This file reads the literature on the apparatus of an explanatory study text, strand by strand: existing reviews, the seminal works, one strand per sub-question, then practice and Q/A sites. Levels: MA meta-analysis, SR systematic review, R review without pooling, E experiment, C correlational or before-and-after study, D descriptive count, A argument. "(secondary)" marks a number read in an abstract, not in the paper. Grey quality scores are in sources; a source under 10 of 20 is named "example only".
0. Existing reviews
| Hartley 1976 |
pretests, objectives, overviews, advance organizers |
turns the research on each into procedural guidelines (abstract) |
R |
| Melton 1978; Hamilton 1985 |
behavioural objectives; adjunct questions and objectives |
effects depend on awareness, clarity, difficulty, number and position of objectives (Melton, abstract) |
R |
| Tippett 2010 |
refutation text, two decades |
refutation text is more likely than expository text to produce conceptual change; no developmental trend (abstract) |
R |
| Guzzetti 1993 |
conceptual change, reading and science education |
effective interventions share one element: they produce conceptual conflict (abstract; no number read) |
MA |
| Schroeder 2022 |
refutation text, between-subjects |
g = 0.41 over 44 comparisons, n = 3,869, consistent and robust across contexts; against expository text alone g = 0.36, k = 30 (Table 3; read in full, PMC8784251) |
MA |
| Danielson 2025 |
refutation text, preregistered |
71 articles, 111 samples, 294 effect sizes; a consistent advantage, no moderator among 26 (secondary) |
MA |
| Zengilowski 2021 |
refutation text, critical |
single topics, the role of testing and durability beyond an immediate posttest are unresolved (abstract) |
R |
| Konrad 2009; Larwin 2012; Larwin 2013; Biggers 2020 |
guided notes |
positive, larger in school-age samples; numbers in strand 6 |
MA; SR |
| Wittwer 2010 |
instructional explanations in worked examples |
21 experiments; minimal benefit overall (abstract) |
MA |
| Yanagisawa 2020; Abraham 2008 |
glosses in second-language reading |
glossed reading teaches more words; numbers in strand 4 |
MA |
| Clinton 2019 |
open against commercial textbooks |
no difference in learning, g = 0.01 (k = 22, N = 100,012), I2 = 98.17 (read) |
MA |
- Located by title and venue only: Aleknavičiūtė 2023 and Pacaci 2024 (conceptual change interventions); Armagan 2017 (conceptual change texts) is in strand 5.
- This search found no systematic review of textbook aids as a set, of study guides or of crib sheets. A 2006 chapter on text-based pedagogical features exists, but its authors could not be verified, so it is not cited.
1. Seminal works
- Hartley 1978, a book read through its ERIC record (grey tier 1): guidelines for writers and typographers, with texts shown before and after revision. Hartley 1977 gives fifty guidelines (abstract) in three groups (prose, tables and graphs, typography).
- Mager 1962 (grey tier 1, ERIC record): a programmed text that teaches to state objectives as observable behaviour, with conditions and a criterion of success.
- Biggs 1996: constructive alignment, objectives, teaching and assessment built on the same verbs, illustrated with one teacher-education unit (abstract).
- Cohen 1987: instructional alignment is claimed to give larger effects at lower cost than other constructs (abstract; the effect sizes were not read).
- Krathwohl 2002 (title and venue only) presents the revised taxonomy of objectives; Wiggins 2005 names backward design (book not opened; example only).
2. Pedagogical aids: prevalence, use, effect (Q1)
| Marek 1999 |
counts aids in introductory psychology texts of 1995-1997; students rate 15 aids, extending Weiten 1996 (abstract) |
D |
| Weiten 1999 |
university, college and high-school students agree: bold terms, glossaries, chapter summaries and self-tests are most useful (abstract) |
D |
| Gurung 2003 |
more than 200 students: bold terms and practice questions used most; reported use and helpfulness do not relate to exam scores; rated helpfulness of key terms relates to worse scores (secondary) |
C |
| Gurung 2004 |
240 students: use of aids not positively related to any exam; a poor first exam raised summary use only a little (secondary) |
C |
| Gurung 2011; Landrum 2012 |
perceived quality of aids, visuals and writing predicted how much was read and exam scores, varying by book; the link to grades is indirect (secondary) |
C |
| Gurung 2012 |
textbook use against quiz scores across 35 institutions: books differed in rated quality, not in quiz scores; grade point average and rated helpfulness predicted scores (abstract) |
C |
| Benbassat 2008 |
230 respondents say boxed text is rarely studied; in a test with 177 readers, boxed and plain text were skimmed alike (secondary) |
D; E |
- No experiment was found that removes one aid from a real textbook and measures learning.
