In 2011, Betsy Sparrow and colleagues at Columbia University conducted a series of experiments known as the "Google Effect" study. They demonstrated that when people know information can be easily retrieved online, they are less likely to encode it into long-term memory. The availability of external access reduces the effort of internalisation — the brain, being pragmatic, outsources storage to the device.
The phenomenon is not new: we have used lists, notebooks and books as external memory for centuries. What has changed is the scale and speed of access. The implication for educational technology is more specific than "less screen time is better": the problem is not the digital tool itself, but whether it is being used in a way that demands cognitive effort — or in a way that replaces it.
Scaffolding versus substitution
Vygotsky described the Zone of Proximal Development — the range between what a learner can do alone and what they can do with support. Well-designed educational tools operate in this zone: they provide enough support to advance beyond what the learner could do alone, but demand genuine effort for that advance to occur. This temporary support — scaffolding — is designed to be withdrawn as the skill consolidates.
In Hangman, the player must generate every letter independently — there is no suggestion list, no automatic hint. This activates the "generation effect" (Slamecka & Graf, 1978): information that the learner produces through effort is retained more durably than information simply presented. The game works precisely because it forces generation, not passive recognition.
Three signs a tool is creating dependency
- The child seeks the hint before trying: if the default behaviour is to ask for help before any independent attempt, the tool is being used as a shortcut. Educational Quiz has no hint button — the structure requires the player to reason before seeing any feedback.
- The skill does not transfer to offline contexts: if a child solves subtraction problems easily in Math Adventure but cannot do the same on paper, consolidation was superficial. Genuine learning transfers across contexts.
- Zero tolerance for screen-free activities: inability to read a printed text or play a board game without intense resistance may indicate that digital stimulation has calibrated reward expectations that analogue activities cannot match at the same speed.
What distinguishes tools that build autonomy
Generation before presentation: in Crossword, each word must be generated from a definition — there is no letter bank to choose from. In Hangman, each letter is retrieved from memory, without a limited set displayed on screen.
Error with consequence: wrong letters in Hangman bring the game closer to over. This cost inhibits random guessing and incentivises real vocabulary retrieval. Without penalty for error, the rational behaviour is to try every option until guessing correctly — which consolidates nothing.
Progression that requires prior consolidation: Enchanted Maze and Chess are examples where advancing genuinely requires the previous skill to have been internalised — the difficulty cannot be bypassed without the necessary cognitive resources.
Technology as a bridge, not a destination
The most productive use of digital educational tools is as a bridge between concepts and competencies. A well-structured cycle: exposure to the concept (in class or reading), practice with immediate game feedback, review of errors without a screen (writing, speaking, explaining), then a new round at higher difficulty. The screen appears at two points in this cycle — it does not replace the other two.
Used this way, games like Math Adventure and Educational Quiz function as diagnostic generators: they reveal exactly where knowledge is still unstable, without the emotional weight of a formal test. The output of the game is not just the score — it is information about which concepts need offline reinforcement.