Dissertation Abstract Examples: 6 Models and a Formula

By Dr. Patricia N., DNP, PhD in Nursing Science. Reviewed by Dr. James W.. 2026-06-16.

Key takeaways

• Use the six-sentence formula: problem, purpose, method, results, interpretation, implication
• Include actual numbers or named themes, not just claims of significance
• Keep to 150 to 350 words and check your graduate school template
• No citations, no undefined abbreviations, past tense for completed work
• Write the abstract last, after Chapter 5 is final

The short answer

A dissertation abstract is 150 to 350 words and contains six sentences of work: problem, purpose, method, key results with numbers, interpretation, and implication. Write it last, and write it without citations or abbreviations that are not defined.

The six-sentence formula

Sentence 1: the problem and its scale.

Sentence 2: the purpose of this study, using the verb your design implies (examined, explored, evaluated, developed).

Sentence 3: design, sample, setting.

Sentence 4: instruments and analysis.

Sentence 5: the main finding with actual numbers or themes.

Sentence 6: what it means and who should act on it.

Example 1: Quantitative (198 words)

Teacher attrition in urban districts costs American public education an estimated $8.5 billion annually, and school leadership is the most frequently cited modifiable factor. The purpose of this quantitative study was to examine the relationship between principals' transformational leadership behaviors and teacher intent to remain. A correlational design was used with 312 teachers across 12 urban districts in the Midwest. Leadership was measured with the Multifactor Leadership Questionnaire and retention intent with a four-item scale; data were analyzed using hierarchical multiple regression controlling for school size, principal tenure, and years of teaching experience. Idealized influence significantly predicted intent to remain (beta = .41, p < .001), explaining 17% of variance beyond the control block, while contingent reward did not (beta = .06, p = .31). Findings suggest that relational leadership behaviors, rather than transactional incentives, are associated with retention intent once school stability is controlled. District leaders investing in retention should prioritize principal development in modeling and inspirational communication over incentive redesign.

Example 2: Qualitative (176 words)

Patient-safety research consistently links nurse burnout to adverse events, but the mechanism connecting exhaustion to unsafe practice remains unclear. This interpretive phenomenological study explored how registered nurses experience and respond to conditions they perceive as unsafe. Thirty-one nurses across four hospital systems participated in semi-structured interviews of 45 to 75 minutes, analyzed through six-phase thematic analysis with a second coder on 20% of transcripts. Three themes emerged: the normalization of workarounds, silence as self-protection, and moral residue that accumulated across shifts rather than resolving between them. Participants attributed unsafe practice less to exhaustion itself than to the informal shortcuts exhaustion made necessary. Findings indicate that safety interventions targeting individual resilience may miss the workaround pathway through which staffing pressure becomes clinical risk. Hospital leaders should audit informal workarounds as a leading indicator rather than treating incident reports as the primary safety signal.

Example 3: Mixed methods (211 words)

Supply chain disruption since 2020 has produced a large practitioner literature on resilience strategies but little evidence on sequencing. This mixed methods study evaluated which resilience strategies US manufacturers adopted and why, using a sequential explanatory design. In phase one, 148 supply chain managers completed a strategy adoption survey; adoption patterns were analyzed with latent class analysis, identifying three adopter profiles. In phase two, 14 managers selected from the profiles participated in interviews analyzed thematically, with integration through joint displays. Diversified adopters recovered to pre-disruption output 40% faster than single-strategy adopters, and interviews attributed the difference to pre-existing supplier relationships rather than to the strategies themselves. Findings suggest that resilience is a function of relational capacity built before disruption, not of strategy selection during it. Firms should treat supplier relationship investment as resilience infrastructure rather than as procurement overhead.

Example 4: DNP practice project (162 words)

Central line-associated bloodstream infections remained above national benchmarks on two medical-surgical units at the project site. The purpose of this DNP quality improvement project was to implement and evaluate a nurse-led maintenance bundle using the Plan-Do-Study-Act framework across four cycles over 16 weeks. Interventions included daily line necessity review, standardized dressing change documentation, and peer audit. The infection rate declined from 2.1 to 0.7 per 1,000 catheter days, and bundle compliance rose from 61% to 94%. Sustainability was supported through integration into unit onboarding and monthly audit reporting. Results indicate that a nurse-led maintenance bundle with peer audit can reduce infection rates without additional staffing, and the model is transferable to comparable inpatient units.

Example 5: Computational (168 words)

Deep learning intrusion detection for IoT healthcare networks reports benchmark accuracy above 98% but is rarely evaluated across datasets. This study developed and evaluated a hybrid CNN-LSTM architecture for real-time detection in IoT healthcare environments. The model was trained and tested on CICIoT2023 and evaluated cross-dataset on two additional corpora, with ablation analysis isolating the contribution of each component. In-dataset accuracy reached 98.7% with an F1 of 97.9%, exceeding three baseline architectures by 3.2 percentage points while reducing inference cost by 28%. Cross-dataset accuracy fell to 87.4%, and ablation showed the LSTM component accounted for most of the retained generalization. Findings indicate that reported benchmark gains overstate deployment readiness and that temporal modeling contributes disproportionately to transfer.

Example 6: Humanities (154 words)

Scholarship on antebellum reform movements has treated temperance and abolition as parallel but separable campaigns. This dissertation argues that in the upper Midwest between 1840 and 1860 the two were institutionally inseparable, sharing personnel, meeting houses, and publication networks. Drawing on 11 previously uncatalogued church record collections, county newspaper runs, and the correspondence of four itinerant organizers, it reconstructs how reform labor circulated between causes. Chapter three demonstrates that congregations expelled members for liquor sales and slaveholding under the same disciplinary logic. The study revises the standard periodization by locating the fusion of the two movements a decade earlier than existing accounts allow, and it suggests that the organizational history of American reform is better read through local institutions than through national societies.

Word limits and common rejections

ProQuest historically capped abstracts at 350 words; many US graduate schools set 150 to 300. Check your template. Abstracts are most often returned for missing numeric results, undefined abbreviations, citations, future tense in a completed study, and a purpose sentence that does not match the design named two sentences later.

Want a PhD editor to review your draft against these examples? Send it over for a free 500-word sample edit and a fixed USD quote.

FAQ

How long should a dissertation abstract be?

150 to 350 words. ProQuest historically capped submissions at 350 words and many US graduate schools set lower limits, so check your template.

Should a dissertation abstract include citations?

No. Abstracts should stand alone without references, and most style guides and graduate schools prohibit citations there.

When should I write the abstract?

Last, after Chapter 5 is final, because the abstract must report the findings and implications you actually landed on.