3. Objectives and alignment (Q2)
| Klauer 1984 (core) |
objectives raise learning of targeted content and lower the rest |
MA |
| Duchastel 1979 |
the prominence of ideas in the text shaped learning, until relevant objectives were given (abstract) |
E |
| Britton 1985 |
specific objectives lengthened reading of relevant passages and raised their recall, against general or none (abstract) |
E |
| Kaplan 1974 |
objectives given part by part with the text gave more intentional learning than all at once (abstract) |
E |
| Stokes 2010 |
objectives of 16 business textbooks use low-level verbs that differ from what their end-of-chapter material asks (abstract) |
D |
| Larkin 2013 |
grades and satisfaction rose after one programme was aligned; the authors note other causes (abstract) |
C |
| Hailikari 2022 |
students say teaching and assessment shaped their approach; the stated outcomes "did not seem" to matter much (abstract) |
C |
| Trigwell 2014; Wang 2013 |
teachers experience alignment in qualitatively different ways (abstract); Wang 2013 located only |
C |
| Polikoff 2014 |
teachers' alignment to standards had surprisingly weak links to their value-added scores (abstract) |
C |
| Lynch 2020; Hollowell 2017; Sadaf 2019 |
rubric-based redesign: a pilot of two against two courses, all indicators better, none significant (Lynch, abstract); descriptive and perception studies |
C; D |
4. Summaries, key terms and glossaries (Q3)
| Hartley 1979; Cook 1981 |
an end summary gave better recall than one at the start of the same page, or none; in Cook, 40 students, an opening summary on a separate page did as well (abstracts) |
E |
| Mayer 1996 |
three experiments, 163 students: a summary of captions and simple drawings matched or beat the full text on recall and transfer (abstract) |
E |
| Reder 1980 |
studying only a summary beat reading the whole chapter on true-or-false statements of main points and inferences, 75.2% against 65.3% (Experiments 1 and 2, N = 63), at 20 minutes and at a week. The gap closed when the text had three times the study time, and was not significant after 6 to 12 months; a summary studied first did not hurt learning related new material (read in full, pp. 125-132) |
E |
| van der Meij 2012 |
numbered summaries inside the text against a structured abstract: recall d = 1.57, 40 students (secondary) |
E |
| Yanagisawa 2020 |
42 studies, N = 3,802: glossed reading taught 45.3% of words at once and 33.4% later, against 26.6% and 19.8%; in-text glosses and glossaries were the least effective forms (secondary) |
MA |
| Abraham 2008 |
11 studies: computer glosses had a medium effect on comprehension and a large one on incidental vocabulary (secondary) |
MA |
| Stewart 1991 |
marginal glosses in first-language college reading may be of limited use (abstract) |
E |
- All gloss meta-analyses are in second-language reading, and no study of an end glossary in a first-language study text was found; defining terms before use is measured in the core (pre-training, Mayer 2002).
