Max Weylandt.

A Look at Election Polls in Namibia

7 November, 2024

a scatterplot showing polled percentages on the x-axis and election results on the y-axis. In the bottom left most dots are clustered about the 45-degree line, indicating a close fit. Four points are at a significant distance from the line, three for SWAPO and one for PDM.

Over at my namibian elections website, I have a new blog post looking at how well Afrobarometer has traditionally done at predicting election results.

It’s a bit of a tricky issue to study for a few reasons:

there are only three elections with a survey within a reasonable timeframe
the questions change over time, and
the questions are unclear: they ask simultaneously about presidential candidates and party support

Looking across these surveys, a few trends appear:

for smaller parties, Afrobarometer has been remarkably accurate
there has always been a significant bloc of survey respondents who claim to not know their preference, refuse to answer, or say they wont’ vote
In 2009 and 2014, SWAPO overperformed the survey by almost exactly the size of this bloc. It might be that these are uncommitted voters who defaulted to the incumbent once in the polling booth.
In 2019, SWAPO overperformed the survey again, but by less than before. Opposition overperformed significantly for the first time, in particular PDM (12%) and LPM (3%)
If these pattern holds, we’d expect several parties to overperform their Afrobarometer numbers in the election in three weeks’ time. This includes SWAPO, but also probably several opposition parties.

← Back ↑ Top