Minimising Bias in Sampling
The following must be considered:
a) Size and number of samples: The greater the number of samples taken, the greater the probability that sample mean is representative of the community. The size of quadrat should also be appropriate.
b) Random-number generators: Humans are naturally biased. We want to throw a quadrat wherever it looks most interesting. By using a random number generator, you minimize this bias.
c) Counting criteria: These are the parameters used for counting organisms within a quadrat. You need to have firm and consistent rules for the entirety of the sampling process. For example, counting every plant that is even partially in the quadrat, or the stem must be inside the quadrat, even if the foliage isn’t.
d) Calibrating equipment and noting associated precision: We use data loggers when measuring abiotic factors. However, the quality of the data collected will depend on whether the equipment has been serviced and calibrated correctly. Before any sampling occurs, they must be cleaned and checked for accuracy, then cleaned again after the excursion. Remember, if the data is not accurate, we may estimate the wrong tolerance limits for an organism.
