The increasing relevance of the mobile Internet is undoubted. In 2008, the share of users in Germany mounts to 16% among the German population (Source: TNS Infratest Media&Technology, September 2008). The usage intention also reaches a two digit percentage (12%) and announces further growth.

In course of this development, market research agencies are increasingly faced to the need of their clients to deliver information about the universe of the mobile Internet. Classical research methods like telephone surveys or online surveys on the static Internet can of course be used as a means of collecting relevant data. But up to date it is still very difficult to reach the relevant target groups via these methods due to their low penetration rate. This difficulty increases especially when subgroups like visitors of certain mobile websites have to be addressed.

Since more than one year, academic and commercial research has started to develop mobile research approaches where the recruitment and interview takes place directly on the mobile Internet. The main advantage of these approaches is the direct and efficient access to relatively small target groups.

But how representative are the results of these approaches that take place directly on the mobile Internet? Can the data be considered as a valid and reliable source of information? How can market research agencies control for potential biases to deliver useful value add to their clients?

To answer these questions the authors of this study compare the empirical results of a mobile Internet survey and a representative telephone survey. The paper starts with a description of the methodological design of both surveys. In the main section a detailed comparison of basic results of the two surveys is undertaken. Therefore the authors mainly use demographic profiles and information about the usage behaviour on the mobile Internet. The paper concludes with a statement of how results from mobile Internet surveys can be extrapolated to the corresponding population.

The purpose of this paper is to actively contribute to the further development of mobile research approaches.