From first search to final submission.One home for the whole research journey.

A short summary of a paper placed at the top of the manuscript, covering the purpose, method, key results and conclusions (typically 150 to 300 words).

A cyclical method in which practitioners study and improve their own practice through repeated rounds of planning, acting, observing and reflecting.

The aim states what the study intends to achieve overall; objectives break that aim into specific, assessable steps.

A reference list in which each source is followed by a brief note describing its content and its relevance to your research.

Anonymity means no one, including the researcher, can connect data to an identity; confidentiality means identities are known but protected from disclosure.

Analysis of variance: a test that compares the means of three or more groups to see whether at least one differs significantly.

A list of everything consulted during research, whether cited in the text or not, distinct from a reference list of cited sources only.

An in depth investigation of a single person, group, event or context, using multiple sources of evidence.

A non parametric test that examines whether two categorical variables are independent of each other.

The standardised format for acknowledging sources, for example APA, MLA, Chicago, Harvard, IEEE or Vancouver. Consistency across the paper matters more than the choice itself.

A statistic measuring how far two raters' agreement exceeds what chance alone would produce, used for coding decisions.

Your own model of the specific concepts in the study and how you expect them to relate, often shown as a diagram.

A range of values that is likely to contain the true population value, with a stated level of confidence (commonly 95%).

A variable related to both the presumed cause and the outcome that can manufacture a misleading association if left uncontrolled.

The comparison condition that does not receive the intervention, providing the baseline against which its effect is judged.

Selecting participants because they are easy to reach (a lecture theatre, an online panel, your own workplace), not because they represent the population.

A number (r) between −1 and +1 expressing the strength and direction of a relationship between two variables. It does not imply causation.

A statistic for the internal consistency of a multi item scale: how far the items seem to measure the same underlying construct.

A study that collects data from many subjects at a single point in time, providing a snapshot of a population.

Fabrication is inventing data that was never collected; falsification is altering real data so it no longer represents what happened. Both are research misconduct.

Reasoning from a general theory to specific, testable predictions or hypotheses.

The outcome variable the study measures, which the independent variable is expected to influence.

Digital Object Identifier: a permanent, unique ID for a paper or dataset that always resolves to the correct source.

The extent to which findings hold in real world settings rather than the artificial conditions of the study.

A measure of how large a difference or relationship actually is (e.g. Cohen's d, η²), complementing (not replacing) statistical significance.

The branch of philosophy concerned with what counts as knowledge and how knowledge can be acquired and justified.

A qualitative method in which the researcher immerses themselves in a community to observe and describe its culture from within.

The U.S. Family Educational Rights and Privacy Act: legislation governing the privacy of student education records.

A facilitated discussion with several participants at once, used to surface shared views, disagreements and group reasoning.

The EU and UK General Data Protection Regulation: the law governing how personal data must be collected, stored, used and deleted.

Research outputs published outside commercial channels: theses, government reports, preprints, working papers and conference proceedings.

A qualitative approach that builds theory from the data itself, with data collection and analysis proceeding iteratively through constant comparison.

Hypothesising after the results are known: presenting a post hoc explanation as if it were the study's original, pre registered prediction.

Cautious language (e.g. 'the results suggest') used to present claims with appropriate uncertainty rather than overstatement.

A clear, testable prediction about the relationship between variables, stated before data collection.

The standard empirical paper structure: Introduction, Methods, Results and Discussion.

A brief source acknowledgement within the body of the paper (e.g. '(Smith, 2020)'), keyed to the full reference list entry.

The variable the researcher manipulates or treats as the presumed cause; the outcome is expected to change with it.

Reasoning from specific observations toward broader patterns, themes or theories.

The ethical requirement that participants understand what a study involves and agree to take part voluntarily, free to withdraw at any time.

The stance that human reality is constructed through meaning and best understood by interpreting how participants experience and make sense of their world.

Institutional Review Board: the committee that reviews and approves research involving human participants to ensure ethical protection.

The four kinds of data (nominal, ordinal, interval, ratio) that determine which statistics and tests are legitimate.

The honest statement of what a study cannot claim: constraints of design, sample, measures and analysis, and their consequences.

A synthesis of existing scholarship on a topic that identifies what is known, what is contested, and the gap your research addresses.

A study that follows the same subjects over an extended period, observing how outcomes change over time.

The arithmetic average of a set of values; sensitive to extreme outliers.

The middle value when data are ordered from lowest to highest; half the values sit above it and half below.

A statistical synthesis of effect sizes across multiple studies on the same question, producing an overall pooled estimate.

Research that combines quantitative and qualitative approaches to benefit from both numerical measurement and interpretive depth.

The most frequently occurring value in a dataset; the only average defined for nominal data.

