Library / Glossary / The context glossary
Reading education research glossary
39 terms from C21, Reading education research — each defined in the primer's own words. Every term the primer teaches links to the slide that teaches it.
C21 · Evidence and change
Reading education research
“Evidence-based means a randomised trial says so.”
Every term below is defined in the words of reading education research, primer C21 of understanding the context of education, and opens the primer at the slide where it is taught. 3 terms are also defined by another primer in the series; where the two differ, both wordings are given. The whole context glossary holds all of them together.
| Term | Definition | Referred to in | Read further |
|---|---|---|---|
| A | |||
| Action research | Research carried out by people on their own practice, in cycles of planning, acting, observing and revising. The lineage runs from Lewin (1946); the aim is a better practice as much as a publishable finding. |
| |
| C | |||
| Case study | An in-depth study of one setting, programme or person, usually with several kinds of data. Flyvbjerg (2006) argues that the common objection, that one cannot generalise from a single case, misunderstands what exemplars do for a field. | ||
| Correlational and longitudinal studies | Studies that measure how variables move together, and studies that follow the same people over time. Longitudinal data can establish that one thing preceded another, which is necessary for a causal claim but not sufficient. |
| |
| Cost-effectiveness | The effect achieved per unit of resource. Harris (2009) argues that effect-size benchmarks that ignore cost cannot guide administration or policy. An economic evaluation that omits staff time is incomplete. | ||
| Counterfactual | What would have happened to the same students without the programme. That world cannot be observed, so every causal design is a way of building a stand-in for it. C22 What would have happened to the same people without the programme. It cannot be observed, so it is estimated from a comparison or control group. Primer C21 covers the designs used to do this. |
| |
| Critical theory | A family of traditions that treat research as part of exposing and changing unequal arrangements. Its characteristic question is whose interests a practice serves, and its designs are often participatory. |
| |
| D | |||
| Design-based research | Iterative building and testing of an educational design in a real setting, producing both a working design and a theory about why it works (Design-Based Research Collective, 2003). Anderson and Shattuck (2012) review its first decade. | ||
| Difference-in-differences | A quasi-experimental design comparing the change over time in a treated group with the change in an untreated one. It rests on the assumption that both groups would otherwise have moved in parallel. |
| |
| E | |||
| Effect size | A difference between groups expressed in standard deviations, so that results on different scales can be compared. Its arithmetic, including confidence intervals and statistical power, is in the craft series’ Basic statistics guide (18). |
| |
| Efficacy and effectiveness | Efficacy is whether a programme can work under favourable conditions, usually with its developers close by. Effectiveness is whether it does work at normal scale in ordinary hands. Effects are usually smaller in the second. | ||
| ESSA tiers of evidence | Four levels in US guidance: strong evidence needs a well-designed experimental study, moderate a quasi-experimental one, promising a correlational study with controls, and the fourth only a well-specified logic model (US Department of Education, 2016). |
| |
| Ethnography | Extended, close observation of a setting, aiming to describe it as its participants experience it. Heath (1983) is a classic educational example, following children’s language across home and school. |
| |
| External and ecological validity | How far a finding holds beyond the sample studied, and how far the study conditions resemble ordinary life. A trial can have high internal validity and little of either. |
| |
| G | |||
| Grey literature | Reports, theses, evaluations and working papers not published in journals. It matters because published studies report larger effects than unpublished ones (Polanin et al., 2016), so a review that skips it is biased upward. |
| |
| H | |||
| Hawthorne effect | The idea that being studied changes behaviour. McCambridge et al. (2014) reviewed the evidence and concluded that new concepts are needed to study research participation effects, so it is better used as a caution than as a quantity. |
| |
| Heterogeneity | Real variation in effects across studies, beyond what sampling alone would produce. Where it is large, the average of a meta-analysis describes no actual programme, and the spread is the finding. |
| |
| I | |||
| Implementation fidelity | The extent to which a programme was delivered as designed. Durlak and DuPre (2008) found that the level of implementation strongly affects outcomes, which is why a null result can mean the programme was never run. |
| |
| Interpretivism | The position that social reality is constituted by the meanings people make, so research should recover those meanings rather than reduce them to variables. |
