KACD
Articles
- Page Path
- HOME > Restor Dent Endod > Ahead-of print articles > Article
- Statistical Review Article Practical considerations and methods for determining sample size to estimate a population mean: a narrative review
-
Mohamad Adam Bujang,*
-
Restor Dent Endod [Epub ahead of print]
DOI: https://doi.org/10.5395/rde.2026.51.e41
Published online: August 12, 2026
Clinical Research Centre, National Institutes of Health, Sarawak General Hospital, Ministry of Health Malaysia, Kuching, Malaysia
- *Correspondence to Mohamad Adam Bujang, PhD (Sc) Clinical Research Centre, National Institutes of Health, Sarawak General Hospital, Ministry of Health Malaysia, Level 5, Block Clinical Research Centre, Jalan Hospital, 93586 Kuching, Sarawak, Malaysia Email: mohamadadambujang1980@gmail.com
• Received: February 16, 2026 • Revised: April 1, 2026 • Accepted: April 3, 2026
© 2026 The Korean Academy of Conservative Dentistry
This is an Open-Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
- 10 Views
- 0 Download
Abstract
- Sample size determination plays a crucial role in ensuring the validity and precision of research findings. This paper reviews the concept, rationale, and application of the sample size formula for estimating the mean of a population. It outlines the theoretical foundation, practical calculation steps, and appropriate contexts for its use in descriptive and estimation-based studies. This formula is often overlooked, as many researchers focus primarily on hypothesis testing frameworks rather than estimation approaches. Although it is a classical method, it remains highly relevant and essential in contemporary research. Therefore, this paper highlights the importance of this formula across various research settings and provides detailed guidance on how to compute the required sample size accurately. By doing so, it aims to support proper sample size planning, strengthen research design validity, and ultimately enhance the quality and credibility of scientific investigations.
INTRODUCTION
Determining the appropriate sample size is a fundamental aspect of research design as it directly influences the reliability and precision of study findings [1]. One of the earliest and simplest approaches in sample size estimation is the formula for determining the mean of a population. This classical method predates the development of more complex statistical analyses such as the independent sample t-test, paired sample t-test, and one-way analysis of variance. Unlike sample size calculations for hypothesis testing, which focus on detecting differences between groups or effects of interventions, the sample size to determine a population mean primarily serves to achieve a desired level of accuracy and precision in estimating the true mean [1,2].
Despite being one of the oldest and most straightforward sample size formulas, its application is rarely seen in contemporary research literature. One possible explanation is that probably most modern studies place greater emphasis on inferential and modeling techniques aimed at hypothesis testing rather than estimation. As a result, this fundamental approach to sample size determination intended purely for estimation accuracy has been somewhat overlooked, even though it remains highly relevant in studies where the goal is to describe or estimate a parameter rather than to compare or predict.
Therefore, this article aims to review and discuss the concept and application of the sample size formula for determining the mean of a population. It emphasizes when and how this formula should be appropriately applied, especially in descriptive and estimation-based studies. By revisiting this essential yet underutilized approach, the paper seeks to highlight its continuing importance in research methodology and to promote a deeper understanding among researchers of how accuracy-driven sample size estimation contributes to robust and credible scientific findings.
CONCEPT AND CALCULATION
The purpose of determining the sample size to estimate a population mean is to ensure that the sample mean (x̄) provides a sufficiently accurate estimate of the true population mean (μ). Unlike hypothesis testing, which aims to detect significant differences or relationships, this approach focuses purely on estimation accuracy. In estimation, the key concern is how close the sample mean is to the population mean. This closeness is measured by the margin of error (E), which represents the maximum acceptable difference between the estimated mean and the true mean.
