Common Limitations in Research: Examples and How to Address Them

Research studies in the life sciences, education, and clinical fields face recurring limitations that affect the strength and interpretation of their findings. These limitations include small sample sizes, selection bias, measurement error, and design constraints that restrict generalizability. This article explains the most common research limitations, provides concrete examples from published studies, and offers practical strategies for identifying, mitigating, and disclosing these limitations in manuscripts. The guidance applies to students preparing theses, early-career researchers developing their first studies, and experienced professionals reviewing or designing research protocols.

At a Glance

The table below summarizes the most common research limitations, their typical effects on study validity, and practical strategies for addressing each one.

Understanding Research Limitations

A research limitation is any aspect of a study that may reduce the confidence readers can place in its findings or restrict the applicability of those findings to other contexts. Limitations exist in every study, regardless of how carefully the research is designed and executed. The presence of limitations does not invalidate a study, but failing to acknowledge them can mislead readers and undermine the credibility of the research.

Researchers should distinguish between limitations that could have been avoided with better planning and those that are inherent to the research question or methodology. For example, a researcher who fails to conduct a sample size calculation before recruiting participants has an avoidable limitation. A researcher studying a rare condition who can only recruit a small number of affected individuals has a limitation that reflects the realities of the study population.

The National Institute of Standards and Technology maintains a Research Data Framework that provides guidance on documenting research processes and data management practices. This framework emphasizes the importance of transparent documentation throughout the research lifecycle, which includes recording decisions about study design, data collection, and analysis that may later be identified as limitations.

Sample Size Limitations

Small sample sizes are among the most frequently cited limitations in research. A study with too few participants may fail to detect a true effect, produce imprecise estimates, and limit the ability to conduct subgroup analyses. The consequences of small samples extend beyond statistical concerns to affect the practical usefulness of the findings.

Why Sample Size Matters

Statistical power refers to the probability that a study will detect an effect when one truly exists. Studies with low power are more likely to produce false negative results, leading researchers to conclude that no effect exists when the effect is actually present. Small samples also produce wider confidence intervals, which means the true effect size is less precisely estimated.

A cross-sectional study of learning style preferences among clinical students in Nigeria illustrates how sample size constraints affect analysis options. The researchers collected data from 200 medical students and planned to use multivariable modeling to examine factors associated with learning style preferences. However, the categorical structure of the primary outcomes, sparse cells in some modality categories, and sample size limitations prevented them from performing the prespecified multinomial modeling. The researchers had to rely on bivariate associations instead, which limited their ability to control for confounding variables.

Strategies for Addressing Sample Size Limitations

Researchers should conduct a formal sample size calculation during the planning phase of any study. This calculation requires specifying the expected effect size, the acceptable level of statistical significance, and the desired level of power. When the target sample size cannot be achieved, researchers should document the reasons and describe how the reduced sample affects the interpretation of results.

Practical strategies for maximizing sample size include extending the recruitment period, adding additional study sites, and using multiple recruitment methods. Researchers can also consider whether a less resource-intensive design, such as a retrospective record review, could answer the research question with a larger sample.

When reporting results from a small sample, researchers should present effect sizes with confidence intervals instead of relying solely on p-values. This approach gives readers a clearer sense of the precision of the estimates and helps them judge the practical significance of the findings.

Selection Bias and Sampling Issues

Selection bias occurs when the participants in a study differ systematically from the population the researcher intends to represent. This bias can distort study findings in ways that are difficult to detect or correct after data collection is complete.

Common Sources of Selection Bias

Volunteer bias arises when individuals who choose to participate in a study differ from those who decline. Participants who volunteer for research studies often have different health status, education levels, or attitudes than the general population. Convenience sampling, in which researchers recruit participants who are easily accessible, can introduce similar distortions.

Attrition bias affects longitudinal studies when participants drop out at different rates across study groups. If the participants who leave a study differ systematically from those who remain, the final sample may no longer be representative of the original cohort.

