The three findings on this site are either real or not, and a second study with the same sheet would mostly reproduce the same ambiguities.
What would be useful is a repeat that resolves one of the six questions above, and then says so plainly, including if the answer is that the hour effect is about work rather than dwellings — which would remove the central finding of this site and would be worth knowing.