Glossary

A Measurement Tool to Assess Systematic Reviews (version 2). A critical appraisal instrument used to evaluate the methodological quality of systematic reviews. It assesses 16 domains including protocol registration, search strategy, and risk of bias assessment.

The extent to which individuals and organisations are cognitively and emotionally prepared to accept and adopt a particular change. It encompasses beliefs about the need for change, the organisation's capacity to make it, and the personal benefits or costs involved.

A standardised measure of effect size that expresses the difference between two group means in terms of standard deviation units. Values of 0.2, 0.5, and 0.8 are conventionally considered small, medium, and large effects respectively, though context matters. Used in our infographic cards to display the practical magnitude of intervention effects (e.g., d = 0.60 for safety training).

A model developed by Quinn and Rohrbaugh that classifies organisational cultures along two dimensions — internal vs external focus and flexibility vs stability — yielding four culture types: Clan, Adhocracy, Market, and Hierarchy.

A range of values within which the true population parameter is likely to fall, typically expressed at the 95% level. Narrower intervals indicate more precise estimates. In meta-analyses, confidence intervals that exclude zero suggest statistically significant effects.

A correlation coefficient that has been statistically adjusted to account for measurement error (unreliability) in the instruments used. Corrected correlations are typically larger than raw correlations and provide a more accurate estimate of the true relationship between constructs. Also called disattenuated correlations.

A measure of the strength and direction of the linear relationship between two variables, ranging from -1.0 to +1.0. In organisational research, r = 0.10 is considered small, r = 0.30 medium, and r = 0.50 large. Our infographic cards display correlation coefficients as position markers on a scale, so you can quickly see how strong each relationship is relative to these benchmarks.

A study that collects data at a single point in time. Cross-sectional designs can show associations between variables but cannot establish whether one variable causes changes in another. Most organisational culture and engagement research is cross-sectional, which is a common limitation noted in our evidence quality assessments.

A quantitative measure of the magnitude of a relationship or difference. Unlike p-values, effect sizes indicate practical significance — how large an effect actually is, not merely whether it is statistically detectable. Common effect size measures include Pearson's r (correlation coefficient), Cohen's d (standardised mean difference), and variance explained (R²).

An approach to decision-making that integrates the best available scientific evidence with practitioner expertise, organisational context, and stakeholder values. Originated in medicine and increasingly applied in management and organisational development.

Grading of Recommendations Assessment, Development, and Evaluation. A systematic approach to rating the certainty of evidence across a body of studies. It considers study design, risk of bias, inconsistency, indirectness, imprecision, and publication bias.

A statistic that describes the percentage of variation across studies in a meta-analysis that is due to genuine differences rather than chance. Values above 50% suggest substantial heterogeneity, indicating that study results vary more than expected by sampling error alone. Low I² (e.g., 4%) means findings are highly consistent across studies.

A study that tracks the same participants or organisations over multiple time points. Longitudinal designs can reveal whether changes in one variable precede changes in another, providing stronger (though not definitive) evidence for causal relationships than cross-sectional studies.

A statistical relationship in which one variable (the mediator) explains the mechanism through which an independent variable affects an outcome. For example, safety climate mediates the relationship between leadership and safety behaviour — leadership improves climate, which in turn improves behaviour.

A statistical technique that combines the results of multiple independent studies addressing the same research question. By pooling data, meta-analyses produce more precise and generalisable estimates of effects than any single study can provide. Our articles synthesise findings from meta-analyses to present the strongest available evidence.

A statistical relationship in which a third variable changes the strength or direction of the relationship between two other variables. For example, industry type moderates the relationship between safety leadership and safety outcomes — the effect is stronger in high-risk industries.

The shared assumptions, values, beliefs, and norms that shape how members of an organisation think, feel, and behave. Culture operates at multiple levels — from visible artefacts and espoused values to deeply held, often unconscious basic assumptions (Schein, 2010).

In meta-analysis, the range within which the true effect of a future study is expected to fall. Unlike confidence intervals (which estimate the mean effect), prediction intervals capture the full spread of effects across contexts, giving a more realistic picture of variability.

The tendency for studies with statistically significant or positive results to be more likely to be published than those with null or negative findings. This can inflate effect sizes in meta-analyses if not detected and adjusted for (e.g., via funnel plots or trim-and-fill methods).

Behavioural, cognitive, or emotional responses that oppose or hinder organisational change efforts. Research increasingly treats resistance not as irrational obstruction but as a signal containing useful information about the change process and its implementation.

A rigorous, structured method of identifying, evaluating, and synthesising all relevant research on a specific question. Systematic reviews follow a predefined protocol including explicit search strategies, inclusion criteria, and quality appraisal to minimise bias.

A leadership style characterised by four components: idealised influence, inspirational motivation, intellectual stimulation, and individualised consideration. Meta-analyses consistently show moderate positive associations between transformational leadership and a wide range of follower and organisational outcomes.

The proportion of variation in an outcome that is accounted for by a predictor variable, expressed as a percentage. For example, if learning culture explains 20% of the variance in innovation, it means 20% of the differences in innovation outcomes across organisations can be attributed to differences in learning culture. Displayed as percentage arcs in our infographic cards.