Every so often a farmer wants to compare the soil test results from two separate labs. A sample is collected from a few spots in the field, mixed in a bucket, and then a portion of the soil is sent to two different labs for comparison. When samples a re split between labs in this manner, there is already some error built in so that the results will likely be quite different.
The Clemson University Agricultural Service Lab is annually involved with a regional soil exchange involving approximately 7 state supported labs and 5 privately owned labs. Before the samples are distributed to all these labs for analysis, the bulk s ample is first completely air dried to insure thorough mixing, run through a grinder that will completely break up the clods and sieve out any rocks or organic debris, and then it is mixed in a large container for several minutes. When the individual sam ples are removed, care is taken that there is no settling of the larger soil particles in the container. Then a number of samples are removed and analyzed by one lab to ensure that the mixing was thorough. All these steps are taken to be sure that the s ample is as thoroughly mixed and uniform as possible. These procedures are a far cry from simply taking two portions from a soil mixed in a bucket. Mixing soil that is moist or too dry is difficult and studies have shown that samples split in this manne r will be significantly different even when measured by the same lab.
After the exchange samples are analyzed by the various laboratories, the results are grouped by the type of soil extractant used for a proper comparison (there are currently three used in the Southeast). Even the results from samples extracted by labo ratories which use the same extractant solution will usually differ by a few pounds-per-acre due to differences in methodology. This is a normal and expected occurrence which you can see on the following table.
| Sample # | Lab A | Lab B | Lab C | Clemson | Lab D | Lab E | Average |
|---|---|---|---|---|---|---|---|
| 1 | 35 | 21 | 38 | 32 | 32 | 24 | 30 |
| 2 | 32 | 16 | 45 | 32 | 35 | 22 | 30 |
| 3 | 58 | 45 | 51 | 59 | 56 | 34 | 51 |
The rating system which indicates if an extracted level of a nutrient is high, medium or low will differ from lab to lab. Even if two labs extract the same amount of a nutrient using the same extractant, one lab may call that amount medium while anoth er lab will indicate that it is a high level for that nutrient.
The fertilizer recommendations may also vary for each lab. The recommendations are ideally derived from Experiment Station data specific for a particular region's climate and soils. Some labs will use data from adjacent states if they don't have data from their own Experiment Station for a particular crop. Some private labs use data from universities.
So, if a farmer gets results back from two separate labs, they may differ because the sample sent to the labs were not mixed properly, the labs may be using different extractants, they may be using a different rating system for the nutrient levels, or they may be following different guidelines for their recommendations. Ultimately, with only two sets of results for comparison, it is extremely difficult to say which is best to follow. Many farmers will choose the lab results that show what they would like to see even though it may not be the most correct for their situation. It is best to stick with one laboratory and develop a long term relationship with their services as many have done with Clemson's Lab. If you have any questions about a sample, the labs will be happy to re-test the sample.
Back to the main page.