A study of first-year medical students transitioning from memorisation-based learning to problem-solving approaches in a computer science course illustrates how participant selection shapes findings. The researchers interviewed 12 students who had completed a mandatory course, and their qualitative analysis revealed consistent patterns in how students adapted their learning strategies. The findings provide insight into the experiences of students who persisted through the course, but they may not represent the experiences of students who struggled more severely or who withdrew from the course.

Strategies for Reducing Selection Bias

Random sampling is the most effective method for ensuring that a study sample represents the target population. When random sampling is not feasible, researchers should use systematic approaches such as consecutive enrollment of eligible participants or stratified sampling to ensure that important subgroups are represented.

Researchers should document the recruitment process in detail, including the number of individuals approached, the number who agreed to participate, and the reasons for nonparticipation. Comparing the characteristics of participants with those of nonparticipants can help readers judge the potential for selection bias.

The NC3Rs Experimental Design Assistant provides a structured platform for planning experiments and identifying potential sources of bias before data collection begins. This tool guides researchers through the design process and highlights aspects of the design that may introduce bias, including allocation methods and blinding procedures.

Measurement Error and Instrument Limitations

Measurement error refers to the difference between the value recorded in a study and the true value of the variable being measured. Measurement error can affect both exposure variables and outcome variables, and it can bias study findings in either direction.

Types of Measurement Error

Systematic measurement error occurs when measurements are consistently too high or too low. This type of error can result from faulty calibration of instruments, poorly worded survey questions, or observer bias. Random measurement error occurs when measurements vary unpredictably around the true value, often due to fluctuations in the measurement process or the characteristics of the participant.

Self-report measures are particularly susceptible to measurement error. Participants may misremember past events, exaggerate or minimize certain behaviors, or provide answers they believe are socially acceptable. A study of writing motivation, self-efficacy, and anxiety among primary school students used a mixed-methods approach to address these concerns. The researchers collected quantitative survey data from 436 students and supplemented this with qualitative interviews with classroom teachers who observed the students' writing processes. The teachers' observations provided a check on the accuracy of the students' self-reported experiences.

Strategies for Reducing Measurement Error

Researchers should use validated measurement instruments whenever possible. Validation studies provide evidence that an instrument measures what it claims to measure and that it produces consistent results across administrations. When no validated instrument exists, researchers should conduct pilot testing to assess the clarity and reliability of newly developed measures.

Training data collectors and using standardized protocols can reduce observer error. Blinding data collectors to the study hypothesis or the treatment assignment can prevent conscious or unconscious bias in measurement.

The EQUATOR Network maintains a comprehensive collection of reporting guidelines for health research. These guidelines specify the information that should be reported about measurement methods, including details about instrument validation, data collector training, and quality control procedures. Following these guidelines helps readers assess the potential impact of measurement error on study findings.

Study Design Limitations

The choice of study design imposes inherent limitations on the types of conclusions that can be drawn from the data. Cross-sectional studies provide a snapshot of a population at a single point in time, which means they cannot establish the temporal sequence between exposure and outcome. Cohort studies follow participants over time and can establish temporal sequence, but they are subject to attrition and require substantial resources.

Cross-Sectional Design Constraints

A cross-sectional study of learning style preferences among clinical students at a Nigerian medical school demonstrates the inherent limitations of this design. The researchers surveyed 200 students using the VARK inventory and identified the distribution of learning style preferences in this population. The findings describe the learning preferences of these students at one point in time, but they cannot show how these preferences develop or change over time. The cross-sectional design also prevents the researchers from determining whether learning style preferences influence academic performance or whether academic experiences shape learning preferences.

Experimental Design Considerations

Experimental designs offer stronger evidence for causal relationships than observational designs, but they have their own limitations. Laboratory experiments may not reflect real-world conditions, while field experiments may sacrifice control over extraneous variables. Quasi-experimental designs, which lack random assignment, are subject to confounding by unmeasured variables.

A study of AI tools for improving argumentative writing skills among engineering students in India used a quasi-experimental mixed-method design. The researchers compared students' writing performance before and after an intervention using DeepL Write and Claude AI. Because the study lacked a control group and random assignment, the researchers could not definitively attribute the observed improvements to the AI tools instead of to other factors such as practice effects or increased motivation.

