This review examines the effect of a dual task on the gait parameters of older adults with a mean gait speed of 1.0 m/s or greater, and the effect of type and complexity of task. A systematic review of Web of Science, PubMed, SCOPUS, Embase, and PsycINFO was performed in July 2016. Twenty-three studies (28 data sets) were reviewed and pooled for meta-analysis. The effect size on seven gait parameters was measured as the raw mean difference between single- and dual-task performance. Gait speed significantly reduced with the addition of a dual task, with increasing complexity showing greater decrements. Cadence, stride time, and measures of gait variability were all negatively affected under the dual-task condition. In older adults, the addition of a dual task significantly reduces gait speed and cadence, with possible implications for the assessment of older people, as the addition of a dual task may expose deficits not observed under single-task assessment.