How Reliable is Haphazard Sampling

Current audit standards sanction the use of both statistical and nonstatistical sampling methods. Based on results of a recent survey it appears that nonstatistical sampling methods predominate with haphazard sampling being the method of choice for most audit sampling applications. Two recent accounting research studies tested the composition of haphazard samples and found them to exhibit multiple selection biases. One study found that increasing sample size was not an effective procedure to eliminate this selection bias. Based on these two studies and research reported in the psychology and biology literatures it appears that automatic subconscious process may preclude the selection of unbiased samples when using haphazard selection. Given these findings auditors should consider the use of random or systematic sampling rather than haphazard sampling.