| |
| Intraclass correlation | A summary of how far outcomes within a class or school are correlated, which determines how many clusters a trial needs. For school achievement it is typically around 0.2 (Hedges & Hedberg, 2007). |
| |
| M | |||
| Maryland Scientific Methods Scale | A scale written for crime prevention and widely borrowed, grading designs from level 1, correlation at a single point in time, to level 5, random assignment. It covers internal validity only (Sherman et al., 1998). |
| |
| Meta-analysis | A statistical synthesis that pools effect sizes across studies. It buys precision only if the studies are estimating something comparable; otherwise it produces a precise number for a quantity that does not exist. |
| |
| Mixed methods | Combining qualitative and quantitative data in one study or programme. Johnson and Onwuegbuzie (2004) argue for it as a third methodological movement rather than a compromise between two camps. | ||
| P | |||
| Peer review | Assessment of a manuscript by other researchers before publication. It filters obvious error but does not verify data, reproduce analyses, or correct for the incentive to publish positive findings. |
| |
| Positivism | The position that there is a regular social world that can be measured, and that research should seek generalisable relations within it. Most quantitative education research operates on assumptions of this kind. |
| |
| Practitioner inquiry | Systematic, intentional study of one’s own professional practice. Cochran-Smith and Lytle (2009) describe it as a stance towards teaching and knowledge rather than a project with an end date. |
| |
| Pragmatism | The position that the research question, not a prior philosophy, should decide the method, and that knowledge is judged by what it enables people to do. | ||
| Preregistration | Publicly recording hypotheses, design and analysis plan before the data are seen. Nosek et al. (2018) describe it as separating the generation of a hypothesis from its test. |
| |
| Programme evaluation | Judging the merit, worth or value of a programme for the people who must decide about it. It has its own discipline, standards and models, covered in Programme evaluation (C22). C22 The systematic assessment of a programme's merit, worth, probity and significance, for people who have to decide something about it (Stufflebeam, 2007). It is a field with its own models, standards and professional bodies. |
| |
| Publication bias | The tendency for positive and statistically significant results to be published more readily. Polanin et al. (2016) estimate the gap between published and unpublished effects at about 0.18 standard deviations. |
| |
| Q | |||
| Qualitative and quantitative research | Research working chiefly with meanings, accounts and observations, and research working chiefly with counts and measures. The distinction describes data, not rigour, and many studies use both. |
| |
| Quasi-experiment | A comparison study without random assignment, using matching, a threshold, or changes over time to approximate a counterfactual. The What Works Clearinghouse (2022) allows such designs to meet its standards only "with reservations". |
| |
| R | |||
| Randomised controlled trial | A study in which units are assigned to conditions by chance, so that the groups differ only by chance before the intervention. In education the units are often whole classes or schools. |
| |
| Realist evaluation | An approach that asks what works for whom in what circumstances, treating context as part of the finding rather than as noise (Pawson & Tilley, 1997). C22 Evaluation that asks "What works for whom in what circumstances and in what respects, and how?", on the claim that an outcome follows from a mechanism firing in a context (Pawson & Tilley, 2004). |
| |
| Regression discontinuity | A design that compares units just above and just below a cut-off for receiving something, on the argument that they are otherwise alike. Introduced by Thistlethwaite and Campbell (1960). |
| |
| Replication | Repeating a study to see whether the finding holds. Makel and Plucker (2014) found replications to be 0.13% of articles in the top 100 education journals, and less likely to succeed when the authors did not overlap. | ||
| Research paradigms | The assumptions a researcher brings about what exists, what can be known and how to find out. Guba and Lincoln (1994) set out the main competing positions. |
| |
| Research–practice partnerships | "Long-term collaborations between practitioners and researchers that are organized to investigate problems of practice and solutions for improving schools and school districts" (Coburn & Penuel, 2016). |
| |
| S | |||
| Systematic review | A review that states its search, inclusion criteria and appraisal method in advance so that it can be checked and repeated. It may or may not include a meta-analysis. |
| |
| W | |||
| What Works Clearinghouse | A US body that rates a finding as meeting its standards without reservations, with reservations, or not at all. The rating turns chiefly on how assignment to condition occurred (What Works Clearinghouse, 2022). |
| |
No term matches. Try fewer letters.
Nearby