The smaller the margin of error desired, the larger the required sample size. This concept is rooted in the central limit theorem (CLT), which states that the sampling distribution of the sample mean approximates a normal distribution when the sample size is sufficiently large, regardless of the shape of the population distribution. Hence, the variability of the sample mean (its standard error) depends on the population standard deviation (σ) and the sample size (n) [2,3]. The general formula for determining the sample size to estimate a population mean is:
where n = required sample size, Zα/2 = Z value corresponding to the desired confidence level (for example, 1.96 for 95% confidence), σ = population standard deviation (or an estimated standard deviation from a pilot study), E = desired margin of error (the maximum acceptable difference between the sample mean and the true mean).
For example, if a researcher wants to estimate the average systolic blood pressure in a population with a 95% confidence level, assuming a standard deviation (σ) of 15 mmHg and a margin of error (E) of 2 mmHg, the required sample size would be:
Thus, at least 217 participants would be needed to estimate the population mean systolic blood pressure within ±2 mmHg of the true mean with 95% confidence.
The confidence level determines how certain researchers want to be that the sample mean will fall within the specified margin of error (E) from the true population mean. It is expressed as 1 – α, where α is the probability of making a Type I error (rejecting a true value). The most commonly used confidence level is 95%, corresponding to α = 0.05, which gives a Zα/2 value of 1.96.
A 95% confidence level means that if the same study were repeated many times, 95% confidence intervals (CI) calculated from those samples would contain the true population means. Some studies, especially those involving high precision such as clinical or laboratory-based measurements, may adopt a 99% confidence level (α = 0.01, Zα/2 = 2.58). Choosing the confidence level involves balancing precision and practicality. A higher confidence level provides greater assurance in the results but increases the required sample size. As a common guideline, a 95% confidence level is widely acceptable.
The population standard deviation (σ) reflects the variability of the measurement within the target population. The ideal approach for estimating σ is to refer to values reported in the literature. Since σ is rarely known beforehand, it is usually estimated from a pilot study, previous research, or an estimated guess by experts [1].
Conducting a pilot study allows researchers to estimate the standard deviation based on real data from the intended population. This approach provides not only practical but also feasible results [1]. When pilot data are unavailable, researchers may refer to similar studies conducted among comparable populations. However, it is important to ensure that the population characteristics, measurement tools, and study settings are closely aligned with the current research [1,3,4]. In a worst-case scenario, researchers can use a conservative (larger) standard deviation to ensure the calculated sample size is sufficient. This approach prevents underestimation, which could result in inadequate precision [1].
Using pilot studies to estimate parameters for power calculation is not favored in some research settings [5-7]. However, the dilemma is that researchers can only determine the true standard deviation once recruitment is completed and the data have been analyzed. Moreover, identifying an appropriate estimate of the standard deviation from the literature that accurately reflects the same patient characteristics and study scope can be challenging. Besides that, an educated guess may either overestimate or underestimate the value. Obtaining the standard deviation from a pilot study provides a more feasible basis, and this approach is supported by the CLT [7,8]. Therefore, this study proposes that researchers add an additional 1% to 2% to the standard deviation estimated from a pilot study to minimize the risk of underestimation. The accuracy of σ is crucial because the sample size increases proportionally with σ². Consequently, an inaccurate or underestimated σ may result in an insufficient sample size and unreliable estimates.
On the other hand, the margin of error (E) represents the maximum acceptable difference between the sample mean and the true population mean. It directly reflects the precision that researchers aim to achieve. In other words, smaller margins of error indicate researchers aim to provide more precise estimates, but eventually require larger sample sizes to prove it.
The choice of margin of error should be guided by proper consideration. The ultimate reason is based on clinical or practical relevance [1,3,4]. In clinical research, the margin of error should reflect the smallest difference that is meaningful or clinically acceptable. For example, when estimating the mean glycated hemoglobin (HbA1c) level for a population, what would be the reasonable margin of error in this case? Well, the clinical and practical relevance are actually associated with the measurement scale of the variable (ie, HbA1c) and feasibility considerations such as cost, time, and resources. Let us first discuss the aspect of the measurement scale.
