Reading research about young people without being misled

Research about young people is quoted more often than it is read. A journal article reports a careful, narrow finding; a summary turns it into a claim about a generation; and the claim then circulates for years with the qualifications stripped off. Anyone who works with young people — a teacher, a youth worker, a policy officer, a journalist — needs a way to read the original and judge what it actually supports. The method is not statistical training. It is a short sequence of questions, always asked in the same order.

Start with who was studied

The sample decides the reach of every conclusion. Three things matter: how many, who, and how they were found. A study of a few dozen young people in one programme can describe that programme well and cannot describe a country. A study recruited through schools misses those who are not in school, which in youth research is precisely the group most policy is aimed at.

Age bands matter more here than in adult research, because a two-year gap in adolescence is a large developmental distance. A finding about fifteen-year-olds may not hold for eighteen-year-olds, and papers that pool a wide age range often bury this in a table rather than in the discussion.

Then ask who is missing. Young people in care, in custody, in remote areas, out of school, or not fluent in the language of the questionnaire are systematically under-represented, and their absence usually pushes the results in a predictable direction.

What the method can and cannot show

Most youth research is self-report: a questionnaire or an interview. That is not a weakness, but it has known behaviour. Answers about sensitive conduct move depending on who is present, whether a teacher administered the form, and whether anonymity was explained credibly. Questions about frequency invite rounding. Questions about attitudes capture what a respondent is willing to state, which is itself useful data if it is not read as behaviour.

The second distinction is design. A cross-sectional study photographs one moment and can show that two things occur together. A longitudinal study follows the same people and can begin to show order in time. Only a design with a comparison group and some form of allocation can support a claim that an intervention caused a change. When a paper's own method section says association and the abstract says effect, believe the method section.

Qualitative work should be read with a different question. Its value is not representativeness; it is the ability to show how something works and what words young people themselves use. Judge it on whether the analysis is traceable back to the material, and on whether alternative readings are considered rather than by the number of participants.

Where distortion enters

Four places, reliably. The abstract, which is written to be quoted and is the least hedged part of the paper. The press release, which adds a population claim the paper never made. The single number, lifted out of a table where its confidence interval lived. And the age of the data, because fieldwork often precedes publication by a long stretch, and a finding about how young people communicate can be describing a technology landscape that no longer exists.

The last one is worth a habit: before citing, check when the data were collected rather than when the article appeared. It is normally stated in one sentence in the methods.

A short checklist

Who was in the sample and how were they recruited. Who is missing. Was it self-report, and under what conditions. Cross-sectional, longitudinal or controlled. Does the paper claim association or cause. When were the data actually collected. Did the authors state limitations, and do those limitations touch the finding you want to use.

Seven questions, a few minutes each. They will not make anyone a methodologist, but they are enough to stop the most common error in this field, which is not bad research. It is a modest, well-conducted study being asked to carry a claim about an entire generation that it was never built to support.

FAQ

Impacts of the loss of ACYS on youth services and young people