Reading your own week

You have a week on paper, or you soon will. This lesson is what to do with it, and the first thing to say is that the answer is not to add it up.

A total is the least informative thing a record contains. It tells you what you already half knew, it varies from week to week for reasons that have nothing to do with you, and the only thing it can be compared against is the prediction you wrote before you started, which is what the next two pages are about. The information is somewhere else, and finding it takes two passes over the same sheet of paper.

The first pass looks for disagreements. The second sorts every hour into three columns. Both produce a number or a written line, and you should finish it with a figure for how many hours of your week you actually decide about. Every later lesson in this course works on that figure and not on your week.

Pass one: put the prediction beside the record

Lesson 1 asked for four numbers before you began, and asked you to keep them. This is what they were for.

Write your four predictions in one column and your four recorded totals beside them. Then a third column, the difference, with its sign. That's the whole instrument.

Lesson 1 gave you the measured version on United States and Belgian national samples, published in 2011: employed respondents "tend to overestimate their work hours by 5–10 percent in relation to the work hours they report in their time diaries", with the gap widest among those estimating the longest weeks.¹ You now have the same comparison on a sample of one, which is worse evidence about people in general and much better evidence about you. Lesson 1 set out why the two are different instruments rather than one instrument used with more or less care, and the short form is that a record is kept in sequence against a day that has to come to twenty-four hours, while an estimate is a calculation made in a few seconds with nothing constraining it.²

Expect three things, and none of them is a failure.

Some categories will agree closely. Good. You have learned that your picture is roughly right there, and you can stop attending to it. This course expects most readers to find at least one category like this, often sleep, because sleep has a clock at both ends. That is an expectation rather than a measurement.³

One or two will be out by a lot. Lesson 1 gives you the measured expectation per category: paid work and housework tend to come in high, sleep and free time tend to come in low. What the research does not tell you, and what this course will not pretend to, is which of your own four it will be.

And something will turn up that was not in your four categories at all. This course expects that to be the commonest outcome and the most useful line on the page, which is again a judgement rather than a finding.³

Why the surprises carry the information

Now the part worth being precise about, because it's what makes a week's record different from a week's diary kept for its own sake.

You predicted first. That single step turned the record from a description into a comparison, and a comparison can be wrong in a way a description cannot.

Where the prediction and the record agree, you have learned that your model of your own week works there. That's a real finding and it lets you stop looking.

Where they disagree, you have found a specific place your model of your own life is wrong, and the record names it. Not "I am bad with time", which is not a finding and not actionable, but "I believed Thursday evening was three hours of study and it was forty minutes twice".

That's the only kind of thing one week can establish, and it is worth being clear about why. One week is a sample of one. Logic and Argument lesson 5 gave you the general form of this: a sample that small supports nothing about a rate, because a single observation cannot tell you what varies and what does not. What it can support is an existence claim, which is a different shape of statement altogether. "At least one of my four categories is badly wrong" needs one week. "I spend eleven hours a week on domestic work" needs many, and no reader of this course is going to collect them.

So read your record for existence claims. They're cheap, they're true, and they are the only thing on the page that a second week is unlikely to overturn.

Two weeks, two people, and the category nobody predicts

Ade is constructed, and so is Hannah below, and so is every figure attached to either of them.⁵ The callout after this section says why.

Ade predicted four numbers: 30 hours of paid work, 12 of domestic work, 52 of sleep, 20 of leisure. He works four days a week at a builders' merchant and spends part of most weeks helping his mother, who is seventy-nine and lives two streets away.

His record came back close on three of them. Paid work 31 hours 30, sleep 49, domestic work 13 and a half. Leisure came to 9.

Now the arithmetic almost everybody skips, and it is worth doing slowly. A week has 168 hours. He slept for 49 of them, so he was awake for 119. His four categories account for 103 hours, but 49 of those are the sleep, which leaves only 54 of his waking hours inside a box he predicted.

So 65 waking hours sat outside all four, which is more than nine hours a day.

That number is not a scandal and it is not unusual. Most of it is the ordinary business of being alive: eating, washing, getting places, the twenty minutes between one thing and the next, standing in a queue, the half hour after you meant to get up. Everybody's residue is large. The four categories you chose were never going to cover a week, and the exercise was never asking them to.

The finding is the largest named thing you can pull out of that residue. For Ade it came to eleven hours, and it was one recognisable activity: phone calls to the surgery, a form for the council that needed a document he had to find, two trips to the pharmacy, an hour on hold with the energy supplier about his mother's account, and a Tuesday afternoon at an appointment. He had recorded every one of them accurately. He had simply never had a name for them, so before the week began there was no box to predict them into.

