Free scoping read. A written assessment of your manuscript in two business days, at no cost.

Start one
Contents

Academy · Research Data, Analytics & AI · SD from Range Value Only

Estimating the Standard Deviation (SD) from a Range Value Reported Without the Minimum and Maximum

When a study reports only a single range value, such as ‘range = 14 units’, without the underlying minimum and maximum, the Hozo-style formulas cannot be applied because they require both values directly. In such cases, a control-chart divisor that depends only on sample size provides an alternative estimate of the standard deviation (SD).

Pubrica AcademyWritten by a biostatistics reviewer Reviewed bya meta-analysis methodologist Published10 October 2026 Last checked10 October 2026 Reading time7 minutes Sources10 peer-reviewed

Most SD-estimation methods based on a reported range, including the classic approach of Hozo, Djulbegovic and Hozo, require the individual minimum and maximum values, not only their difference, because these methods also use (min + max) / 2 to estimate the mean. A narrower but genuine extraction problem arises when a source reports only the range as a single number, for instance a range of 14 units of blood pressure change without the specific minimum and maximum readings that produced it. In that situation, the usual range-based formulas cannot be applied, and the only workable alternative is a method that treats the range as the expected spread of n draws from a normal distribution and divides it by a sample-size-specific constant.

Learning objectives

  • Why a bare range value, reported without the minimum and maximum, precludes the standard Hozo-style formulas
  • The control-chart divisor method and its dependence on sample size
  • The scope of the method, which recovers an SD estimate but not a mean

Common misconceptions

  • “Knowledge of the range is equivalent to knowledge of the minimum and maximum for the purpose of estimating SD.” The standard Hozo formula requires the minimum and maximum as two separate values in order to estimate the mean as (min+max)/2. A single range value does not retain that information; a study reporting only ‘range = X’ therefore supports an SD estimate through a different, sample-size-dependent method, but does not support any estimate of the mean from the range alone.

Consequences of losing the individual minimum and maximum

The Hozo, Djulbegovic and Hozo method for estimating a mean and SD from a reported median, range, and sample size relies on the minimum and maximum as separate inputs: it estimates the mean as the midpoint (min + max) / 2 and bases its SD estimate on the distance from that midpoint to each endpoint1. A reported range value on its own, such as "range: 18 mmHg" with no stated minimum or maximum reading, provides none of that endpoint information; it tells a reviewer only how far apart the two most extreme observations were, not where either of them actually sat.

This situation is more common than might be expected, particularly in narrative summaries, conference abstracts, and tables that report a range column without the underlying raw figures, and in older papers that predate the now-common practice of reporting the full five-number summary2. Because the mean cannot be recovered from a range reported alone, the mean must be estimated separately, often from another statistic reported in the same paper, while a distinct, sample-size-based method is used for the SD.

The control-chart divisor method for estimating SD from a bare range

The statistical principle underlying this method is that the expected range of n independent draws from a normal distribution increases in a predictable, tabulated manner as n increases, a relationship long used in statistical process control to estimate the SD of a process from the range of small, repeated samples3. The resulting divisor, often denoted d₂ in quality-control tables, converts a reported range directly into an SD estimate:

SD ≈ range / d₂(n)

Representative values of d₂ increase with n: approximately 1.13 for n = 2, 2.33 for n = 5, 3.08 for n = 10, and 3.47 for n = 15, continuing to rise more slowly as n grows further3. The work of Walter and Yao on the use of reported ranges in systematic reviews applies the same logic, deriving expected-range relationships for a normal distribution that allow an SD to be back-calculated solely from a reported range and sample size, without requiring the mean4.

Increase of the divisor with sample size

That the divisor increases with n may appear counterintuitive, since it means that the range of a larger sample must be divided by a larger number to yield the same SD estimate. The explanation is that the expected range of a sample itself increases with sample size, because a larger sample is more likely to include observations further into the tails of the distribution3. A sample of 5 draws from a normal distribution with SD = 10 will typically show a smaller observed range than a sample of 50 draws from the same distribution, because the larger sample has more opportunities to include an extreme value, even though the underlying SD is unchanged.

