The science behind measuring the mental load

In 2019, sociologist Allison Daminger published an analysis in the American Sociological Review that changed how the mental load can be discussed. Instead of a vague block of "invisible work", she describes a cognitive cycle of four to five phases that every household task moves through:

1. Anticipate — spotting a future need before it becomes a problem. "We are going to have to sign the kids up for the February holiday club."
2. Identify — searching, comparing, gathering the possible options.
3. Decide — choosing between those options.
4. Execute — doing the concrete thing, the one you can tick off.
5. Monitor — checking it actually got done, and staying accountable if it did not.

The decisive contribution of this breakdown: the phases are not distributed the same way within a couple. Execution is often the most shared. Anticipation and monitoring, almost never. That is where the gap sits between "we share everything" and the exhaustion of exactly one of the two.

Mental Loadless measures the mental load phase by phase. That is the whole difference between counting actions and understanding who carries the household.

A task list records who did what. That is accurate information, and it is the wrong information.

Take a real task: signing the child up for school lunches. B thinks of it in September, B finds the form, B decides how many days, B reminds A to drop it off, A drops it off, B checks on Monday that the registration went through. In a to-do list, the line is ticked by A.

A task executed by A but anticipated, planned and verified by B is B's load. That is what the Mental Loadless cognitive graph captures: each task carries, alongside its executor, who anticipated it, identified it, decided it and monitored it. Plus its domain (8 life domains), its cadence, its emotional weight, and the negotiated standard — "it is done when…", agreed by both, which defuses the re-checking described by Mederer as far back as 1993.

And none of it asks anything more of you

The obvious risk of such a model would be to make creating a task unbearable. There is not one extra field to fill in: the creator is presumed to be the anticipator, the assignee the executor, the rest is inferred, then refined by a weekly heuristic pass that uses no AI at all. A 12-question onboarding establishes the likely referent for each domain. You use the app normally; the measurement builds itself behind you.

The MLI (Mental Load Index) expresses the share of the household's mental load carried by each adult, on a 0-100 scale that sums to 100. "We're at 68/32" is a sentence you can say out loud — that is deliberate.

For every owned phase, the calculation multiplies the phase weight, the cadence weight, the domain weight and the task's emotional weight; the total is then normalised.

The weights, version 1.0

• Anticipate — 1.5
• Monitor — 1.25
• Identify — 1.0
• Decide — 0.75
• Execute — 0.5

Execution weighs the least. That is a deliberate inversion of every family organisation app on the market, which counts precisely the opposite. A decision genuinely made together counts as zero on both sides: a shared call is nobody's load.

These numbers are not a trade secret. The methodology is versioned (v1.0) and published inside the application itself, on the "The science behind Mental Loadless" screen. You can read it, disagree with it, and know exactly what changed when it evolves. A measurement you cannot audit does not deserve the trust it asks for.

The computation runs server-side, as a weekly snapshot, and self-reported feelings are encrypted.

Alongside the overall MLI, a second indicator looks at two phases only: anticipate and monitor. Never execution.

It is the numerical answer to the most common sentence in couples counselling: "I carry all of it in my head, even when he — or she — is the one doing it." As long as execution is counted along with everything else, that experience is statistically invisible. Isolate it, and it becomes data.

Equity is not measured against a 50/50 ideal. The couple sets its own target — a deliberate 60/40 during a heavy period is a legitimate choice — and the app measures the gap to that target. A 70/30 chosen and held is a contract honoured, not a problem. A period of imbalance can even be declared explicitly, from two weeks to six months, with no "reason" field to fill in: during that time, banners and nudges go quiet. The app never guilt-trips a consented asymmetry.

Once a month, each partner estimates two things with two sliders: the share they believe they carry, and the share they believe the other carries. Those estimates are then set against the measurement.

• What I believe I carry
• What I am believed to carry
• What is measured

The gap between those three numbers is often more instructive than the measurement alone. It turns a disagreement ("but I do just as much as you") into an observable object, external to both people. It is no longer your word against theirs: it is three numbers on the same screen.

On top of that sits a trend signal: a 30-second weekly check-in (rumination, energy, satisfaction) and an eight-week curve. Below two data points, the app shows nothing rather than a "stable" that does not exist. Check-ins are encrypted.

Measuring how a load is divided inside a couple is a delicate act. Word choice is not decoration: it is part of the product, and it is enforced automatically across the app's copy, in all nine languages.

The banned words

• "Score", "ranking", "less than", "should" — anything that installs a grade, a winner and a party at fault
• All clinical vocabulary: burnout, depression, diagnosis, symptom. An app has no business putting a medical label on a person

The product's words

• Distribution and balance. A load is divided; it is not won.

The positioning fits in one sentence, and it structures every product decision: the app measures and organises; the professional interprets. We supply usable numbers and neutral language. What those numbers mean inside your relationship belongs to you — and, if you want, to a therapist or a mediator.

A factual one-to-two-page PDF, generated server-side, summarises the past month: MLI, equity against the target you set, cross-perception, check-ins, domain transfers. It is designed to be put on the table of a couples therapist or a family mediator, instead of the twenty minutes usually spent trying to summarise.

It carries an explicit disclaimer: the app measures, the professional interprets. It is produced from the same data as the monthly report shown on screen — so the PDF cannot tell a different story from the one you have already seen. It is available in nine languages, and it is strictly single-household: even in shared custody, it never contains the other parent's data.

Households that want to can contribute to an anonymous collective barometer: aggregated distributions, computed once a month, giving a picture of the mental load across families at the scale of the user base.

• Strict opt-in: two separate toggles (academic research / barometer). Absence of an answer counts as refusal — never consent by default. Revocable with one tap.
• K-anonymity: nothing is published below 10 households in a category. A statistic that would let you recognise one family is not a statistic, it is a leak.
• No individual data ever leaves.

The longer-term aim is an annual state of the mental load in families, published and methodologically defensible. It will only ship when the thresholds are genuinely met.

• Daminger, A. (2019). "The Cognitive Dimension of Household Labor". American Sociological Review, 84(4). — The phase cycle (anticipate, identify, decide, execute, monitor) our entire model rests on.
• Weeks, S. & Ruppanner, L. (2025). Work on the cognitive dimension of household labour and its effects on parental wellbeing.
• Dean, L., Churchill, B. & Ruppanner, L. (2022). "The mental load: building a deeper theoretical understanding of how cognitive and emotional labor overload women and mothers". Community, Work & Family, 25(1). — The distinction between cognitive and emotional labour.
• Mederer, H. J. (1993). "Division of Labor in Two-Earner Homes". Journal of Marriage and Family, 55(1). — The distinction between housework and its management, the origin of our negotiated standard.
• Haicault, M. (1984). "La gestion ordinaire de la vie en deux". Sociologie du travail, 26(3). — The founding French-language text, which named the mental load before the term spread.

These references are also cited inside the app, on the methodology screen.

Mental Loadless is available on the App Store and Google Play, in 9 languages, hosted in the European Union. The full methodology, with its weights and its version number, is readable in the app before you have any data at all.