People often ask about the differences in soil testing labs, what tests they need for a certain situation and how to interpret the analyses. All those are good questions, but they are irrelevant if the sample is not taken correctly. The results are only as good as the soil sample sent to the lab. I'll describe how to take a good soil sample and why each component of taking the sample is important.
The ideal sampling technique is to pull a soil sample on at least every 10 acres. Large fields would need to be subdivided to accommodate this. On these 10-acre fields, collect soil from 10 to 15 locations, from the surface level down to a depth of 6 inches. If sampling row crops or pecan trees, pull an additional sample in the same manner from a depth of 6 to 12 inches. As they're collected, put the 10 to 15 cores of soil into a clean bucket and thoroughly mix them together. Then, remove about one pint of soil from the bucket for the soil sample. Do whatever you like with the remainder of the soil.
What happens if you deviate from this method? First, let's look at field size. One acre of soil 6 inches deep weighs 2 million pounds. If you follow my sampling method, the sample for 10 acres represents an area of 20 million pounds of soil with a one-pound sample. If the sample represents more than 10 acres, you can calculate how many pounds of soil that one-pound sample represents. There are usually several different types of soil (different textures, depths, etc.) in a large field. Trying to represent too many acres with one sample is probably the most common problem in soil sampling.
Sampling at the correct depth is also important. Usually, most of the phosphorus and potassium in a soil are in the upper 6 inches. Most university soil test calibration work is based on a sampling depth of 6 inches. If the soil is dry and you can only collect a sample 3 inches deep, the levels of phosphorus and potassium will appear higher than they actually are. This can cause you to not apply these nutrients when you actually need them for optimum yield.
On the other hand, if the soil is nice and moist and you can push the sampling probe down 9 inches, the levels of phosphorus and potassium will appear lower than they actually are since the sample is diluted with 3 inches of soil that probably has lower levels of these nutrients than did the top 6 inches. This could cause application of P and K when it wasn't really necessary. The 6- to 12-inch sample is usually needed when you suspect that there is carryover nitrogen from the previous year or if you're dealing with particularly deep-rooted crops.
Problem areas should be sampled separately from the rest of the field. If the problem area is included in the regular sample, it makes the regular sample look a little worse and you still won't know why the problem area is bad.
Remember that the success of a soil-testing program is directly related to how well the samples are collected.