This is also why using a single constant such as "divide by 4" regardless of sample size, a rule of thumb sometimes seen in informal guidance, is not a reliable substitute for the sample-size-specific divisor: it happens to be roughly correct only in a narrow band of sample sizes and systematically misestimates SD outside that band35. Where a review includes several studies of varying sample size, the correct divisor for each sample size should be obtained individually rather than one fixed constant applied across the whole dataset.

Scope and limitations of the method

This divisor approach yields an SD estimate only. Unlike the Hozo formula, it provides no route to an estimated mean, because a range reported alone carries no information about where the sample was centred, only about the distance between its extremes14. If a mean-difference meta-analysis also requires a mean for this study, the mean must be obtained elsewhere in the same source, for example a reported median used as a proxy, or the study must be excluded from quantitative pooling and handled narratively, consistent with the general guidance of the Cochrane Handbook on incomplete outcome data6.

The normality assumption underlying the expected-range tables is also a genuine constraint: like the IQR-based methods, this divisor approach assumes that the underlying data are approximately normally distributed, and it will understate the true SD for markedly skewed outcomes in which extreme values occur more often than a normal distribution predicts4. As with every imputed SD discussed across these methods, the resulting figure should be labelled as an estimate in the extraction table and tested in a sensitivity analysis before it is permitted to influence a pooled result materially67.

Worked example using the d2 divisor

Suppose a small pilot study of n=10 reports only "range = 24" for a continuous outcome, with no stated minimum, maximum, or median. Looking up the d₂ divisor for n=10 gives approximately 3.08, so the estimated SD is 24/3.08 ≈ 7.793. If a reviewer mistakenly applied a flat "divide by 4" rule instead, the result would be 24/4 = 6.00, a meaningfully smaller figure that understates the true spread by roughly 23% for this particular sample size, illustrating concretely why a fixed constant is not an adequate substitute for the correct, size-matched divisor35.

By comparison, consider a larger study of n=50 that reports the same range of 24. The corresponding d₂ divisor for n=50 is considerably larger, approximately 4.5 to 4.7 depending on the reference table used, giving an estimated SD of approximately 5.1 to 5.3, noticeably smaller than the n=10 estimate despite the identical reported range34. This is the expected and correct behaviour of the method: the same range observed in a larger sample implies a smaller underlying SD, because a larger sample is statistically more likely to extend that far from its own mean simply by including more draws from the same distribution.

It follows that the raw reported ranges of two studies should not be compared directly as a measure of relative variability, since a larger study will tend to report a larger range than a smaller study drawn from an identical population, as a function of sample size rather than of any real difference in spread37. Converting the range of each study through its corresponding divisor, rather than comparing the raw range figures informally, is the only way to compare implied variability fairly across studies of different sizes.

When this method is applied across many studies, maintaining a reference table of d₂ values for the sample sizes that occur among the review’s included studies, rather than looking up each value individually, accelerates extraction and reduces the risk of assigning the wrong divisor to a study3. Such a reference table, kept alongside the review’s extraction protocol, requires little effort and supports the consistent application of this estimation method across every study in the review to which it applies.

Key points

Key points

A range reported without the minimum and maximum cannot support an estimate of the mean. The SD can, however, be approximated by dividing the range by a sample-size-specific control-chart divisor, d2(n), which increases with sample size and assumes approximate normality.

Practical application

Practical application

  1. Authors should confirm whether the study reports the minimum and maximum individually or only the range as a single value. Where the minimum and maximum are reported separately, the standard Hozo-style formula is appropriate; the divisor method applies specifically to a range reported alone.
  2. The d2 divisor corresponding to the actual sample size of each study should be obtained. A fixed constant, such as division by 4, should not be applied across studies of different sizes, because the appropriate divisor increases with n.
  3. The mean estimate should be obtained from elsewhere in the same study. This method yields an SD only; the mean should be located or estimated separately, or the study excluded from quantitative pooling if no other estimate of the mean is available.
  4. The result should be treated as an approximation and subjected to sensitivity analysis. Studies using this method should be flagged in the data extraction table, and a sensitivity analysis should examine whether their exclusion alters the pooled conclusion.

Frequently asked questions

No. The Hozo formula requires the actual minimum and maximum values as separate inputs and uses them to estimate both a mean and an SD. This divisor method works from a single range value and estimates the SD only, providing no information about the mean.

They come from statistical process control theory, specifically the expected value of the range of n independent draws from a standard normal distribution, a relationship tabulated for use in control charts long before its application to systematic review data extraction.