The measurement scale for HbA1c is relatively narrow, feasibly ranging from approximately 3.0% to 20.0%, with the majority of healthy individuals falling between 5.0% and 6.0% [9,10]. A small change of 1%–2% in HbA1c can shift an individual’s classification from normal to abnormal. Therefore, setting a margin of error of 5% would be unreasonable in this context. For example, if the population mean HbA1c is 6.0% with a 5% margin of error, the 95% CI would range from 1.0% to 11.0%, which is clinically unrealistic, spanning values associated with hypoglycemia, normal glucose levels, and severe hyperglycemia. This highlights why a margin of error of 1%–2% is more appropriate for HbA1c measurements. In contrast, for physiological parameters with a wider measurement range, such as systolic blood pressure (approximately 70–200 mmHg), a 5% margin of error is acceptable and meaningful.
Therefore, the margin of error should be set at a reasonable level, neither too wide that it compromises the accuracy and interpretability of the findings, nor too narrow that it imposes an unnecessary burden on researchers by requiring an excessively large sample size. In practice, determining an appropriate margin of error can be somewhat arbitrary. As a general rule of thumb, allowing a deviation of approximately 5% from the estimated mean (x ¯ x ¯ x ¯
DISCUSSION
Next, the review discusses the potential applications of this formula in scientific studies, with the aim of guiding researchers on when to appropriately apply it in their research. At least, there are seven possible scenarios discussed in this paper.
A common application of this formula is in estimating baseline data, which represent the initial measurements before any intervention, exposure, or treatment. For instance, in clinical research, determining the mean blood pressure, HbA1c level, or blood markers among patients either with or without prior therapy provides essential information about their initial health status [11–13]. By applying this formula, researchers can ensure that the baseline estimates truly represent the population under study. Accurate baseline data not only improve the validity of subsequent analyses but also facilitate meaningful comparison across time or treatment groups [11,12]. Insufficient sample size at this stage could lead to unstable or biased baseline estimates, which would compromise the interpretation of intervention effects.
Another key application is in generating normative or reference data, which serve as benchmarks for interpreting individual or patient measurements [14,15]. This is particularly important in clinical, biomedical, and psychological research. For example, determining the mean fasting blood glucose or hemoglobin concentration in healthy adults helps establish normal reference ranges [16,17]. Applying the sample size formula ensures that these reference values are derived with adequate precision. Such normative data are essential for diagnosis and decision-making, as deviations from the norm can indicate disease or dysfunction [18]. Therefore, the accuracy of normative data directly depends on the appropriateness of the sample size estimation.
The same principle applies in quality control studies, where researchers or practitioners aim to monitor and evaluate the average performance of a process [19]. For example, in clinical laboratories, this formula can be used to estimate the mean concentration of a reagent or the average yield of a manufactured product within an acceptable error range [20]. Ensuring that the mean output lies within tolerance limits is vital for maintaining reliability and compliance with standards.
In environmental and public health research, accurate estimation of mean values is critical for surveillance and policy decisions [21,22]. The formula can be used to estimate the mean concentration of pollutants, nutrients, or contaminants such as PM2.5 in air or heavy metals in water [23]. A properly determined sample size ensures that these environmental estimates are representative of actual population exposure levels, reducing uncertainty in health risk assessments.
In biomedical engineering or diagnostics, this formula is useful in assessing the accuracy and precision of instruments [24]. When validating a new device, such as one used to measure weight, researchers often need to estimate the mean reading under standardized conditions [25]. By determining the sample size using this formula, they can ensure the average reading reflects the true performance of the instrument within a desired level of precision.
Healthcare and management studies often require the estimation of mean costs, service utilization, or duration. For example, estimating the average cost per patient or the average length of hospital stay helps planners make informed decisions about budgeting and resource allocation [26]. Applying this formula ensures that such averages are reliable and useful for strategic or policy-level decision-making.
