1. SWAPO will lose ground in the Zambezi

Two factors will work against SWAPO here. For one, the region has always been relatively politically volatile, always taking quickly to newly formed parties (CoD did well there in its first election, RDP too).  Secondly, there’s the fact that the region was renamed, which apparently left many residents upset. Add to that the bizarre story of locals being abducted by the Botswana Defense Force, and I wouldn’t be surprised if people were unhappy.

2. RDP and DTA will battle it out for the South

In 1994, the DTA got the majority of its votes from Khomas and the surrounding regions, especially to the South. This is also where the RDP got a large chunk of its vote. The DTA will want to regain some of that southern electorate, as well as Windhoek voters. The question for the RDP is if they can offset losses they make there (and they will, given that the DTA has mobilized effectively and Hamutenya’s absence from the campaign trail) by making gains in the north. Otherwise they might just lose their official opposition status to the Venaani’s party.

3. Swapo will lose in half of the country, but it won’t matter

While SWAPO’s support is overwhelming overall, the distribution of people means they don’t actually win everywhere. Look at this map of last time around (this is the National Assembly results):

In terms of area, SWAPO only gets a majority in half the country. But of course they rack up huge totals in the tiny, but highly populated constituencies in the north as well as in Windhoek and at the coast. That adds up, and overall they win by 75 percent.



This morning, the IPPR held a press conference to release some early data on the upcoming elections from the latest Afrobarometer round. This is very exciting, because we haven’t really seen scientific, political polls in Namibia. In this post I’m going to briefly look at whether we can get something useful from these polls, or whether we should discard them as unrealistic.

A Primer on Polls
Many people mistrust polls. This makes sense because they intuitively seem untrustworthy: how can I call up a few hundred people and make claims about the whole country based on that? But it really works, even if it seems magical. George Gallup, the godfather of political polls, had a good analogy to explain why:

Sampling public opinion is like sampling soup: One spoonful can reflect the taste of the whole pot if the soup is well-stirred.

If you pick a good sample that represents the population well, it’s fine to draw broad, national conclusions from it. It seems unlikely but it works remarkably well.

So we know that Polls Work. But an individual poll can sometimes be off. This is why sophisticated forecasts — like Nate Silver’s famous one in the U.S. — often average several surveys for better results.

We don’t have that luxury here in Namibia. We will have to make do with just the Afrobarometer. So is it any good?

Continue reading

Update: I added the DTA manifesto too. 

So I decided to generate a quick word cloud from the two biggest parties’ manifestos. Word clouds are fun because they can tell you at a glance what a document focuses on; which words crop up more often than others. Here we can use them as an informal way of looking at the issues these parties talk about in what is really their major declaration of their intentions, policy-wise.

Note though, they can be misleading and are by no means a rigorous appraisal of what these parties’ priorities are.

Here’s why: When making this word cloud, I had to pick to what extent I wanted exact matches for words or put similar ones into the same category: in other words, do “agriculture” and “agricultural” count as two different categories, or as one? What about “farming”? If a word has many different versions, it may not even show up in the count, while others seem large. So, for example, there are many ways of talking about business (industry, entrepreneurship, commerce, etc) while there is only one way of talking about HIV. Even if the focus is on the economy, not health, it may seem like HIV is the main issue because the attention on the economy gets ‘spread’ across different words. Just an example, but still, please don’t take these too seriously.

But still, they’re a lot of fun to look at. Click on an image to see an enlarged version.

First, here’s the one for SWAPO:




And here’s the cloud for the RDP:


Finally, the DTA:


Alternative, more colourful layouts: SWAPORDP, DTA.


More comparisons to follow.