We would rather understate a finding than overstate it. Every summary in the library says what kind of study it is, how strong the evidence is, and what the study can't tell us.
Most research on screens and children is observational. That means it can show a link between two things but not prove that one causes the other. We say so plainly, entry by entry.
Study types
They are listed in no ranked order. A small study can be carefully designed, and a very large one can still only show a link.
Pooling studies shows whether a finding holds up across different samples and settings, and how large the effect tends to be.
Typical limits
Only as good as the studies it pools. It can blend very different ages, settings or ways of measuring screen use.
Reviewers search for all the relevant research, not just the studies that support one view, and rate each one.
Typical limits
Many reviews count results instead of pooling them, so they can say which way the evidence leans but not how big the effect is.
Because chance decides who gets what, differences in outcomes can be traced to the thing being tested.
Typical limits
Trials are often short, and many use college students, so results may not carry over to young children.
These designs get closer to cause and effect than observation alone, using real-world events in place of random assignment.
Typical limits
Groups can still differ in ways the researchers did not measure.
Tracking children for months or years shows what came before what, which a single snapshot cannot.
Typical limits
It is still observational. Screen time is often reported by parents, and other factors in family life may explain the link.
Useful for spotting patterns across large groups quickly.
Typical limits
It finds associations, not causes, and cannot tell which came first.
Lab studies can isolate one factor, such as handwriting versus typing, and measure it closely.
Typical limits
Samples are usually small and the setting is artificial. We treat these as early signals worth watching.
Running the comparison where children actually learn makes the result more practical.
Typical limits
Results are tied to the setting, the age group and the tools used.
These reports weigh large bodies of evidence and often shape how schools and governments act.
Typical limits
They are not new research, and many rely on correlational data.
Qualitative research shows how and why something happens, in people’s own words and experiences.
Typical limits
It describes, rather than measures. Findings may not apply widely.
Evidence strength
Each entry carries one of six strength labels and a sentence explaining it. The label weighs the design, the size of the study and who was studied. A strong trial in college students is still a trial in college students, and the entry says so.
How we choose studies
Every entry is checked against the original journal page or publisher. Policy reports come from major public bodies, and we label them as reports.
We prioritize the questions parents and school boards here are asking: early language, reading, handwriting and devices in the classroom.
Studies where technology helped are in the library too, such as adaptive math software with small-group teaching and well-run 1:1 laptop programs.
Correlation and causation
Say a study finds that children who use screens more in the evening sleep less. There are at least three ways that could happen.
Evening screen use pushes bedtime later or makes it harder to fall asleep.
Children who already struggle to fall asleep reach for a screen while they wait.
A busy family schedule, for example, could mean both later bedtimes and more screen time.
Studies that follow the same children over time, or that assign groups at random, help sort these apart. When a study can't, our summary says “linked to,” not “causes.”
Corrections
Each entry shows the date it was last reviewed. When a summary changes, the date changes with it.