Strategies for Addressing Design Limitations

Researchers should select the strongest design feasible given their research question, resources, and ethical constraints. When a weaker design is necessary, the limitations should be acknowledged explicitly and the findings framed appropriately.

The cross-lagged panel model is commonly used in addiction research to examine bidirectional relationships between variables over time. However, this model has been critiqued for not appropriately adjusting for between-person variance. A study using data from Project MATCH compared the cross-lagged panel model with alternative models and found that the cross-lagged panel model provided the most evidence of significant cross-lagged paths but the poorest fit to the data. This example illustrates how the choice of analytical model can affect substantive conclusions and why researchers should consider multiple analytical approaches.

Qualitative Research Limitations

Qualitative research methods provide rich, detailed accounts of human experiences and social processes, but they have distinct limitations that researchers must acknowledge. The goals of qualitative research differ from those of quantitative research, and the criteria for evaluating qualitative studies differ accordingly.

Generalizability in Qualitative Research

Qualitative studies typically involve small numbers of participants who are selected purposefully instead of randomly. The goal is depth of understanding instead of statistical generalizability. A study of narrative interviewing as a method for understanding people's experiences of taking antidepressants illustrates this point. Narrative interviews place the person being interviewed at the heart of the research and allow them to control the direction, content, and pace of the interview. However, narrative research does not set out to be generalizable and may involve only a small set of interviews.

Methodological Rigor in Qualitative Design

Qualitative descriptive design is widely used in nursing and health science research, but its flexibility presents challenges for achieving methodological rigor. A worked example of qualitative descriptive design published in the Journal of Advanced Nursing explains that this design often borrows methods from other qualitative traditions. Novice researchers may conduct studies using a mix and match of methods without giving proper attention to the principles of qualitative descriptive design. The flexibility of the design can limit the scope of research and affect the quality and impact of the findings.

Strategies for Strengthening Qualitative Research

Researchers should select a qualitative approach that aligns with their research question and apply the methods consistently with the principles of that approach. The Framework Method provides a systematic approach to managing and analyzing qualitative data in multi-disciplinary health research teams. This method is appropriate for research teams where not all members have previous experience of conducting qualitative research, and it can be used effectively with the leadership of an experienced qualitative researcher.

Triangulation, or the use of multiple data sources or methods, can strengthen the credibility of qualitative findings. A study of oral communication strategies for Deaf and Hard of Hearing students used classroom observations, student questionnaires, and semi-structured interviews with teachers and educational specialists. This multi-perspective approach enabled the researchers to triangulate learner, classroom, and professional perspectives.

Data Collection Problems

Problems in data collection can compromise the quality of research findings regardless of the study design. These problems include missing data, inconsistent data collection procedures, and difficulties in recruiting or retaining participants.

Missing Data

Missing data can arise from participant nonresponse, equipment failure, or errors in data entry. The pattern of missingness affects how researchers should handle missing data in their analyses. When data are missing completely at random, the remaining data are still representative of the full sample. When missingness is related to observed variables or to the missing values themselves, analyses that ignore the missing data can produce biased results.

Data Collection Procedure Issues

Inconsistent data collection procedures can introduce systematic differences between study groups or study sites. Researchers should develop detailed protocols for data collection and train all data collectors to follow these protocols consistently. Regular monitoring of data collection procedures can identify deviations early and allow for corrective action.

The Research Data Framework from the National Institute of Standards and Technology emphasizes the importance of documenting data collection procedures and any deviations from the planned protocol. This documentation supports the transparency and reproducibility of research findings.

Strategies for Preventing Data Collection Problems

Pilot testing data collection instruments and procedures can identify problems before the main study begins. Researchers should develop clear definitions for all variables and provide data collectors with written instructions and decision rules for handling ambiguous situations.

Regular data quality checks during the data collection period can identify errors early. These checks may include reviewing data for out-of-range values, inconsistent responses, or unusual patterns that suggest data entry errors or participant misunderstanding.

