The Scientific Method and the Importance of Documenting Results

Abb. 1: The Scientific Method

Autor:	Prof. Dr.-Ing. Ulrich Schneider
Sprache:	🇩🇪  🇬🇧

Abstract

The scientific method is a systematic approach to investigating natural and social phenomena through observation, hypothesis formation, experimentation, analysis, and critical evaluation. An essential component of scientific research is the documentation of methods, observations, data, and results. Accurate documentation enables researchers to evaluate the validity of conclusions, reproduce analyses, identify errors, and build upon previous discoveries. Together, systematic inquiry and transparent recordkeeping support the reliability, reproducibility, and cumulative development of scientific knowledge.

Introduction

The scientific method is a collection of principles and practices used to develop and evaluate knowledge about the world. Although its specific procedures vary among disciplines, scientific inquiry generally involves formulating questions, developing explanations, gathering evidence, and assessing whether the evidence supports those explanations.

Scientific knowledge is provisional: conclusions may be revised when new evidence becomes available or when existing findings are challenged by further investigation. For this reason, science depends not only on conducting experiments and making observations, but also on recording how investigations were performed and how their conclusions were reached.

Documentation provides a traceable record of the research process. Without sufficiently detailed records, observations may become difficult to interpret, analytical errors may remain undetected, and other researchers may be unable to assess or reproduce the reported findings [1].

The scientific method

The scientific method does not constitute a single, universally prescribed sequence of steps. Different fields employ different approaches, including controlled experiments, observational studies, mathematical modelling, field research, and historical analysis. Nevertheless, several activities are common to many forms of scientific investigation.

Observation and question formulation

Scientific investigations frequently begin with observations of a phenomenon, an unresolved problem, or a discrepancy between existing explanations and available evidence. Researchers formulate questions that define the subject of investigation and help determine what evidence is needed.

Clear research questions help establish the scope of a study and distinguish relevant observations from unrelated information.

Hypothesis formation

A hypothesis is a proposed explanation or prediction that can be evaluated against evidence. Depending on the discipline, a hypothesis may predict a relationship between variables, explain an observed phenomenon, or identify conditions under which a particular outcome should occur.

Researchers may also conduct exploratory investigations without formulating a specific hypothesis in advance. Distinguishing initial expectations from conclusions developed after examining the evidence can improve the transparency of the research process.

Research design and experimentation

Researchers select methods appropriate to their questions. These may include laboratory experiments, surveys, field observations, simulations, archival research, or the analysis of existing datasets.

Research design establishes how observations will be collected and evaluated. Relevant considerations may include:

The materials, instruments, and equipment used. The variables and conditions under investigation. Sampling procedures and selection criteria. Control conditions and comparison groups, where applicable. Measurement procedures and calibration. Data collection schedules and protocols. Ethical, safety, and legal requirements.

Documenting the research design allows readers to understand the conditions under which the evidence was obtained.

Data collection and analysis

Researchers collect observations, measurements, experimental outputs, or other evidence and analyse them using methods appropriate to the research question.

Analysis may involve statistical calculations, qualitative interpretation, mathematical modelling, computational procedures, or comparisons with established theories. Decisions made during data cleaning, exclusion of observations, selection of models, or interpretation of ambiguous results can affect the conclusions.

Recording these decisions is important, particularly when they influence the reported findings.

Interpretation and conclusion

Researchers evaluate whether the evidence supports, contradicts, or fails to resolve the original hypothesis or research question. Conclusions should reflect the strength and limitations of the available evidence.

A result that is consistent with a hypothesis does not necessarily prove that the hypothesis is universally true. Alternative explanations, measurement uncertainty, sampling limitations, and methodological weaknesses may remain.

Scientific conclusions can be revised as additional observations, improved methods, or independent studies become available.

Communication and further investigation

Scientific findings are communicated through research articles, reports, datasets, technical documentation, presentations, and other forms of scholarly communication.

Publication allows other researchers to examine the evidence, evaluate the methods, identify potential weaknesses, and investigate related questions. Scientific inquiry consequently operates as an ongoing process rather than a sequence that necessarily ends with a definitive conclusion.

Documentation of scientific research

Scientific documentation is the systematic recording and preservation of information about the purpose, methods, observations, analyses, and outcomes of an investigation.

