Depth versus breadth: How content coverage in high school science courses relates to later success in college science coursework

8,310 undergraduates in introductory biology, chemistry or physics at 55 randomly chosen US four-year institutions (stratified random sample from a frame of 67; 34+ states). Fully controlled models: n = 6,999 (biology 2,760; chemistry 2,902; physics 1,337).

US undergraduates in a first college science course, reporting retrospectively on how their high school science course covered content.

FICSS survey again; three-level HLM (students within courses within institutions), grades on a 100-point scale. 'Depth' = at least one major topic studied for a month or longer (or recurring); 'breadth' = ALL major topics covered (8 key topics for biology and chemistry, 6 for physics). The two measures turned out uncorrelated — 41% of students had depth without breadth — which is the design's main strength. Controls: college year, parents' education, public vs other high school, ZIP-level community SES (median income, home value, educational attainment), last HS English grade, SAT-quantitative (ACT concorded), regular calculus, AP calculus, last HS maths grade, second year of the subject, last HS science grade. The authors test and reject a 'remedial depth' story and show teacher characteristics correlate only r = .04-.05 with depth/breadth. Still observational: students and schools chose the coverage pattern. Read in full text (author-posted PDF with all tables).

This is the strongest quantitative statement of the 'fewer topics, more deeply' rule in high school science, and it is a correlation. In the fully controlled models, DEPTH (at least one topic studied for a month or more) is worth +1.22 grade points in college biology (p=.0011), +0.89 in chemistry (p=.0251) and +1.54 in physics (p=.0061); BREADTH (covering all major topics) is worth −1.15 in biology (p=.0020), −0.74 in chemistry (p=.056, n.s.) and −0.31 in physics (n.s.). In effect-size terms these are TINY — 0.08-0.13 SD — which the paper reports honestly. Expressed against a natural yardstick they look larger: the depth benefit is 25-64% of the value of taking a whole extra year of the subject in high school (a second year is worth +2.57 in biology, +3.64 in chemistry, +2.39 in physics). Practical reading for a founder: the direction (depth over coverage) is consistent with the coherence literature and with Sadler & Tai 2001, but the magnitude is a fraction of a letter grade in one college course, from a self-report survey, and no causal design has ever tested it.

HIGH. Coverage pattern is not assigned — it reflects teacher choice, school resources, tracking and which students are in which class. The controls are the usual selection-on-observables set and include SAT-quantitative and prior grades, which are partly downstream of the same coursework (over-control) while omitting motivation, conscientiousness and parental push (under-control). The authors' own defence — that depth and breadth barely correlate with measured teacher characteristics — bounds one confound and leaves the student-selection one untouched. Consistent with the archive's default: a d ≈ 0.1 association in a non-randomized course-choice study is not a demonstrated lever.

Not independently replicated. It is a within-programme extension of Sadler & Tai (2001), where coverage carried a negative coefficient in the physics-only model, so the finding is internally consistent rather than externally confirmed. No experiment has manipulated depth vs breadth of high school science content and measured downstream attainment.