5. Refutation text (Q4)
| Guzzetti 2000 |
a refutation text alone is not enough; discussion helps; only refutation text showed long-term effects (abstract) |
R |
| Diakidoy 2011 |
against expository text, refutation raised inferences and delayed learning, not memory for the text (abstract) |
E |
| Jin 2025b |
Solomon four-group design, 113 participants recruited online: refutation beat expository text, and a pretest did not inflate the effect (abstract) |
E |
| Jin 2024 |
refutation helped; seductive pictures added to it hurt (abstract) |
E |
| Armagan 2017 |
42 conceptual change text studies, 1995-2010: overall effect 1.18 (secondary) |
MA |
6. Guided notes, study guides, crib sheets (Q5)
| Larwin 2013 |
12 studies, 27 effects, 1,529 students: d = 0.55 (CI 0.34 to 0.75); 0.78 against students' own notes, 0.27 against the instructor's full notes (read, pp. 51-52) |
MA |
| Larwin 2012 |
school and special education, 11 studies, 412 students: d = 1.13, effects from -2.21 to 10.39; on exam scores d = -0.75 (read, Results) |
MA |
| Konrad 2009 |
guided notes raise note accuracy and performance, most for school-age students (abstract); its college part held three studies (Larwin 2013) |
MA |
| Biggers 2020 |
22 adult studies: positive perceptions; gains in some domains, most with complex content; delivery mode made no difference (abstract) |
SR |
| Dickson 2005b |
a required textbook study guide raised multiple-choice scores; completing three quarters did no better than a quarter (abstract) |
E |
| Horner 2007; Cushen 2019 |
instructor-made guides did no better than students' own (Horner); a provided concept list lowered exam scores (Cushen) (abstracts) |
E |
| Saleh 2026; Lieu 2017 |
required guides beat optional ones (169 against 86 students); completing reading guides correlated with exam scores (abstracts) |
C |
| Dickson 2008; Funk 2011; Burns 2014 |
making a crib sheet did not teach; expecting to use one fostered dependency; use rose among weaker students (abstracts) |
E; C |
| Hamouda 2016 |
crib-sheet content tracked scores only on comprehension-level questions (abstract) |
C |
- Duchastel 1983 models the ideal study guide by its functions, with no test (abstract).
7. Explained solutions (Q7)
- Wittwer 2008 (R): instructional explanations often fail to add learning; it gives design guidelines (abstract).
- Wittwer 2010 (MA, 21 experiments): explanations help conceptual more than procedural knowledge and do no better than self-explaining (abstract).
- Richey 2013 (E, three experiments): withholding stepwise explanations gave more conceptual learning; Gerjets 2006 found explanations ineffective (abstracts).
- Mendiolea 2013 (D, six students): learners valued detail, explanation and choice in worked solutions and skipped extension tasks (abstract).
- Access to solutions and its effect on homework is reviewed in the exercise-sets review (Rystrom 2005; MathEd 2021a; MathEd 2022b).
8. Practice
| OpenStax 2026b |
objectives per section, bold terms defined in context, call-outs, section summaries, checks before exercises, answer key at the back |
open textbook; adoption in state-of-the-art.md |
| Carpentries 2026a |
episode with required questions, objectives and key points; challenges with nested solutions |
45 open lessons (Carpentries 2025) |
| Carpentries 2026b; Wilson 2019 |
outcomes, then assessments, then content; 3 to 5 episodes per hour (Wilson) |
training curriculum; book |
| Quality Matters 2026b |
8 general and 44 specific standards; objectives aligned with assessment, materials and activities; 85% to certify |
rubric; adoption is the owner's claim |
| SUNY 2026 (example only) |
50 standards in an open rubric, licensed CC BY 4.0 |
no count given |
| CMU 2025 |
objectives with action verbs, aligned with assessment and teaching; no research cited |
university teaching centre |
- Revision guides: the publisher's page refused access and no study of the genre was found (G3); Q/A threads name solved-problem series without data.
Q/A sites (Academia, Mathematics Educators, Computer Science Educators; 14 threads, 4 under 10 of 20; strings in protocol). Practitioner opinion, no data.
- Notes handed out: some students disengage, others check their notes (MathEd 2014h); one lecturer says provided notes rarely raise exam results (Academia 2017c).
- Writing one's own notes: to choose and omit what a long textbook holds (MathEd 2014i; Academia 2014a).
- Reading the exposition: students use exercises and answers but not the explanation; guiding questions and self-explanation training are offered (MathEd 2014k).
- Misconceptions: analyse the learner's wrong rule in front of the class (MathEd 2017b).
- Crib sheets: making one is said to focus study, which the experiments of strand 6 do not support; a handwritten card stops copying (MathEd 2017c; Academia 2018b). Study guides are best written by the learners (CSEd 2018c).
- Objectives: SMART criteria and Bloom's verbs (Academia 2017b). Glossary place and marking (Academia 2018c; Academia 2022a) and chapter length (MathEd 2014l), slides against notes (MathEd 2014j): example only.