Bias in Research Methodology

Bias refers to systematic errors in the design, conduct, or analysis of a study that lead to incorrect conclusions. Bias can be introduced at any stage of the research process, and researchers must be vigilant in identifying and minimizing potential sources of bias.

Common Types of Bias

Confounding occurs when a third variable is associated with both the exposure and the outcome, creating a spurious association or masking a true association. Randomization is the most effective method for controlling confounding in experimental studies, but observational studies must rely on statistical adjustment or matching.

Information bias results from systematic errors in measuring exposure or outcome variables. Recall bias occurs when participants who experience an outcome remember past exposures differently than those who do not experience the outcome. Interviewer bias occurs when data collectors treat participants differently based on their exposure or outcome status.

A study of matching-adjusted indirect comparisons in comparative effectiveness research illustrates how cross-trial differences can bias indirect treatment comparisons. When head-to-head randomized trials are not available, researchers may compare treatments across separate trials. However, these analyses may be biased by cross-trial differences in patient populations, sensitivity to modeling assumptions, and differences in the definitions of outcome measures. Matching-adjusted indirect comparisons use individual patient data from trials of one treatment to match baseline summary statistics reported from trials of another treatment, reducing or removing the observed cross-trial differences.

Strategies for Reducing Bias

Blinding, also known as masking, prevents participants, data collectors, and outcome assessors from knowing which treatment or exposure group a participant is in. Blinding reduces the risk of performance bias and detection bias.

Randomization ensures that participant characteristics are balanced across treatment groups on average, reducing the risk of confounding. Stratified randomization can ensure balance on important prognostic factors.

The NC3Rs Experimental Design Assistant provides guidance on randomization and blinding procedures and helps researchers identify potential sources of bias in their experimental designs.

Limitations of Specific Analytical Methods

The choice of statistical methods can introduce limitations that affect the interpretation of study findings. Researchers should understand the assumptions underlying their analytical methods and consider whether alternative methods might be more appropriate.

Cross-Lagged Panel Models

The cross-lagged panel model is commonly used in addiction research to examine bidirectional and dynamic associations between constructs. However, this model has been critiqued for not appropriately adjusting for between-person variance. A study using four waves of data from Project MATCH compared the cross-lagged panel model with alternative models, including the Random-Intercept Cross-Lagged Panel Model and a Latent Curve Model with Structured Residuals. The cross-lagged panel model provided the most evidence of significant cross-lagged paths but the poorest fit to the data. The alternative models found little evidence of prospective within-person associations and more evidence for between-person associations.

This example demonstrates that the choice of analytical model can substantially affect substantive conclusions. Researchers should consider multiple analytical approaches and report the sensitivity of their findings to the modeling choices.

Indirect Comparisons

Indirect comparisons of treatments across separate trials can be biased by cross-trial differences in patient populations, differences in the definitions of outcome measures, and sensitivity to modeling assumptions. Matching-adjusted indirect comparisons use individual patient data from trials of one treatment to match baseline summary statistics reported from trials of another treatment. After matching, treatment outcomes are compared across balanced trial populations.

An important assumption of matching-adjusted indirect comparisons, as in any comparison of nonrandomized treatment groups, is that there are no unobserved cross-trial differences that could confound the comparison of outcomes. Researchers using this method should acknowledge this assumption and discuss its plausibility.

Consensus Methods

The Nominal Group Technique and Delphi Technique are consensus methods used in research directed at problem-solving, idea-generation, or determining priorities. Each method has distinct limitations. The Nominal Group Technique requires participants to personally attend a meeting, which may be difficult to organize and may limit attendance due to geography. The Delphi Technique can take weeks or months to conclude, especially if multiple rounds are required, and may be complex for lay people to complete.

Researchers should consider the research question, the perception of consensus required, and practical factors such as time and geography when selecting a consensus method.

Writing the Limitations Section

The limitations section of a research paper provides an opportunity to demonstrate the researcher's understanding of the study's weaknesses and the steps taken to address them. A well-written limitations section enhances the credibility of the research by showing that the researcher has critically evaluated the study.

What to Include

The limitations section should identify the most important limitations of the study, explain how each limitation might affect the interpretation of the findings, and describe any steps taken to mitigate the limitation. Researchers should avoid simply listing limitations without explaining their implications.