In early or exploratory studies where no formal hypothesis has been developed, researchers may simply wish to understand the central tendency of a new variable, such as a novel biomarker and physiological parameter [27,28]. The formula provides a structured way to determine an adequate sample size to estimate the mean with confidence. The resulting data can guide further hypothesis-driven research and model development.
IMPORTANT NOTES
Out of the seven scenarios discussed, some can directly apply the formula for estimating a population mean in the study proposal or protocol, particularly when the primary objective is to determine the mean value of a variable in a population. An example includes determining reference parameters, such as blood biomarkers or genomic markers, in specific or healthy populations. However, other scenarios may not require this formula to be stated in the sample size justification, especially when the primary interest lies in hypothesis testing, such as comparing mean differences, examining associations, or evaluating diagnostic accuracy.
For instance, when calibrating instruments to assess agreement for a numerical variable, researchers should use a sample size formula based on the intra-class correlation coefficient. In such cases, the formula for estimating a population mean may serve as a secondary approach to assess the instrument’s precision in estimating true values. A similar principle applies in cost-effectiveness and resource-planning studies. When the primary focus is hypothesis testing (eg, comparing cost outcomes between groups using an independent-samples test), the corresponding hypothesis-based sample size formula should be used. Nonetheless, the formula for estimating a mean can still be valuable for determining the minimum number of participants needed to estimate key continuous outcomes such as cost or length of hospital stay with adequate precision.
PRACTICAL SCENARIO
Now, let us discuss the implementation of the sample size formula to determine the mean for a population. Say a researcher aims to determine the mean fasting blood glucose level among healthy adults in a specific population to establish normative (reference) data. The population is defined as healthy adults above 18 years old in a specific district, Kuching (one of the districts in Sarawak, Malaysia). The objective is to estimate the average fasting blood glucose level (in mg/dL) with a 95% confidence level, ensuring that the sample mean does not differ from the true mean by more than ±5 mg/dL. The steps for sample size calculation are guided by a guideline introduced in a previous study [1].
Step 1. To understand the objective of the study; Determine the formula
The goal is to estimate the population mean with an acceptable level of accuracy rather than to test a hypothesis.
Step 2. To decide the appropriate statistical analysis
Hence, the sample size formula for estimating a mean is used.
Step 3. To estimate or calculate the sample size; Apply the formula
• Alpha (α) = 0.05; thus, confidence level (1 – α): 95% → Zα/2 = 1.96
• Estimated population standard deviation (σ): Say from a previous study, the standard deviation of fasting blood glucose in healthy adults is about 12 mg/dL
• Desired acceptable margin of error (E): 5 mg/dL
Based on calculation, n = 22.13 ≈ 23
Thus, the required minimum sample size is approximately 23 participants.
Step 4. To provide additional allowance to cater for the possibility of a nonresponse rate
In real research, it is advisable to inflate the sample size to account for possible missing data or measurement errors. If a 10% nonresponse or unusable data rate is anticipated: 23/0.9 = 25.56 ≈ 26.
Step 5: To write the sample size statement
The sample size for this study was determined based on the five-step approach recommended in a previous study [1]. As the objective was to estimate the population means with acceptable precision, the formula for estimating a single population mean was selected. The required parameters were specified, such as a significance level of α = 0.05 (95% confidence level), an estimated population standard deviation (σ) of 12 mg/dL obtained from a previous study, and a desired margin of error (E) of 5 mg/dL. Applying these inputs yielded a minimum required sample size of 22.13 or 23 participants. To allow for a 10% potential nonresponse or incomplete data rate, the calculated sample size was adjusted to 25.56, which was rounded up to 26 participants.
A minimum sample size of 26 participants is sufficient based on the initial calculation. This estimation assumes that the study population is homogeneous, meaning that biological characteristics such as sex, age, and ethnicity do not significantly influence fasting blood glucose levels. However, this assumption must be justified with scientific evidence. If there is reason to believe that these characteristics may affect the outcome, researchers should adjust the sample size accordingly.
