This raises so many questions.<p>I'd be interested in how many people in the study from each group die off during the study... is there a larger number of people in the unmarried group dying off, thereby increasing the chance that the married group would age to the point where dementia is more likely?<p>The question about married people being more likely to be able to afford health care (via insurance) would seem to be a uniquely US issue... I'd be interested to see the health impact of marriage on different countries, different cultural groups etc.<p>Culture would seem to impact greatly on what marriage entails... many groups appear to have a strong link between marriage and cultural activities, others it allows the couple to be a self contained social group, removing the link to external socialisation.<p>The 50% figure seems to be pretty strong, but at least from the article it is hard to draw too many conclusions without knowing how wide the sample set of participants are.