# Sex differences in age-related hearing loss
Pure-tone audiometric thresholds from 4,084 adults screened for known external causes
of hearing loss, collected at the Health Promotion Center, Daegu Fatima Hospital,
Daegu, Republic of Korea, between January 2004 and April 2010.
## Contents
| File | Description |
|---|---|
| `README.md` | This file |
| `hearing_thresholds.csv` | Sex, age, and pure-tone thresholds at six frequencies in both ears; one row per subject |
*(Rename `hearing_thresholds.csv` above to match the filename actually uploaded.)*
## Data file
One row per subject. There is no subject identifier column, and rows are not
linked to any other record. Subjects appear exactly once; repeat testers were
excluded before this dataset was assembled.
Number of rows: 4,084 (910 men, 3,174 women).
### Columns
| # | Column | Type | Units / coding | Description |
|---|---|---|---|---|
| 1 | `Sex` | integer | 0 = female, 1 = male | Sex of the subject |
| 2 | `Age` | integer | years | Age in whole years at the time of testing (18–80) |
| 3 | `R 0.25` | integer | dB HL | Right-ear hearing threshold at 0.25 kHz |
| 4 | `R 0.5` | integer | dB HL | Right-ear hearing threshold at 0.5 kHz |
| 5 | `R 1` | integer | dB HL | Right-ear hearing threshold at 1 kHz |
| 6 | `R 2` | integer | dB HL | Right-ear hearing threshold at 2 kHz |
| 7 | `R 4` | integer | dB HL | Right-ear hearing threshold at 4 kHz |
| 8 | `R 8` | integer | dB HL | Right-ear hearing threshold at 8 kHz |
| 9 | `L 0.25` | integer | dB HL | Left-ear hearing threshold at 0.25 kHz |
| 10 | `L 0.5` | integer | dB HL | Left-ear hearing threshold at 0.5 kHz |
| 11 | `L 1` | integer | dB HL | Left-ear hearing threshold at 1 kHz |
| 12 | `L 2` | integer | dB HL | Left-ear hearing threshold at 2 kHz |
| 13 | `L 4` | integer | dB HL | Left-ear hearing threshold at 4 kHz |
| 14 | `L 8` | integer | dB HL | Left-ear hearing threshold at 8 kHz |
**Notes on values**
- All 12 threshold columns are hearing threshold levels in decibels hearing level
(dB HL), referenced to audiometric zero per ANSI S3.6-1989.
- Thresholds were determined in 5-dB steps, so every threshold value is a multiple
of 5. Observed range: [MIN] to [MAX] dB HL.
- Higher values indicate poorer hearing. Negative values are valid and indicate
hearing better than audiometric zero.
- `Age` has been rounded to the nearest whole year. No date of birth, test date, or
any other identifier is included.
- There are no missing values; all cells are populated.
## Data collection
Audiometry was performed as part of a routine health check. Testing used a Beltone
Model 112 audiometer (Beltone Electronics Corporation, Chicago, IL, USA) with TDH-50P
headphones, in a sound-treated chamber (Model IAC 400A, Industrial Acoustics Company
Inc., Bronx, NY, USA). Subjects pressed a hand-held response key on hearing a tone.
Each ear was tested separately. The better ear was tested first; if the subject
reported equal hearing in both ears, the right ear was tested first. Testing began at
1 kHz and proceeded to 2, 4 and 8 kHz, returned to 1 kHz as a reliability check, and
finished with 0.5 and 0.25 kHz. Thresholds were obtained with three ascending and
descending crossings in 5-dB steps.
## Subject selection
Of 16,884 individuals who voluntarily completed a hearing questionnaire (10,286 men,
6,598 women), subjects were excluded if they:
- were foreign nationals, repeat testers, or had an incomplete questionnaire;
- reported otorrhea in either ear;
- reported use of intravenous antibiotics, diuretics, or chemotherapeutic agents;
- reported head injury;
- reported work in a noisy environment, or military service;
- had asymmetric hearing, defined as a between-ear difference greater than 15 dB at
0.5, 1, or 2 kHz;
- were younger than 18 years (to ensure completion of puberty in female subjects);
- were older than 80 years (no men exceeded 80 years, so women above 80 were excluded
to match the age range across sexes).
4,084 subjects remained.
## Usage notes
This dataset was screened extensively for known external causes of hearing loss and
may therefore serve as a reference range for age-related hearing loss in the absence
of those causes. Two consequences follow for reuse:
- **It does not represent the general population.** The exclusions above, the
health-check setting, and the single-centre recruitment.
In particular, exclusion of military service and noisy occupations
removes a substantial and sex-asymmetric portion of the male population in Korea,
and the sexes were screened by different effective filters.
- **Right and left thresholds are not independent.** Subjects with a between-ear
difference greater than 15 dB at 0.5, 1, or 2 kHz were excluded by design, so the
two ears are more similar than they would be in an unselected sample. Analyses that
treat ears as independent observations will understate uncertainty.
Age is recorded as a single cross-sectional measurement per subject. Differences in
threshold across age in these data are between-person differences and include birth
cohort effects; they are not within-person rates of change.
Portions of these data have appeared in prior publications. Some of the male and female
data were used in Kim et al., *Clinical and Experimental Otorhinolaryngology* (2010),
and all of the male data were used as a control group in Kim et al., *International
Journal of Audiology* (2017).
## Ethics
All methods were approved by the Institutional Review Board of Daegu Fatima Hospital
(DFH09OT032) and were conducted in accordance with the Declaration of Helsinki.
Informed consent was obtained from all participants.
## Conflicts of interest
The author declares no competing interests.
## Citation
Kim, S. (2026). Sex differences in age-related hearing loss (version 1.0.0).
*PhysioNet*. RRID:SCR_007345. https://doi.org/10.13026/xxxxx