Notice that his leisure shortfall is also eleven hours, and that this is not a coincidence. Every hour that goes into an unnamed activity had to come from somewhere, and in his case what it came out of was the box he was least sure about. That is worth watching for on your own sheet: the category that shrinks most is often the one you were using as a mental remainder rather than a real estimate.

Ade's finding is not that he has too little leisure. It's that eleven hours of his week belonged to an activity he had no name for, and that the activity is not domestic work and is not paid work. It is administration performed on somebody else's behalf, and this course expects it to be the largest unnamed thing in the week for a good many of its readers, which is the course's own expectation rather than anything measured.³

Why the missing category is so common is the third core idea of this lesson, and it's slightly uncomfortable. The four categories you chose before you started didn't just organise what you found. They shaped what you were able to see. Every hour that did not fit went into whichever box was nearest, or went unrecorded, and the closer you stuck to your four the less you learned.

The practical form of that is not "the residue is small", because it never is. It is this: go into the residue and name the biggest thing in it. Most of what is there is the ordinary business of living and needs no attention. What you are looking for is the one recognisable activity, like Ade's eleven hours, that had no box because you had never given it a name.

Pass two: fixed, discretionary, contested

Now the sort, and a warning about it first.

This three-way distinction is this course's own, not a finding from the research.³ Nothing in the literature this course read defines a "discretionary hour" or measures how many of them anybody has. The course uses it because the rest of the course needs a name for the hours a technique could possibly act on, and because the alternative, which is to write as though every reader has an open calendar, would make the course useless to most of the people reading it.

Take every recorded hour and put it in one of three columns.

Fixed. Hours you do not decide about this month. A shift you are rostered onto, the school run, a dialysis appointment, the hours your body needs for sleep, the trip between two of those. The test is not whether the hour is pleasant. It's whether a decision of yours this month could move it.

Discretionary. Hours where the decision is genuinely yours. Not "hours I have nothing planned in", which is a different and much larger set, but hours where you could do a different thing without somebody else's agreement and without a cost you are unwilling to pay.

Contested. Everything you hesitated over. And rather than forcing it into one of the other two, write the condition beside it.

That third column is the one that does the work, and it is worth an example. "Saturday morning at my mother's, fixed unless my brother takes one in three." "Tuesday 7 to 9, the class, discretionary but I've paid for the term." "The hour after the children are down, discretionary on the nights I'm not too tired to use it, which is about half."

Each of those was a vague feeling before it was written, and each is now a statement with a condition in it. A condition can be asked about, checked, or found to be false. A feeling cannot. You don't have to act on any of them today. You have to get them out of your head and onto the sheet, because lesson 6 is about deciding what not to do, and it operates entirely on this column.

Worked: the week whose total was fine

Hannah is self-employed, doing the books for four small businesses, and she works around a six-year-old. She predicted 12 discretionary hours. Her record came back at 11 hours 20 minutes, which is close enough that she nearly stopped there.

The total was not her finding. When she marked the discretionary time on the sheet it came in thirty-one separate pieces. The longest was 55 minutes. Nineteen of them were under fifteen.

Eleven hours in three blocks and eleven hours in thirty-one pieces are not the same resource, and no total will ever tell them apart. Some things she wants those hours for, such as the bookkeeping qualification she has been meaning to start, don't fit in fourteen minutes at all. Others fit perfectly well.

This is lesson 5's subject arriving three lessons early, and it's deliberate, because you're looking at your own sheet now and it's the last moment where noticing it costs you nothing.

One measured figure, with its scope attached, because the shape is not unique to Hannah. A field study published at CHI 2005 followed 24 information workers at a single outsourcing company, observed to the second, and found that the average time spent on a working sphere "was 11 min. 4 sec. (sd=18 min. 9 sec.)" before the person switched or was interrupted, with 57.1 percent of those segments interrupted.⁴ The spread there is larger than the average, which is the study making Hannah's point on its own data. That's an office, in one company, about twenty years ago, and it isn't a measurement of anybody's evenings. This course has no measurement of how discretionary hours arrive at home, and the honest thing is to say so and to point you at your own sheet, which you now have and the researchers did not.

So when you count your discretionary hours, count them twice: the sum, and the shape. The shape is how many pieces, and how long the longest one is.

What a week can and cannot tell you

Three limits, stated plainly, because a record is easy to over-read.

A week cannot give you a rate. Not your average week, not your typical Tuesday, not how many hours a year you spend on anything. Those all need repetition you are not going to do, and a claim of that shape from a single week is the error Logic and Argument lesson 5 was about.

