It’s always election season in
the United States: the arguments over the last election are still going
well after campaigning for the next one has started. So while most
attention is focused on this year’s mid-term Congressional elections and their significance for the 2008 presidential election,
Rolling Stone magazine has taken the opportunity to revisit the 2004
election. In this month’s issue, Robert Kennedy Jr examines the 2004
Bush/Kerry contest in Ohio under the heading “Was the 2004 Election Stolen?“

Kennedy’s
answer is clearly “yes”, and he recounts an impressive array of
irregularities that disadvantaged the Democrats in Ohio. Particularly
egregious were the antics of Ohio secretary of state Kenneth Blackwell,
who, in addition to being responsible for implementing the state’s
election laws, was also co-chair of the Bush re-election committee in
Ohio. Guess which role he treated as more important.

But that’s
how American elections work: a patchwork of local rules is administered by mostly partisan officials, offering ample opportunity
for outcomes to be manipulated. Is that unfair? Yes, of course it is. But
both sides of politics have benefited from it in turn; the 1960 victory
of Kennedy’s own uncle, John F Kennedy, depended heavily on the shady
tactics of Democrat bosses.

It’s another question whether the
unfairness amounts to actual fraud, and the evidence that the Ohio
election was really stolen is much thinner. (Farhad Manjoo at Saloncasts a sceptical eye over it.)
Kennedy relies a good deal on statistical analysis, and particularly on
discrepancies between exit polls and the official results. In his most
striking statistic, he says that results in one precinct were so
unusual that “the statistical odds against such a variance are just shy
of one in three billion.”

In reality exit polls are much less
accurate than Kennedy thinks. And his statistics only work on the
assumption that the only relevant error is sampling error. But nobody
actually thinks that: when polls are systematically wrong, there is
probably some systematic cause. It requires a further leap to conclude
that the cause is electoral fraud, and the statistics certainly don’t
establish that there were enough illegalities to overturn Bush’s Ohio
margin of some 118,600 votes.

It will be a good thing if
Kennedy’s article draws attention to the reforms needed to bring US
election procedures into the 21st (or even the 20th) century. But it
will probably fail to convince most people that Bush’s second election,
in contrast to his first, was fundamentally illegitimate.