Shula Marks's review of our Launching Democracy in South Africa: The First Open Election, April 1994 (THES, March 22) contains an all-out attack on our scholarly objectivity mainly, it would seem, because we do not bend the facts to the political line she herself prefers.

Thus she writes that, while the book has "an air of scientific authority", it only "appears meticulous". She finds our analysis "distorted by the almost visceral hostility to the ANC which seeps insidiously through the text" - but adds that "It would, unfortunately, require a detailed textual exegesis to show this". Surely, if such visceral hostility was truly everywhere in the text, it would not be so "unfortunately" difficult to find? The truth is that we invited the various contributors to the book to participate without enquiring as to their political views, taking account only of their scholarly merit.

Professor Marks's particular ire is reserved for our findings in KwaZulu-Natal. Anyone who has worked in this political quagmire knows how hard, and necessary, it is to keep one's balance . Yet Professor Marks is angry precisely because our chapter contains "a careful balancing act". What actually happened was that we carried out two large opinion surveys there which both showed large ANC leads: we published these in mid-campaign, knowing full well that we would draw fire for following what we believed to be the truth. Sure enough, both surveys were furiously condemned by Inkatha and were rapturously received by a delighted ANC, which continued to quote them for months afterwards.

The election itself saw the IFP beat the ANC handsomely in the province and although there was much cheating it became clear that even if all the disputed ballot boxes were awarded 100 per cent against the IFP, their margin of victory would still be very comfortable. On investigation we found that we could sustain neither of the major ANC or IFP complaints of miscreancy by the other, and thus we had to accept that the result, though clearly very imperfect, was probably crudely right - in which case we had to explain why our surveys had got it wrong.

ÖйúAƬ

We then carefully re-analysed our data, carrying out systematic regression analysis, and the result was the hypothesis of the large but hidden IFP vote by those too intimidated to reveal their intentions to our interviewers. This still makes the most sense. Professor Marks mistakenly believes that we are trying to put the main onus for the intimidation on "ANC comrades", thus exonerating the IFP. Not so: all the examples of IFP attacks on ANC supporters during the campaign cited by Professor Marks are taken from the data we ourselves present. All our surveys had shown that IFP supporters were by far the most bullied and timorous but we were careful to point out that we did not know whether this was as a result of bullying by the ANC, the IFP or both. There is still no scientific basis for finally settling that question - and Professor Marks surprisingly fails to understand that when we found large numbers of older, illiterate and poor rural women terrified of the young "comrades", even this does not necessarily mean that the comrades were doing all the bullying. All over the world old ladies of this kind are easily frightened by teenage tearaways, even when they do not wield AK-47s: that is, the IFP's electorate was sociologically simply more prone to feelings of vulnerability and intimidation than the ANC's younger, more male, more urban and better-off electorate.

Professor Marks is annoyed that we should speak of the "loyalty and professionalism of the armed forces" during the election. It is true, of course, that the security forces were involved in many brutal operations in the 1980s and that they frequently showed a clear partisanship towards Inkatha. After 1990, however, it was (as we point out) the KwaZulu police who were the main object of ANC dislike, not the security forces, and by the time of the election the KwaZulu-Natal ANC was strongly demanding a heavier army deployment in the province. In the event it was largely the armed forces who saved the election from collapsing into chaos. The army played a key logistical and protective role, the air force flew a record number of missions to get ballot boxes to remote areas, and both foreign observers and President Mandela paid warm tribute to them. Thus the forces which were necessary to the preservation of apartheid were ultimately crucial in ensuring that the first democratic election (more or less) worked - and thus helped give the ANC their deserved victory. Such ironies of history may not be to Professor Marks's taste, but history is history.

ÖйúAƬ

The problem here is simply one of partisanship. Professor Marks has every right to her own partisan views but that is, of course, what they are and it is naive and unbecoming of her to attempt to impose them as some sort of truth. Scholarly detachment is particularly difficult in South Africa in the face of the heated and various passions which have long characterised our history. We are now in a situation wearily familiar to honest liberals. We published our surveys of KwaZulu-Natal opinion and were denounced by the IFP. In the light of later evidence we came to the painful conclusion that our findings had been faulty, were honest enough to say so, to do more work, and to try to explain what had occurred. So now we are denounced by the other side. We who actually live here would really hope that those who, like Professor Marks, choose to live in easier climes, would weigh in on the side of scholarly detachment, and not of partisan passion.

R. W. JOHNSON LAWRENCE SCHLEMMER Helen Suzman Foundation Johannesburg, South Africa