Surveying general population - within probability sampling approach - is one of the major challenges of applying web survey questionnaires. Various types of web access panels might be a good enough alternative for many surveys, however, when we talk about official surveys and about serious academic surveys this is still far from being the case. Similarly, a probability based panels may be even more promising alternative, however, it similarly requires a considerable infrastructure investments and so far very few countries have it. On the other hand, there are numerous attempts for involving web survey mode within probability sample selection context in one-shot research studies. In this presentation we overview the experience with this type of studies done in Slovenia. In addition, we also present experimental studies that were not done (although anticipated), because it was found that the added value of such experiments would be too little. All presented studies use the approach where persons are selected from the register of population and then invited by ordinary mail - sometimes also with incentives - to do a web surveys, while in further stages the telephone or/and face-to-face interviews were attempted. The main question here is, whether the response rate in web stage is large enough, so that the corresponding benefits (predominantly the savings of resources) can compensate the potential mode-effect problems and particularly the increased administration complexity of such mixed mode designs.