The symmetric, bell shaped distribution assumed by many statistical tests; roughly 95% of values fall within two standard deviations of the mean.

The default assumption that there is no effect or no difference; statistical tests evaluate how compatible the data are with it.

The branch of philosophy concerned with what exists and what reality is like; your ontological stance sets what kind of claims your research can make.

Publishing that makes research free to read online, either in an open journal (gold) or via a repository copy (green).

Defining an abstract concept in terms of concrete, measurable indicators so it can actually be studied.

The probability of observing data at least as extreme as yours if the null hypothesis were true. A small p value is evidence against the null, not proof of your hypothesis.

A coherent set of assumptions about reality, knowledge and method (positivist, interpretivist, critical, pragmatist) that shapes how research is done.

Restating another author's idea in your own words and structure, with a citation. It is not paraphrasing if you keep the original phrasing.

The process by which experts evaluate a manuscript before publication, assessing its rigour, originality and validity.

A qualitative approach that explores how individuals experience and give meaning to a particular phenomenon.

A small scale trial run of your instruments and procedures, used to find problems before the main study pays for them.

Presenting another person's words, ideas or work as your own without attribution (whether deliberate or accidental).

The stance that reality exists objectively and independently, and is best studied through observation, measurement and hypothesis testing.

The probability that a test detects a real effect (conventionally targeted at 0.80). Low power leads to missed findings.

An exploitative outlet that charges publication fees while providing little or no genuine peer review, prioritising income over quality.

Publicly registering a study's hypotheses, sample plan and analysis decisions before data are collected, freezing the plan against hindsight.

Data you collect yourself directly from participants, sensors or observation, as opposed to using existing datasets.

The tendency for significant, positive results to be published while null results go missing, systematically distorting the published literature.

Selecting participants deliberately because they hold the characteristics your question requires, a standard qualitative strategy.

Research that explores meanings, experiences and processes through non numerical data such as interviews, focus groups and observation.

Research that measures variables numerically and analyses them with statistics, aiming for generalisable findings.

A study structured like an experiment, comparing conditions, but without random assignment, which weakens causal claims.

An experimental design in which participants are randomly assigned to intervention or control conditions, minimising selection bias.

The section of an introduction that justifies why the study matters, why now, and why this approach, bridging the gap and the aims.

The full bibliographic details of every source cited in the paper, formatted in one consistent citation style.

A model that predicts one outcome variable from two or more predictors, estimating each predictor's unique contribution while holding the others constant.

The consistency of a measure: whether repeating the measurement under the same conditions produces the same result.

Repeating a study's methods (directly or with extension) to check whether its findings reproduce, the backbone of trustworthy knowledge.

The finding across several fields that many published studies fail to reproduce when independently repeated.

The specific question, population, setting or method that existing research has left unanswered, which your study is designed to address.

The single, clearly phrased question a study sets out to answer; every design decision should serve it.

The process of selecting cases from a population to study. Probability sampling allows generalisation; convenience sampling does not.

Data collected by someone else, for another purpose, that you reanalyse to answer your own question.

Reusing your own previously published work without disclosure or citation, which can misrepresent work as new.

An interview built on a flexible topic guide: core questions asked of everyone, with freedom to probe, reorder and follow unexpected leads.

Language that shows readers where the argument is and where it is going: forecasts at section starts, recaps at ends, and explicit connections between parts.

The asymmetry of a distribution. Highly skewed data may require non parametric tests or transformation.

Recruiting participants through referrals from existing participants, suited to hidden or hard to reach populations.

A measure of how spread out values are around the mean; larger values indicate more variability.

The typical distance between a sample statistic and the true population value; it shrinks as samples grow.

A result unlikely under the null hypothesis, conventionally p < 0.05. Significance says an effect is detectable, not that it is large or important.

Dividing the population into meaningful subgroups (strata) and sampling from each, proportionally, to guarantee their representation.

A rigorous, pre specified search and appraisal of all available studies on a question, designed to minimise bias in selecting and evaluating evidence.

A test comparing means: one sample against a known value, or two groups against each other (independent or paired).

A qualitative method for identifying, organising and reporting patterns (themes) across a dataset, the most widely taught qualitative analysis.

The existing theory a study draws on to define its concepts and explain how they relate, giving the analysis its lens.

The point in qualitative data collection at which new data stops yielding new themes or insights, signalling fieldwork can end.

A single sentence that states the central argument or claim of a paper, usually at the end of the introduction.

Corroborating findings by using multiple methods, data sources or analysts, strengthening confidence in the conclusion.

Type I error: concluding an effect exists when it does not (false positive). Type II error: missing a real effect (false negative).

The degree to which a measure or study actually captures the concept it claims to, meaning internal validity for causal claims and external validity for generalisability.