Civic Education: What Makes Students Learn
Niemi RG, Junn J · 1998
grade Dlongitudinalindependentmixed
Sample
NAEP Civics assessment data (cross-sectional)
Population
US high-school students.
Design
Exhaustive cross-sectional analysis of NAEP Civics data. No causal design: course-taking is self-selected and only a cognitive outcome is available in NAEP. Historically important as the paper that made civic-education research respectable again after Langton and Jennings.
Key findings
Taking a civics course in high school is associated with roughly a 4 percentage point gain on the NAEP civics assessment - about one third of a letter grade. Read as vindication by proponents and as confirmation of civics' weakness by critics (Hart and Youniss 2018). Only a knowledge outcome; NAEP contains no behavioural or attitudinal measure.
Genetic confound
Not assessed in detail; see exclusion note.
Replication notes
Reversed the Langton and Jennings 1968 null on knowledge, and its central claim - that civics coursework raises civic knowledge - is supported by the experimental evidence (Green et al. 2011) and by Campbell and Niemi 2016. Its magnitude (about 4 NAEP percentage points) is small enough that critics read it as confirming civics instruction's weakness.
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| NAEP Civics assessment score | cross-sectional association with civics coursework | about +4 percentage points, ~one third of a letter grade | standardized (NAEP Civics) | high school | none | end-of-treatment | domain-skill |