If such characteristics (eg, age, sex, and ethnicity) are expected to influence the measurements, the population may be divided into distinct subgroups. For example, using three age categories (18–45 years, 46–60 years, and >60 years), two sex groups (male and female), and three ethnic categories (Malay, Chinese, and other Sarawakian ethnic groups), a total of 18 strata would be created (3 × 2 × 3). In this case, a minimum of 26 participants per stratum would be required, resulting in a total sample size of 26 × 18 = 468 participants. Conversely, if evidence suggests that ethnicity does not influence fasting blood glucose levels, stratification may only be required for age (three categories) and sex (two categories), resulting in six strata. In this scenario, the total sample size would be 26 × 6 = 156 participants, with a minimum of 26 individuals in each stratum.
The previous calculations are correct based on a desired margin of error of 5 mg/dL. A larger sample size is required if a smaller margin of error is applied. For instance, reducing the margin of error from 5 mg/dL to 1 mg/dL increases the sample size from 23 to approximately 553.2, or 554 participants after rounding, before accounting for a 10% nonresponse rate. This reflects the principle that higher precision requires a substantially larger sample size. This rationale is supported by previous research, which suggests that estimates derived from sample sizes exceeding 500 participants are generally accurate in representing population parameters [29,30].
Overall, researchers must clearly define the scope and objectives of the study and determine whether stratification by specific characteristics is scientifically warranted. It is also important to recognize that the sample size formula provides the minimum number of participants needed to achieve adequate precision. However, the approach used to obtain this sample, including the recruitment strategy and sampling technique, is a separate methodological consideration that should be carefully planned to ensure validity and feasibility.
RELEVANCE IN MODERN RESEARCH
Although this formula originates from classical statistical theory, it remains highly relevant in contemporary research practice. Its applications extend beyond traditional descriptive studies, where precise estimation of a mean is critical. Despite its broad utility, this formula is often overlooked in modern research because many studies prioritize hypothesis testing as their primary objective rather than precise parameter estimation. As a result, researchers tend to focus predominantly on p-values, which may lead to biased interpretation if not evaluated within the proper statistical context [31,32]. Understanding when and how to apply sample size estimation for mean estimation ensures more reliable conclusions, especially in studies where establishing accurate baseline values or reference ranges is essential.
CONCLUSIONS
This formula continues to serve an essential role in diverse applications. By determining an appropriate sample size based on the desired confidence level, estimated standard deviation, and acceptable margin of error, researchers can ensure that their estimates closely reflect the true population means. Reintroducing and emphasizing this classical formula in right studies and research planning can strengthen methodological practices. It reminds researchers that precision-driven estimation is not merely a statistical exercise but a cornerstone of credible and reproducible science. As research continues to evolve, this foundational approach remains a timeless guide for achieving accuracy and integrity in data-driven disciplines.
-
CONFLICT OF INTEREST
No potential conflict of interest relevant to this article was reported.
-
FUNDING/SUPPORT
The author has no financial relationships relevant to this article to disclose.
-
ACKNOWLEDGMENTS
The author would like to thank the Director General of Health Malaysia for his permission to publish this article.
-
DATA SHARING STATEMENT
The datasets are not publicly available but are available from the corresponding author upon reasonable request.
-
DISCLOSURE OF GENERATIVE AI IN SCIENTIFIC WRITING
ChatGPT (OpenAI) was used solely to improve the grammar, spelling, and language clarity of this manuscript.
- 1. Bujang MA. A step-by-step process on sample size determination for medical research. Malays J Med Sci 2021;28:15-27.ArticlePubMedPMC
- 2. Ahmad WMW, Samsudin NA, Adnan MM, Husein A. Sample size calculations (study design based) using PS software and sampling selection. Pulau Pinang: Penerbit Universiti Sains Malaysia; 2019.
