Who are undecided voters and why have others changed their voting intentions?

Voters queue to cast their ballots at the NSSF Grounds in Nairobi during the 2017 polls. [Stafford Ondego, Standard]

By all accounts, the forthcoming presidential election is going to be a very close contest.

Two factors underlie this assessment: first, that the margin between William Ruto and Raila Odinga is so close (only about two per cent in TIFA’s most recent poll – within its margin-of-error of +/- 2.1 per cent, and second, that those respondents declining to identify their voting intentions (either because they declared they were ‘undecided’ or who chose to remain silent altogether about this) amount to a significant proportion of potential voters: 7 per cent.

In other words, it can be assumed that it is they who will decide who the next president is. Such figures raise two questions: which of the two (main) candidates will obtain the most votes, and will be the one who also obtain the required 50 per cent + 1 for an outright win?

Also shown in TIFA’s poll, both candidates increased their overall rating by about five per cent (Ruto from 39 to 44 per cent, and Odinga from 42 to 47 per cent).

Two factors account for this: a decline in those claiming to be undecided (from 10 to 4 per cent) as well as those who gave no response (from 4 to 2 per cent). 

Combined with the drop by half in expressed intention to vote for George Wajackoyah (also from 4 to 2 per cent), this explains where the combined gain of 10 per cent for the two main candidates came from.

But as suggested, this remaining seven per cent (of those who declined to declare their voting intention) will be critical. Is there anything their identity and attitudes also captured in this survey that might indicate which way they will go?

This article, utilising so far unreleased data from TIFA’s most recent (late July) survey (and purchased by The Standard), proposes to explore (if not definitely answer) the two most relevant questions related to the above: just who are these respondents who failed to reveal their voting intentions, and does their identification give any clue how they will vote, and how many have changed such intentions since the beginning of this year, and for those who have, why have they done so?

Getting inside the numbers

Let us look at these ‘undecided’ figures in more detail. The first point to make here is that the proportion of respondents declining to name a preferred presidential candidate has continued to decrease since the beginning of the year.

For example, according to TIFA, it dropped by about half from 30 per cent in January to just 14 per cent in late June.

This stands in contrast to the relevant figures for the same period prior to the last three elections: in 2007, one per cent; in 2013, five per cent; and in 2017, nine per cent. It should be noted that the surveys, which yielded these figures from the three previous elections were all conducted face-to-face at respondents’ households, in contrast to the three at issue here.

It may be the case that in the former setting where interviewers and respondents are able to establish a more personal” relationship, it would be more “awkward” for a respondent to avoid answering this question.

As noted, the current TIFA figure is five per cent. Yet with only two weeks remaining when this survey was conducted, it cannot be assumed that all such respondents have not, in fact, made up their minds, but may be too shy to reveal their voting intentions, for one reason or another.

Ballot boxes. [Elvis Ogina, Standard]

One way to test this assumption is to pose this question: what proportion of these respondents who claim to be undecided about their presidential vote have, in fact, made up their minds about the other choices they will confront on their ballot papers?

To answer it, TIFA asked all ‘undecideds’ (on their presidential choice) if they have decided who they will vote for governor, senator and MP. Altogether, more than half (58 per cent) reported that they had yet to make up their minds on any of these choices.

By contrast, just over 28 per cent indicated that they had indeed made up their minds about the three. It thus appears that these latter respondents very likely decided not to reveal their presidential preference though they actually have one.

By contrast, it may be assumed that former respondents are being honest, since if they have yet to decide their vote-choice for any of these other positions, their attention-level to the campaigns – and perhaps to the election itself – is obviously low (even if they are classified as ‘likely voters’, given that they said that they will either “definitely” or “probably” vote).

Reason(s) for being undecided 

Another clue about the ‘real’ position of these self-professed undecided voters is provided by the distribution of responses to the next question they were all asked: “What is the main thing that will help you decide which presidential candidate you will vote for?”

Whereas a substantial proportion (28 per cent) identified “more information about the policies/the manifestos of the candidates” as the main factor that would allow them to decide, more than twice as many (61 per cent) said that they are “not sure” what would help them make this decision.

At this late stage of the electoral cycle, how anyone could not be certain what they need to know (or whose opinion they need to hear) in order to make this important decision seems improbable, suggesting that this vague response is nothing more than a ‘ruse’, when they (or at least most of them) have, in fact, made up their minds.

