Our philosophy

A century of memory and literacy research says a lot about how spelling is learned, and most of it is ignored by most spelling products. We read the research and let it constrain us. That includes respecting its gradations: a meta-analysis outweighs a single study, a single study outweighs expert consensus, and a company's claims about its own product (ours included) are not evidence at all. When the literature is unsettled, we treat our choices as bets rather than dressing them up as science. The Learn section lays out the research we work from, with references, so you can check our homework.

The best-replicated result in learning science is that producing an answer from memory strengthens it far more than looking at it again (the evidence is laid out in what makes spelling practice work). So everything here is built around production: you hear a word and you spell it, from nothing, with no autocorrect and no multiple choice. Recognition feels like knowledge and mostly isn't. If a practice activity lets you succeed without retrieving, we regard it as decoration.

Autocorrect made everyday writing painless by making our actual spelling invisible, to ourselves most of all. We think the kinder thing is the truth: an honest grade on what you produced, the correct form shown against your attempt, and a score that moves in both directions. Errors are information: getting a word wrong and then seeing the correction beats never getting it wrong at all, which is the productive-error evidence in the practice guide. A tool that shields you from your mistakes is shielding you from the mechanism by which spelling improves.

Learning research keeps finding that the conditions which feel hardest (being tested, waiting long enough to half-forget, getting things wrong and repairing them) are the conditions that produce durable learning (see the desirable-difficulties and productive-error sections of the practice guide). So we aim practice at the edge of your ability and let you feel the stretch, rather than serving easy wins that make a dashboard look good. Struggling with a word you eventually beat is not a product failure; it is the product working.

The motivation research is two-sided: game elements can support learning, and rewards can also crowd out the interest they were meant to amplify (both halves are cited in spelling and digital learning). We take both halves seriously. We would rather you come back because you feel yourself getting better than because leaving feels like losing something. That means celebration is kept sparse enough to mean something, feedback describes what you did rather than what you are, and we do not cite industry retention folklore as if it were learning evidence, in our marketing or in our own decisions.

Spaced, short, regular practice beats marathon sessions (the spacing evidence is in the practice guide), which aligns nicely with respecting your day. A session should end cleanly, soon, and on purpose, leaving you somewhere worth returning to tomorrow. We measure ourselves on whether the minutes you give us convert into spelling you keep, not on how many minutes we can extract.

Your SpellingIQ is a game rating on spelling ability. It is not a clinical instrument, an IQ test, or a diagnosis of anything. SpellingIQ is also not a substitute for structured literacy instruction: a learner with a suspected reading or spelling disability deserves a qualified professional, and the research on that is a different body of work from the practice science we build on. Most of the learning evidence comes from classrooms and laboratories rather than from voluntary apps, and much of it was measured on children. Where a finding's reach is uncertain, we say so rather than rounding up.

The cheapest way to evaluate a philosophy is to watch it operate: play today's Daily Five, take the placement test, or read about SpellingIQ.