A few weeks ago, Nation Media Group (NMG), a media house in Kenya sent a notice to its journalists asking them to treat, with skepticism, opinion polls trumpeting the rising or falling fortunes of Kenyan political leaders. "We take the view that, not knowing the sponsors of and how the polling was conducted, we risk our reputation if we play up the "findings" of these pollsters," read the notice.
Being an election year in Kenya, the notice follows a Bill seeking to regulate the manner of publication of electoral opinion polls tabled in parliament in October 2011 by Ikolomani Member of Parliament Dr. Boni Khalwale. The Bill among many other requirements compels pollsters to provide information like the sponsor of the opinion poll, methodology used and educational levels of the participants.
I sought the opinion of pollster Dr. Tom Wolf who has conducted opinion polls in Kenya in the last ten years and worked with research company Ipsos Synovate. Must research firms reveal who funds the research?
Dr. Tom Wolf
: The question as to whether survey firms "must reveal" the source of the funds such undertakings require is unclear. Presumably, since in Kenya, at least, there is no law at present requiring this (even if a clause to this effect is part of the draft law on the subject), the question seems to be asking whether such firms 'should' do this.
This, in turn, leads to the question as to why anyone cares who such funders are, and if some people do, who are they? There appears to be two general motivations behind it. One is hostility to such polls altogether, especially of those associated with individuals or policies that receive lower than desired ratings, and who are unhappy about that. Consequently, they may attempt to do whoever is possible to thwart them, or at least to discredit them. The other is a belief - often held even without any clear evidence - that (a) their publication can affect voting behavior, and (b) that this is somehow injurious (rather than helpful) to democracy.So do opinion polls affect a democracy?
: On the face of it, this position seems inherently contradictory, since if "democracy" means 'rule by the people', surely the people have a right to know what the state of public opinion is - including their own, collective voting intentions - during an election season (starting with the nominations/primaries process) so as to make the most 'strategic' (i.e., effective) use of their votes.
Likewise, it is obvious why candidates/political parties should benefit from such opinion-knowledge, since it helps to tell them what sort of candidates, on what sort of issues, they are more likely to do well, and among which parts of the country/sections of the population it makes most sense to invest limited campaign resources.
Presumably, however, at least the most well-funded campaign will undertake such surveys confidentially, and without any compulsion to release their findings to the public through the media.)How practical will it be to reveal who sponsors/funds these polls?
: There is the more practical challenge of complying with such a requirement, should it become law. Unlike the case in most advanced (larger, wealthier) countries where a single media house (TV network or newspaper - or sometimes a pair of them) often fund such polls, in a country like Kenya this rarely happens.
Consequently, the survey firms (except those that receive direct donor or parent-company funding from abroad) must scamper around to find enough sponsors so as to at least break even, given the very significant costs of conducting such surveys where (due to the spatial distribution of over the landscape) the population remains predominantly rural, with some of it in very distant and infrastructure-poor locations).
Such efforts often produce a very complex funding 'mix', with clients having various 'profiles' like: (1) Those who contribute to the firm for its work in general, but such payments are not associated with any particular poll; (2) Those clients who pay for an entire poll (and may or may not the results released to the public through the media); (3) Those clients who pay for a number of specific questions, the results to which only they receive/learn about, while other clients...(4) Pay for a number of specific questions, but allow the results (to at least some of them) to be released to the media; and so on.In such a situation, what would 'disclosure' entail? Would the cost of each question have to be identified with a funding source, would all those who funded a survey - whatever the amount - have to be identified in a list, or would something in-between be required? And how specific would the legislation have to be to cover all such variations?
: Finally (and although there are various other aspects of this issue that might also be considered), there is the question of to whom this information should be revealed.
If (and this is a very big 'if') the main concern is that the public will be influenced by poll results that are incorrect - and especially if their degree of error is the result not of technical weakness but of deliberate attempts to mislead (whatever the evidence that they can actually do so) - it should be the responsibility of the media to examine the professional capacity and integrity of the firm conducting the survey, so that the decision to public/broadcast them depends on the professional reputation of that firm, based on both past performance, and on-going comparisons with competing firms.
In a competitive (capitalist) market, credibility should count. And this also means that new firms that pop up out of nowhere in the heat of an election season should be subjected to thorough vetting before their survey results are accepted for relay to the public. On this basis, there should be no need for any restrictive legislation - except for a ban on the publication of election-related polls during the week or so immediately before voting. But since reputable Kenyan firms already impose such a (two-week) embargo on themselves, the justification for such a law seems lacking.What is the global trend? Do pollsters in other parts of the world say who funded an opinion poll?
: As noted, in advanced countries, media houses always report the funders of a poll with the results and have disclaimers like: "In a poll undertaken by Gallup that was funded by Gallup and The New York Times
..." But in countries like Kenya which still allow polls, the funding-list is usually so long and complex that it wouldn't make sense to try and publish it.Why do you think the Kenyan media is now interested in who funds opinion polls?
: I have two comments. First, the fact that such a memo was issued by the NMG management to its staff does not imply that the entire media industry has taken such a position. Second, I have no idea why NMG did this - and why now, after giving front-page coverage to polls (from reputable firms for so many years), so you better ask them! Unless, this memo aimed only at new/unknown outfits that are assumed to be about to appear (if they haven't started to appear already) as the next election approaches. This is not clear to me.What is the role of opinion polls in a democracy?
What is the impact of opinion polls in Kenya during this election year?
- They give potential candidates, political parties and interest groups, and ordinary people a more realistic idea of how the public feels about such potential/actual candidates and parties, and issues upon which their ultimate ballot-box decisions will be made.
- They (should) prevent candidates and others from making exaggerating, unrealistic claims about the popularity of the above.
- They underscore the freedom of expression simply by conducting such surveys, let alone having the results freely disseminated by the media.
- They may deter vote-rigging (though exit-polls are usually considered to be a more effective tool for this purpose).
- Assuming more than one survey firm is involved, they encourage all such firms to adhere to the highest ethical and technical standards, lest the results they produce be exposed as 'faulty', on whatever grounds.
Given Kenya's new constitution's requirement, the pre-election choice or running-mates could be very important, so that presenting these possible pairs to the public through polls (among other factors) may greatly assist candidates/parties in making such selections.
Secondly, also stemming from the new constitution, if there is a run-off contest for the presidency, it cannot be assumed that that all those who voted for each of those candidates who did not get into the second round will 'follow their wishes' and vote for their preferred choice.
As such, polls will likewise here show which of these first-round losers are more/less effective in 'guiding' his/her followers on the second round, likewise feeding value information into campaign strategies.
A critical variable in the polls impact will depend upon how uniform the poll results are across the various firms releasing them, with significant discrepancies likely in themselves to become sensitive issues in the campaigns (though of course, even absolutely uniform results could also). In addition, (and as suggested by the NMG Memo), much of the above will depend upon how much, and what sort, of coverage the media houses give the polls, and indeed, how much the candidates themselves do.