Example phrasing for common limitations includes:

For sample size: "The sample size of 40 participants provided sufficient power to detect the large improvement in writing scores observed in this study, but the sample was too small to support reliable subgroup analyses by gender or prior academic performance."

For generalizability: "The participants were recruited from a single private university in Chennai, India, and the findings may not generalize to students at public institutions or in other geographic regions."

For measurement: "Learning style preferences were assessed using the self-report VARK inventory, which may be subject to social desirability bias and may not capture the full complexity of how students learn."

For design: "The cross-sectional design prevents conclusions about the direction of the association between learning style preferences and academic performance."

What to Avoid

Researchers should avoid apologetic language that undermines the value of their work. The limitations section should be factual and balanced, acknowledging weaknesses without suggesting that the study is fatally flawed.

Researchers should also avoid using the limitations section to make excuses for poor research practices. Limitations that could have been avoided through better planning should be acknowledged as such, and the lessons learned should be described.

Example Limitations Paragraphs

A study of AI tools for improving argumentative writing skills among engineering students in India used a quasi-experimental mixed-method design with 40 first-year engineering students. The limitations section might acknowledge that the lack of a control group prevents definitive conclusions about the causal effect of the AI tools, that the small sample size limits the precision of the estimates, and that the findings may not generalize to students in other disciplines or institutions.

A qualitative study of first-year medical students' cognitive shifts in computer science education interviewed 12 students. The limitations section might acknowledge that the findings represent the experiences of students who completed the course and may not reflect the experiences of students who struggled more severely or withdrew, that the interviews were conducted at a single institution, and that the reflexive thematic analysis reflects the interpretive lens of the research team.

Common Failure Patterns in Research Reporting

Researchers often make predictable errors when reporting limitations in their manuscripts. Recognizing these patterns can help researchers avoid them and can help readers evaluate the quality of published research.

The Checklist Approach

Some researchers list limitations without explaining their implications for the findings. A statement such as "The study had a small sample size and used self-report measures" provides no information about how these limitations affect the interpretation of the results. A more useful approach explains the specific consequences of each limitation.

The Apology Approach

Some researchers use the limitations section to apologize for the study's weaknesses, using language that undermines the value of the work. Statements such as "Unfortunately, the small sample size limits the generalizability of our findings" suggest that the study has little value. A more constructive approach acknowledges the limitation while emphasizing what the study does contribute.

The Defensive Approach

Some researchers minimize or dismiss limitations, arguing that the limitations do not affect the validity of their findings. This approach can appear defensive and may reduce the credibility of the research. A more balanced approach acknowledges the limitation and explains why the findings remain informative despite the limitation.

The Buried Limitation

Some researchers bury important limitations in the middle of a long discussion section, making them difficult for readers to find. A more transparent approach places the limitations section in a prominent location, typically near the end of the discussion, and gives each limitation adequate attention.

Practical Steps for Addressing Limitations

Researchers can take concrete steps throughout the research process to identify, mitigate, and disclose limitations.

During Study Planning

Conduct a formal sample size calculation and document the assumptions used. Select validated measurement instruments and pilot test any newly developed measures. Develop detailed protocols for participant recruitment, data collection, and data management. Consider using the NC3Rs Experimental Design Assistant to identify potential sources of bias in the experimental design.

During Data Collection

Monitor recruitment and retention rates and document the reasons for nonparticipation and attrition. Conduct regular data quality checks to identify errors early. Document any deviations from the planned protocol and the reasons for those deviations.

During Data Analysis

Examine the pattern of missing data and select appropriate methods for handling missingness. Conduct sensitivity analyses to assess the robustness of the findings to analytical choices. Consider alternative analytical models when the primary model makes strong assumptions.

During Manuscript Preparation

Identify the three to five most important limitations of the study and explain their implications for the findings. Describe the steps taken to mitigate each limitation. Use reporting guidelines from the EQUATOR Network to ensure that the manuscript includes all relevant information about the study methods.

