The internet has changed our world. To many, but not yet all, logging on to the worldwide web is a huge and irreplaceable part of life.

My research team has just completed a new study of who in Britain has internet access at home. Worryingly, we find that a profound digital divide remains.

One third of Britons, we discover, can now log on from home. On average, they spend three hours a week doing so. In the digital world, money talks. For those homes where family income is above £30,000 a year, two thirds of people use the internet. In homes bringing in less than £10,000 a year, less than a tenth of people have web access. Yet it is not just a matter of cash. Education and occupation make an enormous difference too. Someone with a university degree is five times more likely to click on to the internet than a Briton who left school with no qualifications.

There is also a gender digital divide. British men use the worldwide web much more than women. And young people, as one would guess, are far more likely to access the internet than the old. Only 5% of pensioners in Britain have ever logged on.

Much less predictably, there is a strong North-South division. In Greater London, for instance, 42% of citizens log on outside their workplace. In the North, it is only 18%. Nobody currently understands why.

Beyond the sheer arithmetical patterns, there are deeper issues about what the internet is doing, and will do, to our society.

One original stereotype went like this: the typical internet user tends to be a loner, cut off from society, not keen to know people or be sociable, uninterested in and distrustful of community, and not a great one for books or religion or culture.

Our research proves that this stereotypical view is utterly wrong.

First, internet users are better citizens, not worse. A proper analysis shows that internet users are more likely to be members of local community groups and voluntary organisations. This link holds statistically even after holding constant lots of other characteristics about people.

Second, internet users go to church more than the average person. This is a remarkable fact, because those who log on are much younger on average than those who make up the typical church congregation.

Third, internet users have more and closer friends than do people who never surf the net.

Fourth, internet users watch less television than the average Briton: 2.4 hours a day compared to 3.5 hours. They also read books more often. Once again, these patterns are not a mirage caused by the different age and educational backgrounds of worldwide web surfers. These patterns, my team demonstrates, hold even after adjusting for people's backgrounds.

Finally, in our survey work we find that those who log on to the internet say that they are trusting of others. These people are not antisocial loners.

What should an economist or other social scientist make of these findings about the role of the web and the nature of internet users?

One possibility is that a slice of society may, even decades from now, have no access to the internet. If the current digital divide continues, British society will face serious problems. As the years pass, information for our citizens and all kinds of educational material will be directed more and more through the web and home computers. An underclass of non-learners could then appear - rather as happened in the 19th century among those who could not read or write. By some people's standards, home computers are expensive. With wage inequality rising over the long swing of recent history, it is not possible to be sure in the year 2001 that every house will be wired to the internet in our lifetime. It took telephones nearly a century to become ubiquitous in homes, for instance. Possession of a phone never held the educational advantage that internet access offers. Inequality in who can access the web is deeply undesirable.

Something that worries our current generation of politicians is that unemployed people might be hampered by lack of internet access. We test that. Our statistical inquiries do not, in fact, support it. It is true that those without a home computer are more likely to be out of work. But that, we find, is explained by the lower educational and previous income levels of jobless people in Britain. Hence, there is no evidence that simply giving people internet access would get them into jobs more quickly.

It might be argued that blue-collar workers and their families are less in need of internet access than others. That would be shortsighted. We know that manufacturing industry is declining, as a proportion of employment in the whole economy, and that the current one fifth of British jobs that are in manufacturing industry will have shrunk to one job in 10 in Britain by the end of the next decade. By our grandchildren's time, all but a handful of us are going to be white-collar workers.

In the future, our economy will run almost exclusively on ideas, and the internet will be the chief channel down which they will flow. That is why Britain's digital divide must be closed.

Andrew Oswald is professor of economics at the University of Warwick. Visit the author's website at www.oswald.co.uk