Terminal Digit Preference Share
Measure age-heaping on one terminal digit: the percentage of reported ages ending in the chosen digit relative to a uniform ten percent share across digits.
Description
Measure age-heaping on one terminal digit: the percentage of reported ages ending in the chosen digit relative to a uniform ten percent share across digits.
Measure age-heaping on one terminal digit: the percentage of reported ages ending in the chosen digit relative to a uniform ten percent share across digits.1
When to use Terminal Digit Preference Share
Use this diagnostic to identify suspicious digit preference or age-sex patterns before fitting rates, life tables, or population projections.
- Ages
- Reported single years of age as non-negative integers.
- Terminal digit
- Terminal digit under inspection, from 0 to 9.
How the method works
- Count at digit
- Number of reported ages ending in the chosen digit.
- Total ages
- Total number of reported ages examined.
- Share
- Share of reported ages ending in the chosen digit, in percent.
- Deviation
- Share minus the uniform ten percent expectation, in percentage points.
Limitations and assumptions
These indices are screening diagnostics, not corrections. Their thresholds depend on the stated age window and can also respond to real cohort variation, migration, conflict, or selective undercoverage.
Alternative or Complementary measures
Inspect the underlying age distribution and compare more than one diagnostic, such as Whipple, Myers, or the UN joint score. A graphical population pyramid can reveal cohort patterns that a single index compresses.
References
-
Age heaping — Wikipedia contributors
Similar or alternative tools
- Bachi Age-Sex Accuracy Index Calculator
Compute Bachi's age-sex accuracy index, the sum over both sexes of half the total absolute deviation of terminal-digit shares from ten percent across complete decades of single ages.
- Myers Blended Sex Ratio Index Calculator
Compute Myers' blended index of sex ratios, the sum over terminal digits 0-9 of absolute deviations of blended sex ratios from 100, using overlapping age windows to remove digit-weighting bias.
- UN Joint Age Sex Score
Compute the UN Manual IV joint score for age-sex data quality: three times the sum of absolute successive sex-ratio deviations plus the summed male and female age-ratio deviations, classifying below 20 as accurate, 20 to 40 as inaccurate and above 40 as highly inaccurate.