Dissertation & Research Glossary

Abstract

A 150 to 350 word standalone summary of a dissertation covering problem, purpose, method, results, and implications. Written last, with no citations.

Acknowledgements

A 150 to 400 word section after the abstract thanking your chair, committee, funders, participants, and family, in that order.

Alignment

The consistency between problem, purpose, research questions, framework, design, and analysis. Misalignment is the most common reason US committees return a proposal.

Alpha level

The probability threshold for rejecting the null hypothesis, conventionally .05 in social science. Set before analysis, not after.

Analysis of variance (ANOVA)

A test comparing means across three or more groups. Running multiple t-tests instead inflates Type I error.

ANCOVA

Analysis of covariance, an ANOVA that statistically controls for one or more continuous covariates such as a pretest score.

Annotated bibliography

A reference list in which each entry is followed by a short critical summary of the source's argument, method, and relevance.

APA 7

The seventh edition of the American Psychological Association style guide, standard in US psychology, education, nursing, and social science programs.

Assumption

A condition you accept as true without proof, such as participants answering honestly. Distinct from a statistical assumption, which must be tested.

Audit trail

A documented record of qualitative analytic decisions that allows an external reader to follow how raw data became themes. Supports dependability.

Axial coding

The grounded theory stage in which open codes are grouped into categories and relationships between categories are specified.

Bracketing

In phenomenology, setting aside prior assumptions so participant meaning can be described on its own terms.

Candidacy

The status reached after coursework and comprehensive exams are complete and the proposal is approved. Often called ABD, all but dissertation.

Capstone project

An applied final project in professional programs such as DNP, EdD, or MPA, evaluated on practice improvement rather than original theory.

Chair

The faculty member who directs your dissertation, approves each submission, and manages the committee. The single most consequential relationship in a doctorate.

Chicago style

A citation system with notes-bibliography and author-date variants, standard in history, theology, and much of the humanities.

Codebook

A structured list of codes with definitions, inclusion and exclusion criteria, and example excerpts. Required for credible intercoder agreement.

Comprehensive exam

A written or oral examination on your field, passed before proposal approval. Also called quals, prelims, or comps.

Conceptual framework

A diagram and narrative showing the variables or constructs in your study and the expected relationships among them.

Confidence interval

A range that would contain the population parameter in a specified percentage of repeated samples, conventionally 95%.

Confirmability

The extent to which findings are shaped by data rather than researcher bias. Supported by reflexive memos and an audit trail.

Construct

An abstract concept such as burnout or self-efficacy that is measured indirectly through an instrument.

Construct validity

Evidence that an instrument measures the theoretical construct it claims to measure.

Content analysis

A systematic method for coding and counting the presence of categories in text or media, quantitative or qualitative in orientation.

Convergent validity

Evidence that a measure correlates with other measures of the same construct.

Credibility

The qualitative parallel to internal validity, supported by member checking, triangulation, and prolonged engagement.

Cronbach's alpha

A coefficient of internal consistency reliability. Values of .70 and above are conventionally acceptable, .80 and above preferred for research use.

Delimitation

A boundary you chose, such as studying only Midwestern districts. Distinct from a limitation, which is outside your control.

Dependability

The qualitative parallel to reliability, evidenced by consistent procedures and a complete audit trail.

Descriptive statistics

Summaries of a dataset such as means, standard deviations, frequencies, and distributions, reported before inferential tests.

Discussion chapter

Chapter 5 in a five-chapter dissertation, interpreting findings, tying them to theory, and stating limitations and implications.

Dissertation

The original research document required for a doctoral degree. In the US it is doctoral; in much of the UK the term refers to master's work.

Effect size

A standardized measure of magnitude such as Cohen's d, eta squared, or r. Required alongside p values in APA 7 reporting.

Epistemology

Assumptions about what counts as knowledge and how it can be known. Positivism, post-positivism, constructivism, and pragmatism are the common positions.

Ethnography

A design in which the researcher studies a culture-sharing group through extended fieldwork and participant observation.

Exempt review

An IRB determination that minimal-risk research qualifies for the lightest review category. Still requires IRB confirmation before data collection.

External validity

The extent to which findings generalize beyond the study sample, setting, and time.

Factor analysis

A family of techniques used to identify latent structure among items. Exploratory forms discover structure; confirmatory forms test a specified structure.

Five-chapter model

The dominant US dissertation structure: introduction, literature review, methodology, results, discussion.

Gap statement

The sentence in Chapter 2 that names precisely what the literature has not established and links it to your research questions.

Grounded theory

A design that builds theory inductively from data through open, axial, and selective coding with constant comparison.

Hierarchical regression

Regression in which predictors are entered in theory-driven blocks so the added variance of each block can be evaluated.

Homoscedasticity

The assumption that residual variance is constant across levels of the predictor. Checked with residual plots.

Hypothesis

A testable prediction about the relationship between variables, stated in null and alternative form in quantitative work.

Informed consent

The documented voluntary agreement of a participant who understands the purpose, procedures, risks, and right to withdraw.

Institutional Review Board (IRB)

The US committee that reviews and approves human subjects research before data collection begins.

Instrument

The tool used to collect data, such as a validated survey, interview protocol, or observation rubric.

Intercoder reliability

Agreement between independent coders on the same data, often reported as percentage agreement or Cohen's kappa.