Political party self-identification

Another ‘window’ into the minds (or at least the political profiles) of the ‘undecideds’ is gained by getting the answer to another question: what proportion of them do/do not self-identify with any political party?

Azimio la Umoja presidential candidate Raila Odinga. [File, Standard]

The assumption here is that far more of those who do may be expected to have made up their mind about their presidential voting intentions – though as noted above, some who have done so may be shy about revealing this.

The data here indicates that this is indeed the case, with over 90 per cent of those who self-identify with any political party also naming a presidential candidate, as compared with just slightly over half of those who do not so identify named any.

Thus, once again, the latter exhibit a much lower level of electoral interest.

At the same time, these ‘undecideds’ regarding the presidential contest but who do, nevertheless, self-identify with any party are nearly evenly split: three per cent doing so with ODM and two per cent with UDA, and one per cent each with regard to the two opposing coalitions.

Indeed, only with the benefit of credible official results will it be known if at least some of those declining to reveal their voting intentions have actually concealed them – similar to the significant proportions of respondents in the surveys that were “wrong” with regard to Donald Trump’s victory in the 2016 US election and in the UK’s “Brexit” vote the same year.

Assuming that is the case, will Ruto or Odinga benefit most when the real votes are counted? At the same time, many have remarked that, in comparison with most past pre-election polls in Kenya, the proportion of those self-declaring as ‘undecided’ has much higher this time.

Confidence in election integrity

Another factor evidently affecting the naming of a preferred presidential candidate can be put in the form of another question: does confidence in election integrity influence the willingness or propensity to name a preferred presidential candidate?

According to the data, most respondents who did not name a candidate believe that false election results (for whatever positions) are either “very likely” or “somewhat likely” after the August 9 votes are counted: 56 per cent against 52 per cent.

Conversely, rather more of those who named a candidate are “quite certain” that this will not happen as compared with those without a (stated) preferred presidential candidate: 40 per cent against 32 per cent. (More of the latter are also “not sure” if any election ‘wins’ will be stolen or not: 12 per cent against eight per cent.)

Two demographic Variables -- Education and English language competency: The impact of education on the inclination or ability to make a decision about one’s presidential vote is significant.

DP William Ruto. [Samson Wire, Standard]

For example, among those who mention an intention to vote for any candidate, only 26 per cent have not gone beyond primary school, whereas more than double this figure (56 per cent) who do not mention any candidate are within this lower education bracket.

Regarding language ability, only 33 per cent of those who did not name a candidate say they can speak English, as compared with 57 per cent of those who did (clearly largely a correlate of one’s education level).

But these two variables combined suggest that at least part of the failure to name a candidate is likely to be insufficient access to political information as opposed to a propensity to conceal one’s intentions.

Who has changed their presidential preference, and why to probe deeper into the process (surely both individual and communal, these two factors vary across the country both between and within various communities).

TIFA also sought to discover just how many people have changed their voting intentions (limited to the period since the beginning of the year).

Such changes are of three types: from being undecided to supporting a particular candidate; from supporting a particular candidate to being undecided; and to changing support from one candidate to another.

Even without data, some of the most salient reasons can be suggested: because the person dropped out of the race; because one’s most admired or local political/community leader moved into or out of a relationship with one or the other presidential campaign team or alliance; because of campaign or other messaging; or because of the candidates’ choices of running-mates.

In this endeavour, the first requirement was to determine how many people say that they have done this, and who they are, especially in partisan-political terms.

Altogether, a substantial minority –13 per cent – stated that they have done this. Significantly, a clear majority – 57 per cent – reporting having changed from one candidate to another, with slightly more of those now supporting Odinga than Ruto (five per cent) among them.

However, the DP has the ‘upper hand’ in terms of winning over those who had initially been undecided (by a 12 per cent margin).

Proportion of political parties supporters 

Moreover, among those who had previously intended to vote for one presidential candidate but are now preparing to vote for another one or are undecided (as noted above, just five per cent of the total sample), Raila and Ruto have ‘lost’ a nearly equal number of potential voters (36 per cent and 33 per cent, respectively) of these relatively few respondents (just 12 per cent of all those whose presidential voting intentions have changed at all).

