Background: Internet-based self-help is an effective preventive intervention for highly prevalent disorders, such as depression and anxiety. It is not clear, however, whether it is necessary to offer these interventions with professional support or if they work without any guidance. In case support is necessary, it is not clear which level of support is needed. This study examines whether an internet-based self-help intervention with a coach is more effective than the same intervention without a coach in terms of clinical outcomes, drop-out and economic costs. Moreover, we will investigate which level of support by a coach is more effective compared to other levels of support. Methods: In this randomized controlled trial, a total of 500 subjects (18 year and older) from the general population with mild to moderate depression and/or anxiety will be assigned to one of five conditions: (1) web-based problem solving through the internet (self-examination therapy) without a coach; (2) the same as 1, but with the possibility to ask help from a coach on the initiative of the respondent (on demand, by email); (3) the same as 1, but with weekly scheduled contacts initiated by a coach (once per week, by email); (4) weekly scheduled contacts initiated by a coach, but no web-based intervention; (5) information only (through the internet). The interventions will consist of five weekly lessons. Primary outcome measures are symptoms of depression and anxiety. Secondary outcome measures are drop-out from the intervention, quality of life, and economic costs. Other secondary outcome measures that may predict outcome are also studied, e.g. client satisfaction and problem-solving skills. Measures are taken at baseline (pre-test), directly after the intervention (post-test, five weeks after baseline), 3 months later, and 12 months later. Analysis will be conducted on the intention-to-treat sample. Discussion: This study aims to provide more insight into the clinical effectiveness, differences in drop-out rate and costs between interventions with and without support, and in particular different levels of support. This is important to know in relation to the dissemination of internet-based self-help interventions. © 2009 Donker et al; licensee BioMed Central Ltd.