Internal consistency

The degree to which items in a scale measure the same construct, commonly reported as Cronbach's alpha or omega.

Internal validity

The extent to which a design supports causal claims by ruling out alternative explanations.

Interpretive phenomenological analysis

A design focused on how individuals make sense of significant lived experience, using detailed idiographic analysis.

Latent class analysis

A technique that identifies unobserved subgroups within a population based on response patterns.

Likert scale

An ordered response format, typically five or seven points, used to measure agreement or frequency.

Limitation

A constraint outside your control that bounds the conclusions, such as self-report bias or a single-site sample.

Literature review

The chapter that synthesizes existing research by theme, critiques its methods, and establishes the gap your study fills.

Longitudinal design

A design that measures the same participants at multiple time points to observe change.

Maximum variation sampling

Purposive sampling that deliberately seeks diverse cases to capture the range of a phenomenon.

Measurement level

Whether a variable is nominal, ordinal, interval, or ratio. Determines which statistical tests are appropriate.

Member checking

Returning transcripts, summaries, or interpretations to participants to confirm accuracy. A core credibility strategy.

Mediation

A relationship in which an intervening variable explains how a predictor affects an outcome.

Methodology

Chapter 3, specifying design, sample, instruments, procedures, analysis plan, and ethical protections in enough detail to be replicated.

Mixed methods

A design that collects and integrates quantitative and qualitative data, with integration points stated explicitly.

MLA 9

The ninth edition of the Modern Language Association style, standard in literature and language programs.

Moderation

A relationship in which a third variable changes the strength or direction of the association between predictor and outcome.

Multicollinearity

High correlation among predictors that destabilizes regression coefficients. Screened with VIF and tolerance values.

Null hypothesis

The statement of no effect or no relationship that a significance test evaluates.

NVivo

Qualitative data analysis software used for coding, memoing, querying, and visualizing textual and multimedia data.

Operationalization

The process of specifying exactly how an abstract construct will be measured in this study.

p value

The probability of observing data at least as extreme as yours if the null hypothesis were true. Never a measure of effect magnitude.

Paradigm

The overarching set of philosophical assumptions guiding a study, such as post-positivism or constructivism.

Phenomenology

A design that describes the essence of lived experience of a phenomenon as reported by those who lived it.

Pilot study

A small preliminary run used to test instruments, timing, and procedures before full data collection.

Plagiarism

Presenting another person's words, ideas, data, or structure as your own. Includes self-plagiarism of prior submitted work.

Positionality

A statement of how the researcher's identity, experience, and assumptions shape access, interpretation, and representation.

Power analysis

A calculation of the sample size needed to detect an effect of a specified size at a chosen alpha and power, usually .80.

Proposal

Chapters 1 to 3 in future tense, defended before data collection begins.

Purposive sampling

Selecting participants deliberately because they can inform the research question, rather than randomly.

Qualitative research

Inquiry that examines meaning, process, and context through text, speech, observation, or artifacts rather than numeric measurement.

Quantitative research

Inquiry that measures variables numerically and tests relationships or differences statistically.

Quasi-experimental design

A comparison-group design without random assignment, used when randomization is not permissible or feasible.

Reflexivity

Ongoing critical self-examination of how the researcher influences and is influenced by the research process.

Reliability

The consistency of a measure across items, time, or raters. Necessary but not sufficient for validity.

Research question

The specific answerable question a study addresses, phrased so the planned analysis clearly answers it.

Results chapter

Chapter 4, reporting screening, descriptives, and findings per research question without interpretation.

Sampling frame

The accessible list from which a sample is actually drawn, which often differs from the target population.

Saturation

The point at which additional data produce no new codes or themes. Should be defined operationally, not asserted.

Secondary data

Data collected by someone else for another purpose, such as NHANES or IPEDS, reused in your analysis.

Snowball sampling

Recruitment in which existing participants refer additional participants, common with hard-to-reach populations.

SPSS

Statistical software widely used in US social science and nursing programs for descriptive and inferential analysis.

Structural equation modeling (SEM)

A framework for testing measurement and structural relationships among latent variables simultaneously.

Theoretical framework

An existing theory, with its propositions and boundary conditions, used to explain and interpret your phenomenon.

Thematic analysis

A six-phase method for identifying, developing, and reporting patterns of meaning across a qualitative dataset.

Thick description

Detailed contextual reporting that allows readers to judge transferability to their own setting.

Transferability

The qualitative parallel to generalizability, supported by thick description rather than statistical sampling.

Triangulation

Using multiple data sources, methods, analysts, or theories to strengthen confidence in findings.

Turnitin

A similarity-detection service used by most US universities. A similarity score is not by itself evidence of plagiarism.

Turabian

The student-oriented adaptation of Chicago style, common in theology, history, and divinity programs.

Type I error

Rejecting a true null hypothesis, a false positive. Its probability is the alpha level.

Type II error

Failing to reject a false null hypothesis, a false negative. Its probability is one minus power.

Validity

The extent to which an instrument or design supports the inferences drawn from it.

Variance inflation factor (VIF)

A multicollinearity diagnostic. Values above 5 to 10 signal that predictors are too highly correlated.

Viva

The oral defense of a dissertation, called a defense in US programs and a viva voce in UK and Commonwealth systems.