Good Research Practice

Good Research Practice (GRP) is a set of principles and standards governing the ethical and scientific quality of research across all disciplines. GRP encompasses study design, data collection and management, analysis, reporting, authorship, and regulatory compliance. While GCP specifically addresses clinical trials, GRP provides a broader framework applicable to all types of research, from laboratory studies to observational research to social science investigations.

Public trust in research depends on the integrity of the research process. Research misconduct (fabrication, falsification, and plagiarism) and questionable research practices (such as selective reporting, p hacking, and inappropriate authorship) undermine this trust and waste resources. GRP standards exist to reduce these problems and to promote research that is ethical, transparent, reproducible, and of high scientific quality. Institutions, funding agencies, and publishers increasingly require adherence to GRP standards. For individual researchers, following GRP protects their careers (research misconduct can result in retraction, funding suspension, and loss of position) and helps their work contribute reliably to the body of knowledge.

GRP is not set out in one document. It is described in codes of conduct and policies that institutions and funders adopt, for example the European Code of Conduct for Research Integrity (ALLEA, revised edition 2023), which applies to all scientific and scholarly fields and rests on four principles (reliability, honesty, respect and accountability); the Singapore Statement on Research Integrity (2010), a non regulatory statement of four principles and 14 responsibilities; and, in the UK, the Concordat to Support Research Integrity, refreshed in 2025, to which universities and funders commit. In the US, the federal research misconduct policy (2000) and, for Public Health Service funded research, 42 CFR Part 93 define the misconduct that is formally investigated as fabrication, falsification and plagiarism, excluding honest error and differences of opinion. Most other GRP expectations are enforced through institutional policies, funder conditions and journal requirements.

GRP standards address the entire research lifecycle. In the planning phase, GRP requires a clear research question, a sound study design, a priori statistical analysis plans, proper ethical approvals, and pre registration of study protocols when appropriate. During data collection, GRP requires standardized procedures, quality control checks, and proper training of research personnel.

Data management under GRP includes secure storage, regular backup, documented data cleaning procedures, and retention of raw data for verification. The FAIR principles (Findable, Accessible, Interoperable, Reusable) guide data sharing. Analysis should follow the pre specified plan, with any deviations documented and justified. Exploratory analyses should be clearly labeled as such.

Reporting under GRP follows the principle of complete and transparent disclosure. Reporting guidelines (such as CONSORT for trials, STROBE for observational studies, and PRISMA for systematic reviews) provide structured checklists. All outcomes specified in the protocol should be reported, including null or negative results. Selective outcome reporting is a violation of GRP.

Authorship under GRP follows established criteria (such as the ICMJE guidelines), requiring that each author meet all four ICMJE criteria, which include a substantial contribution to the work and accountability for it, that everyone who meets them is listed as an author, and that other contributors are acknowledged. Ghost authorship and gift authorship violate these standards.

Conflict of interest disclosure, proper citation of prior work, and responsible handling of human and animal subjects are also GRP requirements. Many institutions have Research Integrity Officers or offices that oversee compliance with GRP standards and investigate allegations of misconduct.

GRP applies to all research, regardless of discipline, funding source, or publication venue. It should guide researchers from the initial planning of a study through data collection, analysis, reporting, and archiving. Institutions should provide GRP training to all researchers, including students and early career investigators. GRP standards should be referenced in grant applications, institutional policies, and publication checklists.

A common error is treating GRP as relevant only to funded or published research, when the same standards apply to student projects, pilot studies, and internal reports. Another mistake is focusing on regulatory compliance (such as ethics approval) while neglecting scientific rigor (such as pre registration and transparent reporting). Selective reporting of outcomes (publishing only significant results) is a well documented violation of GRP, contributing to publication bias and wasted resources. Poor data management, including failure to document data cleaning decisions and failure to retain raw data, is another common deficiency. Finally, some researchers confuse GRP with GCP, not recognizing that GCP is a specialized quality standard for clinical trials, written into law in some jurisdictions (for example, the EU Clinical Trials Regulation requires trials of medicinal products to follow GCP principles), while GRP is the broader set of expectations for all research activities, applied mostly through codes of conduct, institutional policies and funder conditions.

A university implements a comprehensive GRP program that includes: mandatory GRP training for all graduate students and postdoctoral researchers (12 hours, with annual refresher modules), a requirement that all studies collecting new data from human participants be pre registered before data collection, institutional data management standards requiring 10 year retention of raw data, an anonymous reporting system for suspected research misconduct, and a Research Integrity Office that reviews 5 randomly selected studies per department per year for GRP compliance. In the first year, the program identifies 3 instances of incomplete record keeping and 2 instances of retroactive changes to analysis plans, which are addressed through targeted training rather than disciplinary action.