- 3. Sadiq IZ, Usman A, Muhammad A, Ahmad KH. Sample size calculation in biomedical, clinical and biological sciences research. J Umm Al-Qura Univ Appl Sci 2025;11:133-141.ArticlePDF
- 4. In J. Introduction of a pilot study. Korean J Anesthesiol 2017;70:601-605.ArticlePDF
- 5. Teresi JA, Yu X, Stewart AL, Hays RD. Guidelines for designing and evaluating feasibility pilot studies. Med Care 2022;60:95-103.ArticlePubMedPMC
- 6. Kraemer HC, Mintz J, Noda A, Tinklenberg J, Yesavage JA. Caution regarding the use of pilot studies to guide power calculations for study proposals. Arch Gen Psychiatry 2006;63:484-489.ArticlePubMed
- 7. Kwak SG, Kim JH. Central limit theorem: the cornerstone of modern statistics. Korean J Anesthesiol 2017;70:144-156.ArticlePDF
- 8. Zhang X, Astivia OLO, Kroc E, Zumbo BD. How to think clearly about the central limit theorem. Psychol Methods 2023;28:1427-1445.ArticlePubMed
- 9. Mastura I, Chew BH, Lee PY, Cheong AT, Sazlina SG, Jamaiyah H, et al. Control and treatment profiles of 70,889 adult type 2 diabetes mellitus patients in Malaysia: a cross-sectional survey in 2009. Int J Collab Res Intern Med Public Health 2011;3:98-113.
- 10. Teh XR, Yeoh AS, Abdul Samad A, Ismail M, Sivasampu S. Determinants of HbA1c variability among type 2 diabetes mellitus patients in Malaysian primary care setting. Sci Rep 2025;15:25379.ArticlePubMedPMCPDF
- 11. Ramli AS, Selvarajah S, Daud MH, Haniff J, Abdul-Razak S, Tg-Abu-Bakar-Sidik TM, et al. Effectiveness of the EMPOWER-PAR intervention in improving clinical outcomes of type 2 diabetes mellitus in primary care: a pragmatic cluster randomized controlled trial. BMC Fam Pract 2016;17:157.ArticlePubMedPMC
- 12. Khoo SY, Lai WH, On SH, On YY, Adam BM, Law WC, et al. Resting-state electroencephalography (EEG) microstates of healthy individuals following mild sleep deprivation. Sci Rep 2024;14:16820.ArticlePDF
- 13. Hong J, Jalaludin MY, Mohamad Adam B, Fuziah MZ, Wu LL, Rasat R, et al. Diabetic ketoacidosis at diagnosis of type 1 diabetes mellitus in Malaysian children and adolescents. Malays Fam Physician 2015;10:11-18.
- 14. Tan SH, Guan CA, Bujang MA, Lai WH, Voon PJ, Sim EU. Identification of phenomic data in the pathogenesis of cancers of the gastrointestinal (GI) tract in the UK biobank. Sci Rep 2024;14:1997.ArticlePubMedPMCPDF
- 15. Bujang MA, Hon YK, Lai WH, Yap EP, Tiong XT, Ratnasingam S, et al. Health-related quality of life with six domains: a comparison of healthcare providers without chronic diseases and participants with chronic diseases. J Clin Med 2024;13:5398.ArticlePubMedPMC
- 16. Doles N, Ye Mon M, Shaikh A, Mitchell S, Patel D, Seehusen D, et al. Interpretating normal values and reference ranges for laboratory tests. J Am Board Fam Med 2025;38:174-179.ArticlePubMedPMC
- 17. Tirosh A, Shai I, Tekes-Manova D, Israeli E, Pereg D, Shochat T, et al. Normal fasting plasma glucose levels and type 2 diabetes in young men. N Engl J Med 2005;353:1454-1462.ArticlePubMed
- 18. Peña ED, Spaulding TJ, Plante E. The composition of normative groups and diagnostic decision making: shooting ourselves in the foot. Am J Speech Lang Pathol 2006;15:247-254.ArticlePubMed
- 19. Dunbar SB, Koretz DM, Hoover HD. Quality control in the development and use of performance assessments. Appl Meas Educ 1991;4:289-303.ArticlePubMed
