The short version · about five minutes
Mostly unglamorous, and mostly free.
The evidence is clearest about a handful of unglamorous things: keep the baby safe, fed and healthy; respond to her warmly and quickly (within a few seconds, not instantly); talk with her, face to face, quite a lot; and protect the parents’ own sleep and mental health, which turns out to be a good deal more than a footnote.
Most of what is sold as a brain-booster has weak or no evidence behind it. Meanwhile, the biggest effects in the whole literature come from removing a few specific harms rather than from adding enrichment. And children are sturdier than we fear - in an ordinary loving home, “good enough” is very close to the best that can be done, which I found reassuring to learn and suspect you will too.
The World Health Organization and UNICEF boil the whole field down to five things they call nurturing care (health, nutrition, safety, responsive caregiving and early learning), and I have sorted the cards on this site by those five.
Three ideas that organise everything else
If you remember nothing else.
- Removing harm does more than adding enrichment.
- The largest effects ever measured come from the absence of neglect, smoke, lead, alcohol and hitting, and from catching hearing loss early. Nothing on the enrichment side comes close. If you only have energy for one list, it is the short one of things not to do.
- RCTNelson et al. (2007)MetaLanphear et al. (2005)MetaGershoff et al. (2016)
- The mechanism, nearly every time, is live back-and-forth.
- Almost everything that works is some form of live, responsive exchange between the baby and a person. Products that try to replace it (videos, apps, flash cards) do not work; products that create an occasion for it (a book, a song) do.
- ExperimentKuhl et al. (2003)ObservationalRomeo et al. (2018)RCTHigh et al. (2000)
- Children are sturdier than we fear.
- In a home that is not neglectful, most enrichment beyond “good enough” has a small effect, and the strongest trials studied babies at the low end, where the baseline was low to begin with. I do not read that as a reason to do less, so much as a reason to worry less.
- ReviewScarr (1992)ReviewPlomin et al. (2015)MetaWitte et al. (2000)
The five domains of nurturing care
Where each card lives.
- 01Good health
- Skin-to-skin, the well-child schedule, hearing and milestone checks, vaccines on time.
- Skin to skinVisits, hearing test, milestonesTummy time
- 02Adequate nutrition
- Breastfeeding where possible, vitamin D, iron from six months, allergens early.
- Breastfeeding, and the IQ claimsVitamin D, iron, allergens
- 03Safety and security
- Safe sleep, and keeping smoke, alcohol and lead away from her.
- Safe sleepSmoke, alcohol and lead
- 04Responsive caregiving
- Notice and answer her cues. Treat a parent’s depression. Do not hit.
- Notice and answer her cuesA parent's depressionDo not hit
- 05Early learning
- Talk with her, read to her, and let live people rather than screens do the teaching.
- Talk with the babyRead aloud
How to read a card
The moon shows how firm the evidence is.
- Strong
- Several randomised trials, or a large and consistent body of evidence, and the professional bodies agree.
- Good
- Solid evidence with modest effects, or strong evidence in one specific group of babies.
- Mixed
- A real signal, but observational, small, or a good deal weaker than the popular version of the claim.
- Weak
- Little or no evidence of benefit. Usually harmless.
- Avoid
- Evidence of harm, or a safety risk with no benefit to weigh against it.
Sources carry their type
A trial, a survey and a guideline are not the same kind of evidence.
- RCT
- Randomised trial. People (here, babies or families) were assigned to groups by chance, so differences are most likely caused by the thing being tested.
- Meta
- Pooled analysis. A systematic review or meta-analysis: many studies combined, with the combined effect size reported.
- Observational
- Observational study. Families were watched, not assigned. Good for spotting patterns, weaker for proving cause, because families who do X differ in many other ways too.
- Experiment
- Lab experiment. A controlled study of a specific mechanism, usually small and short.
- Guideline
- Professional guideline. A professional body’s formal recommendation, written by a committee that reviewed the evidence.
Every claim on the site links to its source, and every source carries a one-line account of what it actually showed. Here is how I would weigh them.
Next
Where to go from here.
-
12 cards
What works
The things with several trials, or a large and consistent body of evidence, behind them. Each opens into why, the numbers, and what it looks like at 3 a.m.
Open
-
8 cards · 4 to avoid
Mixed, oversold, or harmful
Real signals the popular version has stretched, pleasant things that do a bit less than promised, and four product categories with documented harm.
Open
-
6 patterns · 3 caveats
What carries forward
Which early patterns keep paying off through school and adolescence, and three caveats that should lower your anxiety rather than raise it.
Open