Piet CroucampFew things stir up commentators and analysts at the IEC’s main results centre in Tshwane quite like polls, predictions and forecasts.It seems a generally good point to make that the polls must be wrong, had always been wrong and might even reflect a degree of analytical opportunism to cover up for the likeliness of being wrong.Steven Friedman of the Centre for Policy Studies at the University of Johannesburg is not only sceptical about the forecasts and conclusions of the Ipsos/ENCA poll; he also suggests that Ipsos director Mari Harris is deliberately obfuscating her optimistic and possibly mistaken analysis of the Tshwane vote.He accuses Ipsos of slipping in the so-called “undecided voters” with the tally of the measured support for the ANC which then conveniently forecasts – or predict if you feel the need for some scholarly venom – that the ANC may well lead the polls comfortably in Pretoria.The collective, but morbid murmur from the DA’s well-oiled electoral machine almost measured on the Richter Scale and the shock and awe of another five years of ANC rule in the capital city prematurely rippled through Waterkloof Ridge. While polls indicated a week ago that the ANC was under threat of losing their political monopoly in three Gauteng metropoles, the best possible outcome or scenario for Mmusi Maimane now was coalitions with either the EFF or even the ANC. Not even the smaller, but ideologically more palatable parties such as the FF+ or the UDM would push the DA to the top of the electronic results board at the IEC centre in Tshwane.ANN7 also published a poll, but their reputation as a lower-end of the food chain broadcaster discourages most analysts from seriously considering whatever forecasts or predictions they make. The fact that their results predict 22% of whites to vote for the ANC may not ring a cynical bell in a Sandton Villa or at Nkandla, but at least Ipsos will – in their effort to “clean-up” their data – notice the inconvenient variable in the equation.The fact is: it is very difficult to make predictions about electoral outcomes, even in a country such as South Africa where demographics is a relatively accurate predictor of voting behaviour.White people generally don’t vote for the ANC and 95% of those who vote for the ANC will be black. The coloured voters of the Western Cape carry the burden where the white numbers, which will address the influx of “foreigners” from the Eastern Cape, fall short. And 98% of those who vote for the EFF will be black South Africans.Psychological variables The DA might have more voters who would have been classified as non-white under Verwoerd, but the fact is their campaign will stand or fall by their ability to attract “the black vote”. But we have more complicating variables than demographics.News24 analyst Dawie Scholtz refers to psychographics as an equally important psychological variable which determines why people vote for a particular party. Psychographics may be a helpful frame of reference in answering the question about how a socio-economic entity such as the so-called black middle class will vote or what may influence them to change their political loyalties. Apart from the obvious question about the operational reach and range of a middle class, you will have to divide the middle class variable into urban vis-à-vis rural middle class. Further complicating the concept is the fact that a middle class benefitting from the patronage of an overpaid state sector may well behave politically very different from their counterparts in the private sector who as a general rule earn approximately 36% less for their significantly more productive efforts. Even more esoteric and complicated is the extrapolations about what would change the voting behaviour of a heterogeneous identity such as a black middle class. We suspect the black middle class might only have formal employment in common with the white middle class.Human behaviour not a scienceA clear ideological fault line between black and white may be why it seems as if the black middle class would rather vote for the EFF than for the DA. We also have no evidence that the poorest of the poor and the unemployed, who possibly qualifies as the most neglected constituency of the governing party’s economic policies, are prepared to sever the umbilical cord with liberationism. To predict electoral outcomes is a treacherous endeavour because of the large number of variables involved. Also, human behaviour will probably never be a science.So, to so-called pundits, data, statistics and facts may well be the road less travelled when it comes to analysis, extrapolation and forecasting. It does not even necessarily lead to consensus, and all too often the truth is nothing but the result of the imagined realities of even the most educated and experienced analysts and commentators.Now factor in egos before you make up your own mind about what will be reality and fact in a few days.Disclaimer:News24 encourages freedom of speech and the expression of diverse views. The views of columnists published on News24 are therefore their own and do not necessarily represent the views of News24.