Abstract: Neurolinguistic Programming (NLP) is a relatively new meta-model of human behavioral organization and change. The NLP meta-model states that both verbal and non-verbal communications represent underlying sensory experience. Further, humans will tend to develop and use a predominantly visual, auditory, or kinesthetic sensory system, or primary representational system (PRS), in communicating their experiences with others. One NLP meta-tactic designed to enhance therapeutic influence involves matching a person's PRS with congruent therapeutic suggestions. This study tested the effectiveness of this meta- tactic on one frequently desired therapy outcome -- relaxation. Thirty-six subjects, 18 males and 18 females, were classified as to PRS by the predominant use of visual, auditory, or kinesthetic predicates (verbs, adverbs, adjectives) in response to six reflective questions. The three PRS groups consisted of 12 subjects each, balanced for sex. Individual subjects were then exposed to three different PRS relaxation tapes over a period of three consecutive days. Sequence effects were counterbalanced. A split-plot, repeated measures research design was used to analyze change scores on three response measures: electromyographic recordings taken from the right frontalis muscle; the A-state scale of the State-Trait Anxiety Inventory; and a semantic-differential-type relaxation scale. No main effect differences were found for the PRS groups or PRS relaxation tape factors. One significant interaction between groups and tape conditions on the relaxation scale also proved to be inconsistent with the NLP model. A discussion of methodological and related issues follow. Given the complexities of testing a meta- model such as NLP, the unreliable findings from adjunctive research areas, and the paucity of available research on the NLP model, implications of this study and speculations regarding the overall model are difficult to assess. Disparate research findings are discussed in terms of possibilities for future study.