So you’ve got this thing that’s supposed to count every step you take. Based on that, it can tell you how many calories you’ve burned, and therefore maybe how much you can eat for dinner.
But what if it’s wrong? What if you’re a lot more active than it says… or a lot less? Mitesh Patel, too, was wondering about the accuracy of pedometers.
So he and a bunch of his colleagues at the University of Pennsylvania got some of the most popular wearable pedometers and smartphone pedometer apps and put them to the test.
Overall, they report today in the Journal of the American Medical Association, the accuracy of pedometers is good. And Patel recommends them.
“I would say that if you already have a smartphone, there are many free apps that can let you start tracking your step counts within a few minutes,” Patel told me in an email. “If you are selecting a wearable device, most of them are accurate.”
However, some devices work better than others. One miscounted by 20 percent.
To check the accuracy of the pedometers, the researchers asked 14 volunteers to walk on treadmills at three miles per hour.
Each volunteer wore the following:
• On their waistbands: a Yamax Digi-Walker SW-200, a Fitbit Zip and a Fitbit One.
• On their wrists: a Fitbit Flex, a Jawbone UP24 and a Nike Fuelband
• In one pocket: an Apple iPhone 5s running Fitbit, Withings Health Mate and ProtoGeo Oy Moves.
• In the other pocket: A Samsung Galaxy S4 running ProtoGeo Oy Moves
An human observer tallied each step that each volunteer took. When the volunteers reached 500 steps, the researchers compared the number of steps counted by the devices to the number counted by the observer.
The 14 volunteers repeated the experiment once again for 500 steps and twice for 1,500 steps. So each device was tested a total of 56 times.
Below you’ll see the numbers on the accuracy of the pedometers. The “mean” is the average number of steps the device counted. For example, the Digi-Walker SW-200 counted 505.1 steps on average when it should have counted 500.
The “standard deviation” gives you an idea of how widely each device ranged in counting steps from one test to another. The devices with the smallest standard deviations were the most consistent and vice versa.
The 500-Step Trials
|Device/App||Mean Number of Steps||Standard Deviation|
|iPhone 5s Fitbit App||514.2||57.1|
|iPhone 5s Health Mate App||516.4||70.8|
|iPhone 5s Moves App||530.9||56.1|
|Galaxy S4 Moves App||466.3||81.8|
The 1,500-Step Trials
|Device/App||Mean Number of Step||Standard Deviation|
|iPhone 5s Fitbit App||1569||167.5|
|iPhone 5s Health Mate App||1570.7||156.9|
|iPhone 5s Moves App||1581.9||188.6|
|Galaxy S4 Moves App||1423.5||225.7|
By these numbers the Fitbit One and Fitbit Zip essentially tied as the most accurate and consistent devices, both for 500 and 1,500 steps. The Nike Fuelband came in last in both categories and both trials. But most of the others were close enough.
How can you make use of these devices? Let’s say you want to lose some weight. By burning an additional 500 calories a day you could lose a pound a week.
The energy you expend with each step depends on your weight, pace and stride. Many of these apps will estimate that for you, but for most people it’s in the ballpark of 10,000 steps. So it’s doable.