According to literature some groups are often underrepresented in survey research. For instance people with a lower economic status, people in urbanized areas and lonely people. Internet panel research forms an extra threshold. People without Internet access and people who dont want to join for a longer period are less inclined to participate. In the Netherlands in 2007 a probability based Internet panel of 5000 households was recruited and all recommendations from literature to get an optimal response have been implemented; Longitudinal Internet Study for the Social Sciences (the LISS panel). We will discuss how older people, mostly without Internet access, were persuaded to participate in the panel. Furthermore we will consider the quality of the data of the older panel members (wave nonresponse, item nonresponse) over a year and compare the quality of the answers on some questions from the face-to-face survey SHARE (Survey of Health, Aging and Retirement), which were also asked in the LISS panel.