Documentation may take the form of laboratory notebooks, field notes, experimental logs, electronic spreadsheets, databases, source code, software notebooks, photographs, instrument records, protocols, or formal research reports.

The appropriate form depends on the discipline and the nature of the investigation. Regardless of format, records should be sufficiently complete, accurate, organised, and understandable to support subsequent examination of the research.

Recording methods and procedures

Researchers should document the procedures used to obtain their results, including relevant equipment, materials, experimental conditions, instruments, measurement techniques, and analytical methods.

Where applicable, records should identify instrument settings, software versions, calibration procedures, environmental conditions, and deviations from the original protocol.

Such information helps distinguish findings that arise from the phenomenon being investigated from differences introduced by changes in experimental conditions or research procedures.

Recording observations and raw data

Observations and raw data provide the evidential basis for subsequent analysis. They should be recorded as accurately as practicable and preserved separately from processed or interpreted results.

Records should identify the units of measurement, relevant timestamps, sample identifiers, data sources, and contextual information needed to interpret the observations.

When data are corrected, transformed, filtered, or excluded, the changes and their justifications should be documented. Preserving original records, where feasible and appropriate, helps researchers identify errors and reconstruct the analytical process.

Documenting analysis

Data analysis frequently involves choices that can affect the outcome of a study. Documentation should therefore describe the procedures used to transform data into results.

Depending on the research, this may include statistical methods, mathematical equations, assumptions, computational scripts, model parameters, software dependencies, and criteria for including or excluding observations.

For computational research, retaining the relevant data, code, parameters, and information about the computing environment can be particularly important. A description of the general method alone may be insufficient to reproduce results generated by complex analytical workflows [1].

Reporting results and uncertainty

Research documentation should distinguish observations from interpretations and conclusions. Results should be presented with appropriate units, graphical representations, statistical summaries, and descriptions of uncertainty.

Where relevant, researchers should report measurement precision, confidence intervals, sampling limitations, model assumptions, and other sources of uncertainty.

Unexpected findings, unsuccessful experiments, and results that contradict the initial hypothesis may also be scientifically informative. Recording them helps reduce the risk that the research record presents an incomplete picture of the investigation.

Importance of documenting results

Reproducibility

Reproducibility is an important objective of transparent research. In computational science, the term commonly refers to obtaining consistent results by applying the same analytical procedures to the same input data and computational conditions.

Adequate documentation allows researchers to reconstruct the steps that produced a result, inspect analytical decisions, and determine whether the reported output follows from the supplied evidence.

Without appropriate records, reproducing an analysis may be difficult or impossible, even when the original findings appear plausible.

Replicability and independent verification

Replicability concerns whether further investigations addressing the same scientific question produce sufficiently consistent findings when new data are collected.

A documented study provides a basis for designing an independent investigation that follows the original methods or deliberately modifies them to examine the robustness of the findings.

Replication attempts may produce different outcomes for legitimate reasons, including sampling variation, differences in populations, changing environmental conditions, measurement limitations, or weaknesses in the original explanation. Such differences can help refine scientific understanding.

The National Academies distinguishes computational reproducibility from replicability based on new data, while noting that terminology can vary among disciplines.

Error detection and correction

Detailed records support the identification of errors in measurements, calculations, data processing, equipment operation, and interpretation.

For example, an incorrect unit conversion may affect a numerical result. A record of the original measurement, conversion formula, and subsequent calculations allows the error to be traced and corrected.

Documentation also helps researchers distinguish isolated mistakes from systematic problems that may affect an entire study.

Transparency and research integrity

Transparent documentation makes the basis of scientific claims available for scrutiny. It can help identify selective reporting, unexplained exclusions, inappropriate analytical decisions, and discrepancies between reported methods and actual procedures.

A complete research record supports responsible scientific conduct by preserving an account of what was done, what was observed, and how conclusions were derived.

Documentation alone cannot guarantee honesty or methodological quality, but it provides important evidence for evaluating research practices and investigating disputed findings.

Continuity and cumulative knowledge

Scientific research frequently involves collaboration among individuals and institutions, sometimes over many years. Researchers may need to revisit earlier experiments, use archived datasets, or continue investigations initiated by other teams.

