Sandwich generation meaning, and why it lands differently the second decade

The sandwich generation means one adult caring for a child and an aging parent at the same time. That is the whole of it, and I have set the definition out properly elsewhere.

What it does not contain is a length. It tells you who is in, and says nothing about how long you are in for, and the difference between a year of this and a decade of it is not a difference of degree.

I am about eighteen months in, if I count from the first wrong cupboard. Which is the problem in miniature. There is nothing to count from, so there is nothing to count.

Every answer on the page is a membership test

There are four questions a search engine will tell you people also ask about this phrase. Why it is called that. Whether millennials are it. Whether baby boomers are it. What the average age is.

Those are all questions about who.

I went through the rest of the page in September 2026. Nine results and an automated summary above them. The one at the top, settled on as the best answer to what this word means, is a Pew Research Center report from January 2013. Thirteen years old, and a headcount.

Not one of them says how long.

That is not laziness. It is what happens when a word is spatial. Sandwiched. Squeezed. Caught in the middle. Each is a position, and a position has no tense. You can be squeezed for an afternoon and squeezed for eleven years and the word does not alter by a letter. The name has other problems and I have written about those. This is not the part it gets wrong. It is the part it hides.

The surveys that count you do not ask

How long people spend caring has been measured. Nobody has ever been asked about the overlap.

AARP and the National Alliance for Caregiving surveyed 6,858 American caregivers between August and October 2024. Among the 6,549 caring for an adult, the average length of care was five and a half years; fifteen percent had been at it ten years or more.

Then look at what that question is about. It asks how long you have cared for one adult. The sandwiched appear elsewhere in the report as a headcount: 29 percent of caregivers have a child or grandchild under 18 at home. Duration on one page, membership on another, never crossed.

The Bureau of Labor Statistics does it inside one release. Of the roughly 2,500 American eldercare providers it identified across 2023 and 2024, half had been at it two years or less; the 7.6 million who are also parents of children at home are counted three paragraphs away.

And the instrument behind the quarter-of-American-adults figure, Pew's survey of 8,750 US adults fielded in September 2025, is four questions long. Do you have children under 18. Do you have adult children. Do you have a living parent in these age ranges. Have you given them money this year. That is the whole thing, and none of the four asks how long.

One estimate does exist, and it is not a survey. Diego Alburez-Gutierrez and two colleagues, modeling family structure across 198 countries in 2021, put the overlap across Europe and North America near 3.7 years, barely moving across birth cohorts from 1970 to 2040. What it counts is having a young child and a parent within five years of death. Who is alive, not who is caring, and nobody in it was asked anything.

The gap between a modeled 3.7 years across Europe and North America and what this feels like from inside is the whole subject.

What a long time does was argued about and never settled

The obvious guess is that longer is worse. The field tested that and did not get it.

In 1989, in Psychology and Aging, Aline Townsend and three colleagues followed 112 American adult children caring for an impaired parent, to test what they called the wear-and-tear hypothesis. They found improvement rather than deterioration in most of them, and depression unrelated to how long the caring had gone on. The paper cited as the statement of wear-and-tear reports against it.

Thirty-six years later there is still not one answer. A systematic review in Innovation in Aging in 2025 gathered 41 longitudinal studies measuring caregiver burden at three or more points in time. The most common pattern was a flat average. Burden rose mainly where the person cared for had dementia, and fell among those caring for somebody discharged after a sudden illness.

Nobody can tell you what ten years will do to you, and not for want of looking.

A decade in the provisional tense

If you believe you are in a phase, you decide provisionally. And you go on deciding provisionally for as long as the belief holds, which may be years after it stopped being true.

You do not take the role. You do not book the thing in April. You do not start the course, because the course runs two years and you will not still be doing this in two years. None of it feels like sacrifice at the time. It feels like sensible sequencing. You are about to be free, so you hold the good decisions back until you are.

Do that for ten years and you have not postponed a life. You have lived one, in the subjunctive, and never furnished it.

The hours are the smaller cost. The larger one is that the not-building is invisible.

This is personal observation rather than anything established. What I can tell you is that I am doing it now. There is a box in my hall of things for the house we will sort out later, and later has moved three times.

You do not wait it out. You become someone inside it

One finding about time I keep going back to, and it is not about being tired.

The long study of Alzheimer's caregivers at the University of California, San Francisco, which I have quoted here before, produced a second paper in 1996. Marilyn Skaff, Leonard Pearlin and Joseph Mullan followed 456 of the same spouses and adult children and asked about their sense of mastery, meaning whether they felt they still had a hold on their own life.

Among those who went on caring, mastery declined. Among those who moved their relative into a facility, it held steady. Among those whose relative died, it rose.

Read the first group again, because it is the one nobody is in a hurry to tell you about. Continuing is the condition under which it erodes. Not the hours. Not a particular hard week. The going on.

Something you do for long enough stops being a thing happening to you and becomes a description of who you are.

The same artifact as the one-year mark

I have argued in the matrescence chapter that it does not end at one year: the twelve-month mark is three unrelated clocks stopping near each other, none of which was measuring the woman.

This is the same defect from the other end of the family, inverted. There, a date was imposed that the evidence did not support. Here nobody sets one, so everybody supplies one, too ordinary to hear as a claim.

Once your mom is settled. When you get through this year. After the assessment. Each is an end date, offered by somebody who does not have one, about a thing none of those surveys asks about.

Imposed or absent, it comes to the same. You are measured against a line you did not draw.

Where this stops being mine to talk about

A long transition is not a license to wait, and I do not want any of the above read as one.

Tell a woman a thing lasts years and you have handed her a reason to sit with something that may not be that thing at all. So, plainly. If nothing lifts, if you cannot function, or if your own thoughts frighten you, that belongs with a doctor or a psychologist this week rather than after it. I am not a clinician and no website can tell you which of these you are in. I have set out what one woman's account can and cannot do elsewhere, and what the caring does to the person doing it is its own subject.

What I still do not know

Whether the second decade is the same mistake in better clothes.

There is a number in the title of this piece, and nothing I have read marks a threshold at ten years. So if the argument above is right, handing you a decade is another end date supplied by somebody who does not have one, and I am the somebody.

I have not worked out how to write about duration without accidentally issuing one, and I am not going to pretend otherwise in order to finish this well.

What to carry

No survey puts a clock on this, and that is not a gap in the reading. It is the thing itself.

So the question worth taking away is not when it ends. Nobody can answer that for you, and the people with answers are answering something else. It is whether you are willing to furnish a life you have been told is temporary.

You did not lose yourself the first time. You became someone else, and you are going to do it again, more slowly, with nobody announcing it. This chapter of the site calls these thresholds. A threshold is a thing you cross. This one you live on.