Thus, with this design, not only is every unit in the population not eligible for inclusion in the sample, but the composition may be affected by the personal biases of the researcher as to who he/she believes should be interviewed.