One of the common challenges of disability research is the small number (n) of subjects/participants available. Furthermore, researchers often have to break down their samples into separate groups due to participants' unique characteristics. Conventional research design and statistical methods cannot function appropriately in small-n disability research due to a lack of statistical power or a violation of the assumption of statistical methods. Fortunately, progresses have been made in the past two decades in research design and statistical methods when the n is small. The single most effective way to increase statistical power to detect treatment effect in randomized designs, for example, is to employ a within-subjects design, which reduces error variance substantially (Venter & Maxwell, 1999). When a between-subject design has to be employed, other strategies can be adopted to increase the statistical power, such as adding a pre-test data collection or adopting a two-tiered approach to adjust systematic individual differences and measurement error statistically (Kraemer & Thiemann, 1989). For small-n data analysis, newly developed resampling methods, known also as computer intensive methods, provide useful alternatives. Bootstraping (Zhu, 1997), permutation/randomization tests (Good, 1994) and Monte Carlo (Noreen, 1989) are some good examples. After a review of general principles in statistical power analysis and effect size, this presentation will describe, in great detail, research design and statistical analysis with small n, including meta-analysis of single-case designs and confirmatory factor analysis. Practical examples and related software application will be illustrated and introduced during the presentation. Keyword(s): adapted physical activity, measurement/evaluation, research