Well-maintained documentation preserves the context required to understand previous work and reduces dependence on individual memory. It also enables later researchers to compare findings across studies, develop new hypotheses, and combine evidence through systematic reviews or meta-analyses.

Efficient use of resources

Experiments and field investigations can require substantial time, funding, materials, and specialised equipment. Documentation helps researchers understand which procedures have already been attempted, which conditions were tested, and what difficulties were encountered.

These records can reduce unnecessary duplication, support troubleshooting, and make it easier to plan subsequent investigations. Repeating an experiment can still be scientifically valuable when undertaken to verify a finding or investigate uncertainty.

Principles of effective scientific documentation

Although documentation requirements vary across disciplines, several principles are broadly applicable.

Accuracy

Records should reflect actual observations, procedures, and analytical decisions. Measurements and results should not be altered to fit expectations.

Completeness

Records should contain enough information to understand the investigation and assess how its results were obtained.

Clarity

Methods, units, terminology, abbreviations, and calculations should be explained sufficiently for the intended scientific audience.

Traceability

Data and results should be linked to their sources, collection procedures, transformations, and analytical steps.

Timeliness

Observations and procedures should be recorded as close as practicable to the time they occur, reducing reliance on later recollection.

Consistency

Standardised formats, naming conventions, units, and recording procedures can improve the comparison and interpretation of records.

Preservation

Records should be stored in ways that protect their integrity and maintain their usability over time. Appropriate backups, version control, file formats, and access arrangements can support long-term preservation.

Reproducibility

Where feasible, documentation should enable other qualified researchers to reconstruct the relevant methods and evaluate the reported results.

These principles can be adapted to the practical and ethical requirements of individual research fields.

Data sharing and open science

Open science promotes the accessibility, transparency, and reuse of scientific knowledge and research outputs. Depending on the study, these outputs may include publications, research data, protocols, software, source code, analytical workflows, and supporting documentation.

Repositories and recognised data-management practices can help preserve research outputs and associate them with publications through persistent identifiers and descriptive metadata.

The FAIR data principles—findability, accessibility, interoperability, and reusability—provide a framework for improving the usefulness of research data [2].

However, documentation and transparency do not always require unrestricted public release of every record. Research involving personal information, confidential material, culturally sensitive knowledge, protected species, or security-sensitive information may require access controls or other safeguards. Data sharing should follow applicable ethical, legal, and governance requirements.

The UNESCO Recommendation on Open Science emphasises the importance of open access to scientific knowledge and research outputs, subject to appropriate safeguards and responsible data governance [2].

Limitations and challenges

Maintaining comprehensive scientific records requires time, resources, training, and appropriate infrastructure. Researchers may encounter difficulties when working with large datasets, specialised equipment, proprietary software, changing file formats, or complex collaborative workflows.

Some observations cannot be repeated because the original phenomenon was temporary, historical, or unique. Other studies may be impossible to replicate exactly because environmental conditions, populations, or technologies change.

Furthermore, detailed documentation does not automatically establish that a conclusion is correct. A study can be meticulously recorded and still contain flawed assumptions, inadequate measurements, biased sampling, or an inappropriate analytical method.

A failure to reproduce or replicate a result is not, by itself, proof of misconduct. Differences may reveal uncertainty, contextual dependence, methodological weaknesses, or previously unrecognised features of the phenomenon.

The purpose of documentation is therefore not to guarantee identical outcomes in every investigation, but to make the research process sufficiently transparent that the evidence, reasoning, limitations, and possible explanations can be critically evaluated.

Conclusion

The scientific method provides a framework for investigating questions through evidence, critical evaluation, and the revision of explanations. Documentation is integral to this process because it preserves the connection between observations, methods, analyses, and conclusions.

Accurate and accessible records support reproducibility, independent verification, error correction, research integrity, and the cumulative development of knowledge. They also enable future researchers to understand previous investigations and assess whether conclusions remain justified in light of new evidence.

Scientific progress consequently depends not only on discovering new information, but also on maintaining a reliable record of how that information was obtained. Documenting results is an essential part of making scientific knowledge transparent, testable, and useful to others.

References

1. National Academies of Sciences, Engineering, and Medicine (2019). Reproducibility and Replicability in Science. Washington, DC: The National Academies Press. https://doi.org/10.17226/25303.
2. UNESCO (2021). UNESCO Recommendation on Open Science. Official text.