Because the expected range of a sample increases with sample size: more observations provide more opportunities to include an extreme value, so the larger range typical of a larger sample reflects the same underlying SD as the smaller typical range of a smaller sample, and a larger divisor is therefore needed to recover the same SD.

It assumes approximate normality, like the IQR-based methods; for markedly skewed data, this method will generally understate the true SD because normal-distribution expected-range tables do not account for heavier tails.

References

  1. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample. BMC Medical Research Methodology. 2005;5:13. read 10 October 2026. View on BMC Medical Research Methodology
  2. Wan X, Wang W, Liu J, Tong T. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Medical Research Methodology. 2014;14:135. read 10 October 2026. View on BMC Medical Research Methodology
  3. Walter SD, Yao X. Effect sizes can be calculated for studies reporting ranges for outcome variables in systematic reviews. Journal of Clinical Epidemiology. 2007;60(8):849-852. read 10 October 2026. View on Journal of Clinical Epidemiology
  4. Higgins JPT, Li T, Deeks JJ (editors). Chapter 6: Choosing effect measures and computing estimates of effect. In: Cochrane Handbook for Systematic Reviews of Interventions, version 6.4. Cochrane, 2023. read 10 October 2026. View the Cochrane Handbook, Chapter 6
  5. Deeks JJ, Higgins JPT, Altman DG (editors). Chapter 10: Analysing data and undertaking meta-analyses. In: Cochrane Handbook for Systematic Reviews of Interventions, version 6.4. Cochrane, 2023. read 10 October 2026. View the Cochrane Handbook, Chapter 10
  6. Altman DG, Bland JM. Standard deviations and standard errors. BMJ. 2005;331(7521):903. read 10 October 2026. View on The BMJ
  7. Luo D, Wan X, Liu J, Tong T. Optimally estimating the sample mean from the sample size, median, mid-range, mid-quartile range, and/or range. Statistical Methods in Medical Research. 2018;27(6):1785-1805. read 10 October 2026. View on Statistical Methods in Medical Research
  8. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414):557-560. read 10 October 2026. View on The BMJ
  9. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. read 10 October 2026. View on The BMJ
  10. Mulrow CD. Rationale for systematic reviews. BMJ. 1994;309(6954):597-599. read 10 October 2026. View on The BMJ

Every source above is a peer-reviewed article, a publisher’s record of one, or a standard issued by the body that owns it. Each carries the date it was last read. Where a standard moves, this article is revised rather than left up with a note.

How to cite this article

Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.

Pubrica Academy. (2026, October 10). Estimating the Standard Deviation (SD) from a Range Value Reported Without the Minimum and Maximum. Pubrica. https://pubrica.com/academy/estimation-of-standard-deviation-with-range/

This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.

Bare-Range-to-SD Worksheet

This method applies when a source reports only a range figure, with no stated minimum or maximum.

[ ] Confirmed the source states only a range value, not individual min/max
[ ] Recorded the study's exact sample size n
[ ] Looked up the matching d2 divisor for that n
[ ] Computed SD = range / d2(n)
[ ] Identified a separate source for the mean, or flagged the study for narrative-only handling
[ ] Labelled the SD as imputed and flagged for sensitivity analysis

Pubrica AcademyWritten and maintained by Pubrica’s subject matter experts. Articles are revised when the underlying standards change, and each article states the date on which it was last reviewed.

How Pubrica can help

Pubrica’s data extraction specialists apply sample-size-matched control-chart divisors when a source reports only a range, and separately obtain or flag the mean estimate so that a meta-analysis does not rest on an undeclared assumed value.

Find something else

Search the Pubrica Academy

Access is free and requires no account. Enter a topic, or open the full library to filter by category, sub-topic or resource type.

Support for estimating SD from a range reported without minimum and maximum values

A scoping assessment is free and takes two business days. It provides a written assessment of what the work requires and which specialist would undertake it.

Get a free quote

Regulations

Check your project against this rule

  1. 1Your work
  2. 2Your details
  3. 3Send
What do you need?

Material you send is held under signed confidentiality, with access limited to the assigned team and encrypted transfer. It is not used to train anything and is not passed to third-party tools. Ask for the agreement before you send anything you would rather cover first.