DXA scanning is considered the gold standard for effective body fat measurements. The DXA scan is able to output decently reliable numbers, but is not absolutely precise and has been known to have error rates in total body fat of 4%  or up to 8-10% [7,8]. However, the price of a scan is not cheap, often costing a couple hundred dollars. The cheap alternative that is often recommended is the skin caliper. Although able to produce a rough estimate of body fat percentage, the skin caliper will not be able to reliably track change in body fat percentage over time.
Skin calipers rely on measuring the thickness of the subcutaneous tissue, the tissue between the skin and muscle, in order to give a body fat measurement. However, the thickness of subcutaneous tissue does not always directly correlate with fat cell size. Changes in muscle thickness can change the geometrical organization of fat cells . Therefore, an increase in muscle thickness of a certain area can cause the thickness of the subcutaneous tissue to decrease, while having no change in actual fat mass. For example, in a 5 week study conducted in 10 healthy women in which they only performed three periods of right knee extensions, there was found to be a significant decrease in adipose tissue thickness in the right leg when measured by ultrasound and skin skin calipers.
One might make the assumption that because thickness decreased, then fat mass also decreased. However, this was not found to be the case. There was found to be NO significant change in average fat cell weight in both the right and left leg. Average fat cell weight was found by using a microscopic method from a biopsy of both legs . It is hypothesized that this inconsistency in fat measurement is caused by a compression of the extra-cellular matrix between fat cells. 
Another study confirmed the inaccuracy of the skin caliper by comparing the measurements with that of an MRI. There was found to be no correlation between the two measurement techniques. This is possibly due to the inability of skin calipers to measure the volumetric changes in subcutaneous fat as opposed to the MRI machine. Skin calipers only measure the distance from the belly of the muscle to the skin . Two other studies had the same conclusion that the two techniques did not agree well. The first only found a weak correlation of (r = 0.205, P < 0.001) while the second found one of (r2 = 0.17, SEE = 2.2, P = 0.040) [4,5]. Although I would not dissuade the casual fitness enthusiast from using skin calipers to track their own progress, I would encourage people to be wary of the studies that use skin calipers as a means of measuring changes in body fat.
Email me @ firstname.lastname@example.org