Afrobarometer, a nonprofit polling project, reported that surveys conducted since 2000 have revealed a slow but steady deterioration in support for democracy, with particularly steep declines in Nigeria, Tanzania, Botswana and Uganda.

Sixty-one percent of Africans surveyed said they regard democracy as preferable to other political systems. But the pollsters found that flaws in many democracies across the continent have undermined support for the concept.

"Democracy's not been working politically," said Robert Mattes, one of the pollsters, speaking from Cape Town. Support for democracy is falling fastest in "places where you've had one-party rule for a long time or places where elections haven't been working." [...]

Powerfully negative trends were also found in Zimbabwe, which is six years into a political and economic crisis that has turned one of Africa's foremost success stories into one of its biggest failures.

All five of the countries mentioned have winner-take-all, single-member district systems. Other countries like Lesotho, also mentioned, have proportional or semi-proportional systems. Winner-take-all elections are especially prone to choicelessness, therefore low turnout and disengagement. Under those systems, there can only be two viable parties in any given district. This is problematic in multi-party systems, especially where partisan identity is closely bound to cultural identity. The opposition is likely to split its votes among many candidates, electing members with less than majority support, effectively shutting the majority of voters out of representation. Where the situation persists, it can result in unresponsive, one-party government. In turn, this can drive down faith in democracy.

Andrew Reynolds' 1993 work on Malawi touches on the point:

Applying the Criteria to Malawi: The Case for PR

The fundamental reason why drafters of a new Malawian constitution should prefer PR over plurality to elect their new parliament is that PR will largely avoid the anomalies that plurality often produces (especially when used under multi-party conditions). The lack of a clear link between national votes polled and parliamentary seats can be devastating to stability in newly democratizing, politically fragile states.

Distorted results could push Malawi to a civil war of the likes seen in Angola and Mozambique. The study of election results under plurality shows that these fears are by no means groundless, as single party governments are often elected under plurality systems with less than a majority of the popular vote. In new multi-party democracies where there is distrust among groupings, losing groups in these situations may easily resort to extra-parliamentary, destabilizing tactics if they feel cheated by the electoral system.

A system of PR will be inclusive enough to bring minority interests into the legislative arena of government even if they are still excluded from the decision-making structures. Whether they are Malawians from the North, a certain type of religious group or one of the smaller linguistic groups, they will have the opportunity under PR to organize and play a role in national party politics.

It would be interesting to investigate the opposite: whether approval of democracy can be correlated with proportionality of legislative representation.