The paper examines major criticisms of AD/HD (Attention Deficit/Hyperactivity Disorder) neurofeedback research using T. R. Rossiter and T. J. La Vaque (1995) as an exemplar and discusses relevant aspects of research methodology. J. Lohr, S. Meunier, L. Parker, and J. P. Kline (2001), D. A. Waschbusch and G. P. Hill (2001), and J. P. Kline, C. N. Brann, and B. R. Loney (2002) criticized Rossiter and La Vaque for (1) using an active treatment control; (2) nonrandom assignment of patients; (3) provision of collateral treatments; (4) using nonstandardized and invalid assessment instruments; (5) providing artifact contaminated EEG feedback; and (6) conducting multiple non-alpha protected t tests. The criticisms, except those related to statistical analysis, are invalid or are not supported as presented by the authors. They are based on the critics' unsubstantiated opinions; require redefining Rossiter and La Vaque as an efficacy rather than an effectiveness study; or reflect a lack of familiarity with the research literature. However, there are broader issues to be considered. Specifically, what research methodology is appropriate for studies evaluating the effectiveness of neurofeedback and who should make that determination? The uncritical acceptance and implementation of models developed for psychotherapy, pharmacology, or medical research is premature and ill-advised. Neurofeedback researchers should develop models that are appropriate to the technology, treatment paradigms, and goals of neurofeedback outcome studies. They need to explain the rationale for their research methodology and defend their choices.