They base their findings on almost 1000 pedometers, tested out by 35 volunteers between 20 and 60 years of age.
Each volunteer was given 30 cheap pedometers as well as a sophisticated automated step count log against which to compare the performance of the gadgets.
They wore five cheap pedometers and the automated step count log each day for a period of six days. They counted the steps recorded daily with each of the pedometers and compared the figures against those of the automated log.
A variation in performance of 10% was considered acceptable, but only one in four of the pedometers fell within this range.
Three out of four either exceeded or fell below 10%. And in more than one in three, the variation was greater than 50%.
And in almost two thirds of these, the pedometers overestimated the actual steps taken.
This is important, say the authors, because an error of 20% in 10,000 daily steps adds up to 2 000 steps, so either 8 000 or 12 000 steps will be recorded.
Pedometers have become very popular as a cheap and easy way of boosting fitness or losing weight, they say. "The wide accessibility of pedometers needs encouragement," they say.
But they warn: "Inexpensive [pedometers] provide incorrect information on step counts, which makes them inappropriate for physical activity promotion targets." And they suggest that a quality kite mark would be helpful for consumers and patients.
British Journal of Sports Medicine