A week cannot tell you whether it was typical. Nothing inside the record can, which is why the fix is not statistical. Name the week. Write one line at the top of the sheet: what was going on. School holidays, a deadline, a visitor, two days ill, an ordinary week as far as you can tell. That line costs thirty seconds and it is what stops you, three months from now, quoting a figure from an unusual week as though it described your life.

And note that an unusual week is still worth having. The surprises survive it, because a surprise is about your picture and not about the rate. Ade would have found his eleven hours in almost any week.

A week cannot tell you what to do. It has no view on whether nine hours of television is too many, and none on whether eleven hours of caring for somebody is too many either. That judgement is yours, and this course isn't going to make it for you: lesson 1 names the question of whether a person should aim to do more or to do less, and lesson 6 is where it is faced, because that is the lesson it would otherwise decide silently.

Three things people get wrong about their own record

"The total is the finding." It's the least informative line on the page. It has no comparison attached unless you predicted first, it moves for reasons outside you, and it cannot distinguish Hannah's thirty-one pieces from three clean blocks. The differences and the residue are the finding.

"I should keep doing this forever." The cost is real, in attention rather than in minutes, and it holds steady. The return falls sharply: your second week tells you much less than your first, because the first one used up all the surprises. A further week needs a reason of its own, and reasons are easy to have, such as a new job, a term that works differently, or a month where you plan to change something and want the before.

"My week is typical." You can't know that from inside it, and the sentence usually means "I would like the total to be fair". Name the week instead, which is both more honest and more useful, and read it as the week it was.

Practice

Connections

Back. Lesson 1 measured the gap between estimate and record on national samples and asked you to open the same comparison on yourself; this lesson reads the result. The limits on what one week supports come from Logic and Argument lesson 5, on what a sample can carry, and lesson 4's distinction between "some" and "all" is exactly the difference between an existence claim and a rate. Using AI Effectively lesson 11 gave you the habit of asking what a measurement was taken on before believing what it says, which is why the 2005 office study above arrives with its scope attached.

Forward. Lesson 3 is why your estimates of single tasks run short, which is this lesson's gap at the scale of one job rather than one week. Lesson 4 plans inside the discretionary hours you just counted. Lesson 5 is about the shape of those hours, which Hannah's sheet has already shown you. Lesson 6 works entirely on the contested column, and is written for the reader whose week is mostly fixed. Lesson 8 returns to what a record is worth and to the argument about whether the gap is real.

Go deeper

• The American Time Use Survey. Worth twenty minutes to see what a category looks like when an organisation has had to define one and defend it. This course has read the article that uses its data, not the survey's own category documentation, so treat the visit as a look rather than as a reading. United States only.
• The overestimated workweek revisited (Monthly Labor Review, June 2011). Free. This course has read the standfirst and pages 43 to 45. Page 44 is the part relevant here: it sets out why a sequential record and a summary estimate are different instruments rather than one instrument used with different care.
• No Task Left Behind? Examining the Nature of Fragmented Work (CHI 2005). This course has read the abstract, the research setting and the results on fragmented work. Ten pages of a field study in one office, and the interesting thing about it is how hard the authors worked to measure something most people assume they already know.

Sources

1. John P. Robinson, Steven Martin, Ignace Glorieux and Joeri Minnen, "The overestimated workweek revisited", Monthly Labor Review, June 2011, pages 43 to 45. Read in part. Supports: the 5 to 10 percent overestimate and that it is concentrated among those estimating the longest weeks. United States and Belgian data.
2. The same source, page 44, for the account of why a sequential record and a summary estimate are different instruments: respondents keeping a diary "are not asked to make complex, vague, or subjective calculations, but simply to recall their activities sequentially for a specific period", and the method carries "an implicit constraint on the numbers of hours in a day". Lesson 1 sets this out in full and this lesson relies on it rather than restating it.
3. The three-way sort into fixed, discretionary and contested is this course's own, stated as such in the body where the reader meets it. It is not drawn from any source in research/SOURCES.md, and no study read for this course defines or counts discretionary hours. It exists because the rest of the course needs a name for the hours a technique could act on.
4. Gloria Mark, Victor M. Gonzalez and Justin Harris, "No Task Left Behind? Examining the Nature of Fragmented Work", CHI 2005, pages 321 to 330. Read in part: the abstract, the research setting, and the results on fragmented work, pages 321 to 325. Supports: the average of 11 minutes 4 seconds in a working sphere before a switch or an interruption, and that 57.1 percent of working sphere segments were interrupted. Scope: 24 information workers at one outsourcing company, published at CHI 2005; the fieldwork is not dated in what this course has read. It measures an office and not a household, which the body says at the point of use.
5. Ade and Hannah are constructed, along with every figure attached to them, and the lesson says so in a callout at the point the first one appears. No source in this course's research describes an individual's week.