Conflicting Social Media Studies

Compare the studies on design before comparing their headlines. Study A is larger in reported effect but smaller in sample, urban, and apparently correlational: it links heavy use with self-reported anxiety but cannot show which came first. Study B is larger and covers more provinces, and it controls for sleep and exam pressure-plausible confounders-but its telecom funding is a reason to ask about independence, not to dismiss it automatically. Frame the evidence as mixed and conditional: ‘Heavy use is associated with anxiety in one urban sample, but the association weakens when sleep and academic pressure are accounted for in a larger national sample.’ Your personal month-long experiment is an anecdote, not proof; use it at most to illustrate why the question matters. This framing satisfies the rubric because it neither ignores the conflict nor pretends the studies settle causation.

Thesis: ‘Social media is not uniformly harmful to Nepali youth; its effects appear to depend on dose, sleep, and offline pressures, so policy should target harmful use patterns rather than ban platforms.’ Counter-paragraph: ‘The strongest case for “more harm than good” is that heavy use is associated with anxiety among Kathmandu students in Study A. However, Study B's larger, controlled analysis suggests that much of that association fades once sleep duration and exam pressure are taken into account, so the correlation is not proof that social media causes anxiety.’ This uses both studies without overclaiming. It does not say Study B “disproves” Study A; it says the association is conditional. It also does not hide the funding issue. If you mention the telecom funder, do so as a limit on confidence, not as a conspiracy: ‘Industry funding warrants scrutiny of the design, but it does not by itself make the data false.’

Avoid: ‘Social media causes anxiety in Nepali youth.’ Avoid also: ‘Controlling for sleep proves social media is harmless.’ Both exceed the evidence. A cautious rewrite of the first is: ‘Among Kathmandu Valley students, self-reported heavy social-media use is associated with higher anxiety, but the survey cannot show whether anxiety drives use, whether use drives anxiety, or whether both reflect a third factor such as exam pressure.’ This keeps the direction of the finding but removes the causal verb. It also names the alternative explanations. If you want a policy implication, phrase it as a hypothesis to test: ‘A school-based trial that reduces evening use and measures sleep and anxiety would be needed before claiming a causal effect.’ This is the kind of sentence the rubric rewards because it shows what the data can and cannot support.