- 20. Abdel GM, El-Masry MI. Verification of quantitative analytical methods in medical laboratories. J Med Biochem 2021;40:225-236.ArticlePubMedPMC
- 21. Janes H, Pepe MS, Bossuyt PM, Barlow WE. Measuring the performance of markers for guiding treatment decisions. Ann Intern Med 2011;154:253-259.ArticlePubMedPMCPDF
- 22. Hladik ML, Markus A, Helsel D, Nowell LH, Polesello S, Rüdel H, et al. Evaluating the reliability of environmental concentration data to characterize exposure in environmental risk assessments. Integr Environ Assess Manag 2024;20:981-1003.ArticlePubMedPDF
- 23. Talbi A, Kerchich Y, Kerbachi R, Boughedaoui M. Assessment of annual air pollution levels with PM1, PM2.5, PM10 and associated heavy metals in Algiers, Algeria. Environ Pollut 2018;232:252-263.Article
- 24. Ibrahim F, Thio TH, Faisal T, Neuman M. The application of biomedical engineering techniques to the diagnosis and management of tropical diseases: a review. Sensors (Basel) 2015;15:6947-6995.ArticlePubMedPMC
- 25. Jamaiyah H, Geeta A, Safiza MN, Khor GL, Wong NF, Kee CC, et al. Reliability, technical error of measurements and validity of length and weight measurements for children under two years old in Malaysia. Med J Malaysia 2010;65 Suppl A:131-137.PubMed
- 26. Ruger JP, Richter CJ, Spitznagel EL, Lewis LM. Analysis of costs, length of stay, and utilization of emergency department services by frequent users: implications for health policy. Acad Emerg Med 2004;11:1311-1317.ArticlePubMed
- 27. Petsani D, Konstantinidis E, Katsouli AM, Zilidou V, Dias SB, Hadjileontiadis L, et al. Digital biomarkers for well-being through exergame interactions: exploratory study. JMIR Serious Games 2022;10:e34768.ArticlePubMedPMC
- 28. Denfeld QE, Purnell JQ, Lee CS, Orwoll ES, Camacho SA, Hiatt SO, et al. Candidate biomarkers of physical frailty in heart failure: an exploratory cross-sectional study. Eur J Cardiovasc Nurs 2023;22:149-157.ArticlePubMedPMCPDF
- 29. Bujang MA, Ab Ghani P, Zolkepali NA, Adnan TH, Ali MM, Selvarajah S, Haniff J. A comparison between convenience sampling versus systematic sampling in getting the true parameter in a population: explore from a clinical database: the Audit Diabetes Control Management (ADCM) registry in 2009. 2012 International Conference on Statistics in Science, Business and Engineering (ICSSBE); September 10-12, 2012; Langkawi, Kedah, Malaysia. IEEE; 2012;p. 1-5.Article
- 30. Bujang MA, Sa'at N, Joys AR, Ali MM. An audit of the statistics and the comparison with the parameter in the population. The 22nd National Symposium on Mathematical Sciences (SKSM22); Shah Alam, Malaysia. AIP Conference Proceedings. Vol 1682. AIP Publishing LLC; 2015;p. 050019.Article
- 31. Bujang MA. The dilemma and wisdom in translating p values: a collaborative approach to strengthening scientific validity. Biomed Res Int 2025;2025:6703756.ArticlePubMedPMCPDF
- 32. Bujang MA. A Power Primer revisited. Indian J Psychol Med 2026;48:206-208.ArticlePubMedPMCPDF
REFERENCES
Tables & Figures
REFERENCES
Citations
Citations to this article as recorded by 
ePub Link
Cite
XML DownloadPractical considerations and methods for determining sample size to estimate a population mean: a narrative review
Practical considerations and methods for determining sample size to estimate a population mean: a narrative review
