This document outlines JMIR Publications’ expectations for reporting statistical results in manuscripts. Statistical reporting should describe results, not reinterpret or extend them.

All statistical reporting should follow these principles:

• Clarity: Results should be easily interpretable by readers.
• Consistency: Values must match across all manuscript sections.
• Transparency: Key statistical details should not be omitted.
• Accuracy: Reported values must reflect the underlying data.

For details on common statistical terms, please refer to the AMA’s Glossary of Statistical Terms. The SAMPL Reporting Guideline is a useful resource for best practices in statistical reporting.

1. Consistency and accuracy across manuscript

Statistical values must be consistent across the abstract, main text, tables, and figures. Common issues noted during production include mismatched values, inconsistent rounding of decimals, or missing data.

Where discrepancies are identified, authors may be asked to verify and correct the source data to ensure alignment across all components of the paper.

2. General guidelines on formatting

Numbers, decimals, and fractions

• Numbers greater than 9999 have a comma to separate thousands, millions, etc (eg, 36,127 or 250,000, but 6500).
• Numerical values less than 1 require a leading zero before the decimal point
  • Exception: The leading zero should not be used for P values, α level, and β values.
• Mixed fractions are not generally used but, on occasion, are used with less formal text, especially for expressions of time (eg, 2½ years).
• Use an en dash (or minus sign) to indicate negative values.

Punctuation

• When reporting results in the main text, related statistics should be placed in parentheses after the result.
• When reporting several results within parentheses, use commas to separate related statistics and semicolons to separate unrelated statistics.
  • Examples: (mean 19, SD 2), (odds ratio 5, 95% CI 0-10; P<.001)
• Avoid using brackets within parentheses when reporting statistics.

Percent signs

• Use the percent sign without a space (18%).
• Repeat the percent sign in ranges (15%-20%).

Ranges

• Use a hyphen to indicate a range.
• If either value in a range is negative, use “to” between the values, rather than a hyphen.
  • Example: odds ratio 1 (95% CI –1 to 3).
  • In tables, if one or more ranges contain a negative value, all ranges in the table should be reported with “to” between the two values to ensure consistency.

Equality and inequality signs

• Do not use spaces around signs of equality (=) or inequality (<, >, ≤, ≥) except in mathematical equations.
• Use the symbols for greater than or equal to (≥) and less than or equal to (≤), rather than an underlined greater than (>) or less than (<) symbol.

Exponents

• When using exponents to indicate very small or large numbers, superscript the exponent; do not precede the exponent with “e” or “^”.
• For more information, refer to the AMA (section 20.3).

Examples:

• 2 × 10⁶
• 10 × 3^–10
• M⋅s⁻¹

Units of measure

• Measurements should be presented in SI units.
• Add a space between the numeral and the unit of measure (eg, 37.5 °C, 5 kg).
• Most units of measure are abbreviated (without expansion) when used with numerals.
  • Exception: The expanded units of measure for time (day, hour, millisecond, minute, month, second, week, and year) should be used in running text.
• Abbreviated units of measure should always be singular (eg, 25 min, 78.5 kg).
• For temperature, use the Celsius scale (°C); convert values in Fahrenheit (°F) to °C.

3. Standard reporting of common statistics

Descriptive statistics

Mean values and associated statistics 

• Report mean values together with the appropriate measure of variability or precision and clearly identify the statistic used (eg, SD, SE, or 95% CI).
• When reporting values derived from the mean and its associated statistic, clearly state the calculation performed.
  • Example: Mean systolic blood pressure was 128 (SD 12) mm Hg.  

Median (IQR)

• Median values must be reported with their IQRs. Preferably, the IQR should be presented as a range reporting the 25th and 75th percentiles, rather than as a single value.
  • Example: The median age was 65 (IQR 64-66) years.

Categorical data: n/N (%)

• N designates the entire population under study. n designates a sample of the population under study. This rule also applies to flowcharts illustrating sample selection.
• Percentages (%) derived from the study sample should be reported with clear numerators (n) and denominators (N).
• Percentages should be rounded to a consistent number of decimal places.
• Do not add a zero after the decimal point of a whole number (eg, 50/200=25%, not 25.0%).
• See AMA section 19.4.2 for specific rules on rounding.

The following examples show the preferred styles for reporting n values and percentages in the main text:

• Of 200 participants, the majority (n=150, 75%) felt that…
• Of 200 participants, 150 (75%) felt that…
• Of 200 participants, 75% (n=150) felt that…
• The majority of participants (150/200, 75%) felt that…

Effect estimates

• Confidence intervals (CIs) must be included for all relevant effect measures (eg, odds ratios, risk ratios, and hazard ratios).
  • Example: odds ratio 3.1 (95% CI 2.2-4.8)
• Confidence limits should be reported as the upper and lower boundaries, with a comma separating the two values
  • Example: mean 30% (95% confidence limit 28%, 32%)

4. P values

Reporting and rounding

• Exact P values should always be reported when available (eg, P=.03, not P<.05).
• P values should be reported to 2 decimal places.
  • Exceptions: (1) when P<.01 (eg, P=.003), (2) when rounding would change the significance level (eg, P=.048), or (3) when the study’s significance level is lower (eg, P<.0004), which is rare and only relevant for studies with very large sample sizes.
• P values cannot equal 1 or 0. P values less than .001 should be reported as “P<.001,” and P values greater than .99 should be reported as “P>.99.”
• Use of asterisks in tables to indicate significance levels (eg, *P<.05, **P<.01, ***P<.001) is discouraged. Exact P values should be reported whenever possible/available.

“Trending” toward significance

• Avoid phrases like “There was a trend (P=.06) showing that...was significant.”
• Instead, state if there was a trend, followed by acknowledging that the results were not statistically significant, or clearly state the results’ significance without using terms like “trending.”

5. Formatting of other common statistical measures

The following table shows how to properly report common statistics and format test names.

Other common test names

Formatting rules for additional statistical tests are specified below:

• F₁-score: italicize F, place 1 as a subscript, use a hyphen before “score,” and start “score” with a lowercase “s”.
• I² for heterogeneity: use an uppercase italicized I.
• Type I error and type II error: do not use the numerals 1 and 2; AMA section 18.7.5.
• R²: uppercase italicized R; do not italicize the superscript.
• z score: lowercase italicized z, without a hyphen.

6. Equations

• Use spaces between all mathematical operators in complex equations (including “=”).
• In the text, whenever possible, characters in equations should be inserted using the Advanced Symbols feature of Microsoft Word.
• Equations that are kept in text should be indented and numbered (with the number in bold and in parentheses after the equation itself) if they meet any of the following criteria: (1) there are 3 or more equations, (2) the equations are related to each other, or (3) the equations are referred to after initial presentation (eg, “As per equation 3, we recalculated...”).
  • Example: y_i = C_i – c_i (1)

7. End note

Authors should always ensure that data are presented accurately and consistently. While noting issues in statistical reporting is a responsibility of the editor/section editor—who should point out missing or incorrectly reported statistics during the review process—copyeditors act as the “second line of defense” and must enforce our guidelines on reporting statistics. Therefore, you may be asked to revise or reformat statistics during the production process.

For any questions related to the reporting of statistics, please contact copyed-support@jmir.org.