Roots Party presidential candidate Prof George Wajackoyah. [Boniface Okendo, Standard]

Nearly all of the remainder previously supported someone who never entered or who dropped out of the race (25 per cent: especially Musalia Mudavadi or Kalonzo Musyoka).

In partisan terms, while an equal proportion of supporters of the political parties, which each of these candidates leads (ODM and UDA) report having changed their presidential voting intentions since the start of the year (11 per cent), far more who continue to self-identify with Jubilee have done so (26 per cent), and almost as many among those who support any of the numerous other parties (22 per cent).

The take-away being that those attached to a party lacking a presidential candidate have been rather more inclined to ‘move’ in one direction or another.

But what, at least as far as the respondents say, accounts for such changes of voting-intentions?

For most (42 per cent), it has been campaign promises and manifesto content, though “the influence of others” and “the choice of running-mate” also register significant numbers (13 per cent and 10 per cent, respectively).

Less impactful among this group as a whole was the failure of an admired leader to join (or withdraw from) the presidential race ( for instance Mudavadi, Kalonzo and a few others at eight per cent) as well as an evaluation of opinion polls (four per cent).

What this contest is really about

Even if this is Ruto’s first attempt at the presidency, the considerable duration of his public career even before his decade-long service as deputy president, means that neither of the two main candidates suffers from any ‘public profile’ deficit.

Yet one factor contributing to the higher than usual proportion who have chosen to remain ‘undecided’ in recent surveys may be a higher level of confusion as to what agendas they both represent as compared with the main contenders for the country’s office previously.

Such ‘confusion’ may stem from the fact that they have largely exchanged their political ‘clothing’, with the deputy president largely campaigning against his president – and thus the status quo, even if throughout their first term, they appeared to be in comfortable lock-step with each other.

The former prime minister conversely usually defends the president, even as he, too, tries to offer credible change for an electorate, the vast majority of which is currently suffering a plethora of economic (among other) woes.

Indeed, in a TIFA survey of June 2021, 50 per cent more respondents identified Ruto rather than Odinga as “the political leader in terms of criticising the Jubilee government and trying to hold it to account”, and in TIFA’s April 2022 survey, some three-quarters of respondents identified Odinga as “Uhuru’s preferred successor”.

Such a situation makes it largely impossible for Odinga to assume the anti-government posture he has assumed in the last five elections, notwithstanding his short-lived absorption into Moi’s KANU government and party in 2001.

Finally, in more systemic terms is the fact that without political parties that exist from one election to the next to which are attached contrasting policies relating to issues that matter to voters and that are fairly consistent over time – which in ‘advanced’ countries often have to do with policies related such issues as taxation and its application to inequality, the level of government penetration into private lives, the level and nature of involvement in foreign affairs, the relationship between national and regional (or in Kenya’s case, county) governments, the degree of independence of various state organs from Executive control or influence and the separation of powers more generally, import and export policies as related especially to the desired degree of self-sufficiency for the production of most consumer goods – including food – and so on.

Agano Party presidential candidate Waihiga Mwaure. [Jenipher Wachie, Standard]

The ‘choice’ of candidates – even when belonging to different parties – tends to fall back onto perceptions of their personalities and identities (including but not limited to their ethnic groups), as well as onto their feelings about the status quo and the conviction that one or the other candidate will do more to change things for the better as defined by that voters (or family, or community).

At the same time (and as is often the case in ‘established’ democracies), some voters will seek to ‘punish’ incumbents as a way of expressing their unhappiness with the status quo, rather than an expression of confidence in any particular set of would-be leaders.

Given the above, with a significant level of confusion as to just ‘what’ either William Ruto or Raila Odinga represents in terms of what Kenya would like after five years of having either of them at State House (of course, depending in part as to the distribution of party strength in Parliament and among the 47 counties), it seems of little surprise that the ‘undecideds’ – and those who admit to having changed their voting-intentions within just the last seven months – are as numerous as such research has revealed.

But with only a few days remining until ‘judgment day’, the ‘mystery’ of just how they will use their votes will soon be resolved, with this group of voters very likely to make the difference – assuming they actually get to their polling stations on August 9.

For as has been pointed, try as they might, they will not find ‘undecided’ as a vote-option in the list of presidential options, nor in those for any other seats.

-Dr Wolf is the lead pollster at TIFA – a Nairobi-based market research and opinion polls firm