A word that covers many things
“Revolutionizing” is a strong word, and in education it is often attached to ideas that turn out to be gradual. Personalized learning is one of the most common of these ideas. The basic notion is appealing: instead of everyone following the same path at the same speed, each learner gets materials, pace and support suited to them. Open education is sometimes presented as a natural home for this, because open resources can be mixed, adapted and reused freely.
This article looks at what personalization can mean in practice and where the claims run ahead of the evidence. It does not review any specific product.
Three different things called personalization
It helps to separate three ideas that are frequently blended.
Choice. The learner decides what to study, in what order and with which resources. This needs no special technology, only a good range of materials and some guidance. Open licenses make a wide range of materials easier to assemble.
Adaptation. A system or a teacher changes what is offered based on how the learner is doing. This might mean extra practice on a weak topic, or skipping content that the learner has already mastered. Adaptation can be done by a human, by a simple rule in a course (if a quiz score is under a threshold, review the unit), or by more complex adaptive algorithms.
Tailoring to the person. Content is shaped by interests, background, language or goals. A mathematics example about football for one learner and about cooking for another is a mild form of this.
Each has different costs and different evidence, and a product that claims to be “personalized” may offer only the first, which is simply a menu.
What open education adds
Openness offers practical assets for personalization.
- Materials can be translated and localized, so a learner gets content in their own language and context.
- Teachers can assemble a unit from several resources to suit a group or an individual.
- Learners can follow a self-directed path, choosing materials across sources.
- Resources with open licenses can be fed into adaptive tools without separate permission, within the license terms.
These are real advantages, but they are enabling conditions, not outcomes. The existence of flexible materials does not ensure that anyone is choosing or adapting them well.
Where the promise outruns the evidence
Some cautions are worth keeping in mind.
Learning styles are not a sound basis. A popular version of personalization tailors lessons to supposed learning styles, such as visual or auditory learners, an idea often discussed as the learning styles myth. Reviews of the research have generally not found support for the idea that matching teaching to a preferred style improves learning. A better foundation is matching the method to the content and to what the learner already knows.
Learners are not always good judges of what they need. People tend to choose what feels easy or familiar. Fully learner-directed paths can leave gaps, especially for beginners. Some guidance usually helps, and a balance between student autonomy and teacher input is common.
Adaptive systems are only as good as their design. An algorithm that adjusts difficulty depends on the quality of questions, the data it has and the assumptions built into it. Studies of adaptive tools show mixed results that depend on the subject, the age of the learners and how the tool is used.
Teacher time still matters. Several studies and reports on technology-based approaches suggest that human contact and feedback often contribute a great deal. Technology that replaces a teacher entirely, including some forms of AI tutoring, is not well supported by the evidence.
Data and privacy come with the territory. Personalization usually requires collecting information about learners, as does most learning analytics. Learners, parents and schools should ask what is collected, who can see it and how long it is kept, and check privacy terms and local rules.
Results are hard to measure. Many claims of improvement come from the companies selling the tools or from small trials. Good evidence typically involves comparison groups and independent evaluation, and it is not always available.
Questions to ask about any “personalized” offer
- What exactly is personalized: the content, the pace, the order, the support?
- On what information does it base its adjustments?
- Is there independent evidence, and does it involve learners like mine?
- What does a teacher or tutor do alongside it?
- What data does it collect, and who controls it?
- Can I take my materials and records with me if I stop using it?
What tends to work in practice
Modest forms of personalization are often the most dependable. Letting learners skip content they know, offering extra practice where it is needed, giving timely feedback, and offering several ways to explain a hard idea are all reasonable. Pairing these with regular check-ins from a teacher or mentor tends to be more useful than relying on automation alone.
A measured outlook
Personalized pathways are a sensible aim, and open resources help make them possible. They are not a revolution that has already happened, and they will not suit every learner or subject. The best attitude is interested skepticism: try things on a small scale, check whether they help, and avoid choosing tools on the strength of promises alone.