EDITORIAL Our Vision for Autism Research One dilemma facing all researchers is the choice of where to submit their next paper. The scientific content of a Journal, its readership, impact factor, the speed of publication and how editors dealt with previous submis- sions are some of the factors that authors consider when choosing a venue for their manuscript. In launching a new publication, editors should convey to potential authors the essence of a Journal in order to attract those papers that will establish its future character. In introdu- cing this second issue we articulate some of the editorial considerations underlying publication priorities. Autism Research was established by the International Society for Autism Research (INSAR) to satisfy the pressing need for a publication outlet for high quality papers dealing with the full range of basic and beha- vioural sciences that underlie research into autism spectrum disorders. The Journal is also an important vehicle for INSAR, whose aim is to advance autism research. The Editors recognize that these goals are well served by publishing both positive and negative findings. Exciting new findings, especially from hypothesis driven research, are the bread and butter of scientific journals, and they are sought for this journal as well. Nevertheless, research into autism is also marked by intriguing findings that have been neither replicated nor refuted, either because replication has never been attempted or because a failure to replicate has not been published. Consequently, in order to support new hypotheses or new data, authors may refer to unrepli- cated findings, which may simply be false positives. Scientific advance in part requires identifying those hypotheses that are not supported by empirical data and on the publication of definitive negative studies. There is usually a bias against the publication of negative findings in research. The field of autism research is one with a large, potentially vulnerable lay audience that actively follows research. Therefore, the Editors consider it important to publish both replications and non- replications to foster the ultimate emergence of a balanced picture. At times, the pressure to report positive findings in the face of nonsignificant results from primary analyses may lead authors to perform ad hoc analyses to support an auxiliary hypothesis. Thus, a hypothesis that is refuted in the main finds support in a subgroup of data. Determin- ing scientific merit in such cases can be difficult for reviewers and Editors because the findings may help describe true heterogeneity, or they may be spurious. Subgroup analyses planned a priori, statistical limitations created by subdividing data and plausibility are all factors that help assess validity in such cases. Statistical power is an important concern, and some studies lack statistical power at the outset. The issue is particularly problematic with respect to genetic studies, in which associations observed in modest-sized samples are given undue weight; these issues will be discussed in more detail in a future editorial. For this reason, we ask authors to provide a statement regarding power in all submitted manuscripts. Studies that are preliminary, or perhaps out of necessity under-powered, should be submitted as short reports and these limitations noted. Another phenomenon in autism research over the decades has been a tendency for some influential hypotheses to lack explanatory power. Autism is a complex disorder involving both deficits in multiple domains of functioning and a very wide range of severity. Authors should to state clearly which aspects of ASDs their hypotheses and data address and which aspects of the disorder and its full range of expression are not accounted for. The phenomenon of ‘‘selective referencing’’ is a troubling issue for scientific publishing in general. Examples include instances in which manuscripts con- tain only citations that support the authors’ views or referencing may appear to over-represent subset of groups, perhaps even driven by geography. Neither scenario provides a fair and balanced basis upon which a proper judgement can be made about the merits of a given study. Referencing should be broadly inclusive. Both supportive and contradictory citations and text should be provided to the reader to place the study at hand in a balanced context. What sort of papers would Autism Research like to publish? In addition to definitive advances in specific areas, manuscripts that address the complexity and heterogeneity in autism directly, or contribute to its understanding seem likely to move the field forward. Another variable that is often neglected in studies is variation in severity of phenotypic expression. Reliably measuring symptom severity, and hence identifying neurobiological and psychological differences across the spectrum, is likely to contribute to identification of INSAR Autism Research 1: 71–72, 2008 71 Published online in Wiley InterScience (www. interscience.wiley.com) DOI: 10.1002/aur.18 & 2008 International Society for Autism Research, Wiley Periodicals, Inc.