Records and Documentation

Maintaining detailed records throughout the research process supports the identification and disclosure of limitations. Researchers should document decisions about study design, participant recruitment, data collection, and data analysis, including the rationale for each decision.

The Research Data Framework from the National Institute of Standards and Technology provides guidance on documenting research data and processes. This framework emphasizes the importance of recording information about data collection methods, data processing steps, and quality control procedures.

Researchers should retain the following types of records:

Recruitment records, including the number of individuals approached, the number who agreed to participate, and the reasons for nonparticipation. Data collection records, including the dates of data collection, the personnel involved, and any deviations from the protocol. Data quality records, including the results of data quality checks and any corrections made to the data. Analysis records, including the analytical methods used, the software and version used, and the results of sensitivity analyses.

Professional Escalation Criteria

Researchers should seek additional expertise when they encounter limitations that they cannot address with their current knowledge and skills. The following situations warrant consultation with a statistician, methodologist, or other expert:

When the sample size calculation requires specialized statistical expertise or when the planned analysis is complex. When the study design involves cluster randomization, adaptive designs, or other advanced methods. When missing data are extensive or the pattern of missingness is complex. When the analytical results are sensitive to modeling choices and the researcher is uncertain which approach is most appropriate. When the researcher is uncertain about the interpretation of the findings or the appropriate framing of the limitations.

The Framework Method for qualitative data analysis is appropriate for use in research teams where not all members have previous experience of conducting qualitative research, but it should be used with the leadership of an experienced qualitative researcher. Similarly, researchers using advanced statistical methods should seek guidance from statisticians with relevant expertise.

Frequently Asked Questions

What is the difference between a limitation and a delimitation?

A limitation is a weakness in the study that may affect the validity or generalizability of the findings. Limitations can arise from the study design, the measurement instruments, the sample, or the analytical methods. A delimitation is a boundary that the researcher deliberately sets to define the scope of the study. Delimitations include decisions about which populations to include, which variables to measure, and which time periods to cover. Delimitations are choices made by the researcher, while limitations are constraints that may be outside the researcher's control.

How many limitations should I include in my paper?

There is no fixed number of limitations that should be included in a research paper. The limitations section should identify the limitations that are most important for readers to understand when interpreting the findings. Most papers include three to five limitations. Including too many limitations can make the study appear weak, while including too few can suggest that the researcher has not critically evaluated the study. Focus on the limitations that have the greatest potential to affect the validity or generalizability of the findings.

Should I mention limitations that I could have avoided?

Yes. Researchers should acknowledge limitations that could have been avoided through better planning or execution. Acknowledging avoidable limitations demonstrates intellectual honesty and helps readers understand the context of the findings. Researchers should also describe what they would do differently in future studies. This information is valuable for other researchers who may be planning similar studies.

How do I write about limitations without undermining my study?

Write about limitations in a factual and balanced way. Explain what the limitation is, how it might affect the interpretation of the findings, and what steps were taken to mitigate it. Avoid apologetic language and avoid dismissing the limitation. Emphasize what the study contributes despite its limitations. A well-written limitations section demonstrates that the researcher understands the strengths and weaknesses of the study and can interpret the findings appropriately.

What is the difference between internal validity and external validity?

Internal validity refers to the degree to which the study findings can be attributed to the intervention or exposure instead of to other factors. Threats to internal validity include confounding, selection bias, and measurement error. External validity refers to the degree to which the study findings can be generalized to other populations, settings, and times. A study can have high internal validity but limited external validity if the participants or setting are not representative of other contexts.

How can I address the limitation of a small sample size in my analysis?

Several analytical strategies can help address small sample sizes. Report effect sizes with confidence intervals instead of relying solely on p-values. Use exact tests when the assumptions of asymptotic tests are not met. Consider Bayesian methods, which can incorporate prior information and may perform better with small samples. Conduct sensitivity analyses to assess the robustness of the findings to different analytical choices. Most importantly, interpret the findings with appropriate caution and avoid overstating the precision of the estimates.

What should I do if I discover a serious limitation after data collection is complete?

