Facial width-to-height ratio (fWHR) is calculated by dividing the distance between your left and right cheekbones by the distance from your upper lip to the midpoint between your brows. A typical result falls somewhere around 1.85 to 2.05, with most people landing close to 1.95. The measurement itself is straightforward, but small errors in head position, camera setup, or landmark placement can throw the ratio off significantly.
The Four Landmarks You Need
fWHR relies on just two measurements, defined by four anatomical points on the face. For width, you measure from the left zygion to the right zygion. These are the most lateral (outermost) points on your cheekbones. You can feel them by pressing gently along your cheekbone arc until you find the widest point on each side.
For height, the upper boundary is the nasion: the midline point where the bridge of your nose meets your forehead, roughly between your brows. The lower boundary is the labiale superius: the midline point on the vermilion border of your upper lip, right at the base of the philtrum (the vertical groove above your lip). Some researchers use alternative lower boundaries, like the midpoint where the lips meet (stomion) or even the bottom of the chin, but the standard and most widely cited version uses the upper lip line.
The Formula
Once you have both distances, the calculation is simple:
fWHR = bizygomatic width ÷ upper facial height
So if the distance between your cheekbones is 142 mm and the distance from your upper lip to the nasion is 73 mm, your fWHR would be 142 ÷ 73 = 1.95.
Measuring Directly on the Face
If you’re measuring on a live person rather than a photo, the standard approach uses a sliding caliper with precision to 0.01 mm. The person should sit upright in a relaxed position with their head in the “Frankfurt horizontal plane,” which simply means looking straight ahead with the bottom edge of the eye socket level with the top of the ear canal. This keeps the head neutral and avoids any tilt that would distort the numbers.
Place the caliper tips on the outermost points of both cheekbones to get bizygomatic breadth. Then measure the vertical distance from the nasion down to the center of the upper lip border. Record both values in millimeters and divide.
Measuring From a Photograph
Most fWHR measurements today are taken from photographs. This is where technique matters most, because camera distance and lens choice directly affect facial proportions.
Research on photographic distortion found that short focal lengths (24 mm, 35 mm, and 50 mm lenses) at close range caused 12% to 19% vertical stretching of the midface. Smartphone cameras produced similar problems: an 18% increase in midface height at 8 inches and a 12% increase at 12 inches. Since fWHR divides width by height, any artificial stretching of height will shrink your ratio and give you a false reading.
The fix is distance. No measurable distortion was found beyond 12 inches from the subject, and researchers typically use photos taken at 5 feet or more as their standard. If you’re using a smartphone, stand at least 3 to 5 feet from the camera and crop afterward rather than zooming in close. A DSLR or mirrorless camera with a focal length of 85 mm or longer at several feet away will produce minimal distortion.
The photo should show the face looking directly at the camera with a neutral expression. The head should not be tilted or turned. Even slight head tilt increases apparent facial asymmetry, with one study finding that 97% of subjects with head tilt showed measurable facial asymmetry compared to only 34% of subjects holding their heads straight. A turned head compresses one cheek and widens the other, making accurate zygion identification nearly impossible.
Software for Photo Measurement
The most widely used tool in published research is ImageJ (currently version 1.54), a free open-source program from the National Institutes of Health. You open your photo, use the line tool to draw from one landmark to another, and ImageJ returns the pixel distance. As long as you use the same photo for both measurements, the pixel distances produce the correct ratio without needing a physical scale. TpsDig2 is another free program commonly used in research, particularly when placing multiple landmarks on a face at once.
Specialized computer vision models trained for facial landmark detection exist and can automate the process, but general-purpose AI tools (like chatbots with image capabilities) perform poorly at precise numerical measurements. For reliable results, stick with manual marking in ImageJ or a dedicated facial analysis program.
Common Measurement Errors
The biggest source of error is inconsistent landmark placement, especially at the zygion. The widest point of the cheekbone is subtle and easy to misjudge by a few millimeters, which shifts the ratio. When multiple research assistants measured the same faces, automatic and manual measurements correlated at r = .86 to .91, meaning they mostly agreed but not perfectly. In at least one case, the difference between manual and automatic measurement changed whether a statistical finding reached significance.
To reduce error, measure each distance two or three times and average the results. If you’re measuring from photos, mark all four landmarks before drawing any lines, so you can evaluate their placement as a set rather than committing to one point at a time. Having a second person independently mark the same photo provides a useful check.
Other common pitfalls include measuring with the mouth open (which shifts the labiale superius downward), using photos where the subject is smiling (which raises the cheeks and shifts the zygion), and using selfies taken at arm’s length (which introduces the midface distortion described above).
Typical fWHR Values
A large study of over 3,400 international athletes found that men averaged an fWHR of 1.98 and women averaged 1.92. The difference is real but small, with considerable overlap between the sexes. Values varied by ethnicity: white men averaged 1.94 while white women averaged 1.89, Black men averaged 2.01 versus 1.93 for women, and East Asian men averaged 2.00 versus 1.98 for women. Asian Indian women actually averaged slightly higher than men in that sample (2.04 versus 1.99), illustrating that the sex difference isn’t universal across populations.
These numbers come from athletes, so they may skew slightly compared to the general population. But they give you a useful reference range: most people fall between roughly 1.75 and 2.20, with values near 1.95 being typical.
What fWHR Actually Tells You
fWHR gained attention in psychology research as a potential signal of dominance and aggression. A 2015 meta-analysis found that people with wider faces relative to their height were judged as more threatening (a moderately strong association) and more dominant. The ratio also predicted actual aggressive and dominant behavior in men, though these effects were small. It correlates moderately with body mass index, meaning heavier individuals tend to have wider-looking faces partly due to soft tissue rather than bone structure alone.
The sex difference in fWHR is statistically significant but modest. The ratio is better understood as one of many facial cues that contribute to social perception rather than as a reliable predictor of any individual’s personality or behavior. Its primary use remains in academic research on face perception, anthropometry, and evolutionary psychology.