Collection of biological samples is the foundation of genetic studies ranging from estimation of genetic diversity to reconstruction of population history. Sample collections are intended to accurately represent the genetic, biological, ecological, cultural, geographic, and/or linguistic diversity of a particular region or population by providing a small, but representative, set of samples. In this study, we analyze human mitochondrial DNA variation in samples collected using four different sampling strategies to represent the same geographic region. Specifically, samples were collected from a village, a rural area, a regional clinic, and a national university in the governorate of Dhamar in Yemen. All samples were assayed for mitochondrial hypervariable region I DNA sequence variation and data were subjected to standard molecular genetic analyses. Our results suggest that analyses in which individual DNA sequences are explicitly compared or evaluated, e.g. phylogenetic and network analyses, may be more sensitive to sample collection design than analyses in which data are averaged across individuals or are analyzed more indirectly, e.g. summary statistics.