If you discover a serious limitation after data collection is complete, document the limitation and its potential impact on the findings. Consider whether the limitation can be addressed through analytical methods, such as statistical adjustment for confounding or imputation for missing data. If the limitation cannot be addressed, acknowledge it clearly in the limitations section and explain its implications for the interpretation of the findings. In some cases, the limitation may be serious enough that the study findings should be presented as preliminary or hypothesis-generating instead of definitive.

How do reviewers and editors evaluate the limitations section?

Reviewers and editors evaluate whether the limitations section demonstrates a critical understanding of the study and its findings. A well-written limitations section identifies the most important limitations, explains their implications, and describes the steps taken to mitigate them. Reviewers are generally more critical of studies that fail to acknowledge obvious limitations than of studies that acknowledge limitations and explain their impact. A thoughtful limitations section can enhance the credibility of the research and increase the likelihood of acceptance for publication.

Related Articles

• Experimental Research Sample
• Experimental Research Sample
• Experimental Research Sample
• Statistical Power And Sample Size
• Statistical Power And Sample Size

References and Further Reading

• Research Data Framework. National Institute of Standards and Technology.
• EQUATOR Network. EQUATOR Network.
• Experimental Design Assistant. NC3Rs.
• NCBI Literature Resources. National Center for Biotechnology Information.
• PubMed. National Library of Medicine.
• Using the framework method for the analysis of qualitative data in multi-disciplinary health research.. BMC medical research methodology, 2013.
• How to use the nominal group and Delphi techniques.. International journal of clinical pharmacy, 2016.
• Qualitative Study.. 2026.
• Matching-adjusted indirect comparisons: a new tool for timely comparative effectiveness research.. Value in health : the journal of the International Society for Pharmacoeconomics and Outcomes Research, 2012.
• Narrative interviewing.. International journal of clinical pharmacy, 2016.
• Thinking.. Nurse education today, 2004.
• Neonatal outreach simulation.. Seminars in perinatology, 2016.
• A Worked Example of Qualitative Descriptive Design: A Step-by-Step Guide for Novice and Early Career Researchers.. Journal of advanced nursing, 2025.
• Evidence on Learning Style Preferences Among Clinical Students in Nigeria Using the Visual, Aural, Read/Write, and Kinesthetic Model: Cross-Sectional Study.. 2026.
• Integrating DeepL Write and Claude AI to enhance argumentative writing competence of engineering students in India.. 2026.
• Enhancing oral competence: Evidence-based strategies for developing foreign language speaking skills in Deaf and Hard-of-Hearing (DHH) learners.. 2026.
• The associations among motivation, self-efficacy, and anxiety of writing skills in primary school students: a mixed-methods approach.. 2026.
• Will the widespread use of large language models in scientific writing undermine scientists' critical thinking?. 2026.
• From memorisation to reasoning: a qualitative study of first-year medical students' cognitive shifts in computer science education.. 2026.
• Limitations of the Lipschitz constant as a defense against adversarial examples. Nemesis/UrbReas/SoGood/IWAISe/GDM@PKDD/ECML, 2018.
• Challenges and Limitations of ChatGPT and Artificial Intelligence for Scientific Research: A Perspective from Organic Materials. Applied Informatics, 2023.
• Experimental designs in management and leadership research: Strengths, limitations, and recommendations for improving publishability. Leadership Quarterly, 2019.
• Limitations of cross-lagged panel models in addiction research and alternative models: An empirical example using Project MATCH. Psychology of Addictive Behaviors, 2021.
• Characteristics, potentials, and limitations of open-source Simulink projects for empirical research. Journal of Software and Systems Modeling, 2021.
• How Do In-Context Examples Affect Compositional Generalization?. Annual Meeting of the Association for Computational Linguistics, 2023.
• Perspective: Fundamental limitations of the randomized controlled trial method in nutritional research: The example of probiotics. Advances in Nutrition, 2018.
• Cohort studies: Design, measures, examples, and limitations. Translational Plastic Surgery, 2026.
• "Applied" research can be useful: An example. Journal of School Psychology, 1976.

This article is educational and does not replace institutional policy, professional advice, or applicable safety and regulatory requirements.