On the new opinion polls

Posted in AL, BNP, democracy, elections, media, politics by jrahman on January 8, 2013

I was in Dhaka during the 2008 election.  The day before the election, I told Asif Saleh that BNP was making a remarkable comeback and the election would be very tight.  I was, of course, way off.  Turns out so were pundits like Nayeemul Islam Khan, Asif Nazrul, Mahmudur Rahman and Nazim Kamran Chowdhury — who all noticed a massive momentum towards BNP.  I was reminded of this episode last November, when Republican spinmeister Karl Rove refused to accept election results as they were coming in — apparently it wasn’t consistent with the momentum (Mittmentum) he had observed. 

I (and more famous Deshi pundits) had an excuse.  We didn’t have any proper opinion poll or survey data to guide our thinking.  One pundit who did see such data — Zafar Sobhan — did predict an Awami landslide, and he was proved right.  Of course, Rove and his ilk didn’t have such excuse.  In America, people like Nate Silver looked at the polls and other relevant information and predicted the final election outcome quite accurately. 

Compared with America (and other advanced democracies), opinion polls are still few and far between in Bangladesh.  But compared with 2008, we now have regular polls by Daily Star and Prothom Alo.  Good luck to anyone who believes they know the public pulse and don’t care for polls.  Personally, I have no idea what the public believes, so I find these polls very interesting. 

Here is the Daily Star survey, done by Centre for Strategic Research.  Here are detailed results of Prothom Alo survey, conducted by ORG Quest (here is its news report, here is the methodology).  As far as I can tell, these polls are done in the same way similar polls are done elsewhere.  There are margins of error, and the polls are indicative of public opinion, not an exact predictor of anything. 

With those caveats in mind, I think these polls should make BNP and Ershad supporters optimistic, while AL should be quite worried.  The polls also hold interesting results for third force enthusiasts.

Who would win the election if one were held now?

Obviously that’s the key question anyone reading these polls most care about.  According to the DS-CSR poll, AL would have received 42% vote against BNP’s 39%.  The PA-ORG poll gives BNP 44% against AL’s 35%.  Given the margins of error, and possible difference in exact questions, let’s round these numbers into fractions.  What these polls tell us is that at least four out of every ten people would vote for BNP had there been an election now, while another four would vote for AL.

This is consistent with the 40-40 politics observed two years ago. 

This is self-evidently good news for BNP, not so much for AL. 

A key question is, where will the other two out of ten people go?  According to the PA-ORG poll, Jatiya Party is supported by 12%, but DS-CSR puts JP’s support at 5%.  Even if the true support for JP is towards the lower end of that range, it seems to me that without JP support, AL is heading for an electoral disaster of at least 2001 proportion.  This should make HM Ershad very happy indeed.

In fact, it gets worse for AL.  According to the DS-CSR poll, only about a quarter believe that a fair election is possible under a party-led government, and again, only a quarter support the cancellation of the caretaker system.  Three-quarters of those surveyed by PA-ORG said that a fair election is not possible without a caretaker government.  That is, even some people who would otherwise vote for AL believes AL can’t hold a free and fair election.

One way to interpret these results is that even if we think AL is supported by something like 40% of possible voters, only about 25% are its hard core supporters. 

Meanwhile, the third force supporters should note that two-fifths of DS-CSR respondents think political culture is worsening, while over half of those polled by PA-ORG believe violent politics will return, and three-fifths expect no compromise between the two parties on the caretaker issue.  One interepretation of these results is that most people would not be surprised by a 1996 or 2007 like situation.

But the third force supporters should also note that both Hasina Wajed and Khaleda Zia have high net approval ratings (approved minus disapproved), suggesting that minus-2 will not work.  It’s also important to realise that these approval ratings aren’t head to head popularity contests — that is, one can’t say either leader is more popular than the other, because the surveys don’t ask that question. 

Both polls also suggest that Nurul Islam Nahid and Obaidul Qadir are considered as successful ministers, while AMA Muhith is considered as a failure.  Those looking for a leftist bias should note that the PA-ORG poll finds as high a disapproval rating for Motia Chowdhury as for Mr Muhith. 

Finally, let me end with what I found to be the most interesting thing in these polls — this chart from the DS-CSR survey that tells us what will drive the next election result.


Listening to or reading various pundits, one might get the impression that the war crimes trial issue will determine the next election.  Well, this is just one opinion poll.  But it is still a more scientific guide than anything you get from pundits.  And according to this, inflation and corruption are the factors that will decide the next election.


4 Responses

Subscribe to comments with RSS.

  1. Rumi said, on January 8, 2013 at 8:48 pm

    1. In 1977 election AL bagged 25% votes. So that 25% will be with AL down the generations.
    2. In your last point regarding issues needing highest attention — was it open ended or closed question? Was WCT among the options? do not know, do you know?
    3. DS and PA sponsored polls show 4-5% variation in reverse directions. Statistically not impossible, but very improbable.
    4. While DS says AL enjoys 42% support while 48% are happy with Hasina. Similar pattern is seen in PA poll too.
    5. Similarly PA sponsored polls say BNP support went up to 44% while the same poll says 58% are unhappy with BNP.
    6. #4 and 5 do not make any sense to me.
    7. I think there are two undercurrents here. One, our newspapers are trying to editorialize the polls to serve their own agenda. They have not yet learnt how to editorialize poll results. That’s why we are seeing such cockamamie reporting. Second, our population is still naive to polls. Many people did not understand the questions or when answering the question they were careful in giving a balanced answer like AL is doing bad but Hasina is not bad. Otherwise how the hell in the world one can support present day Hasina but not support AL or vote for BNP?

  2. shafiq said, on January 9, 2013 at 10:34 am

    Nate Silver famously said in the Colbert Show that he regards most of the Cable TV Pundits lower than Ebola virus. There are some validity to his low esteem. Many Pundits still do not have the necessary statistics savvy to appreciate the remarkable usefulness of professional polls. Many of them went on bloviating as if they have a mysterious connection to the pulse of the public opinion. After the 2012 election, no politician or pundit in America will ever disparage and disregard aggregate polls. Already we are seeing that Republicans and Democrats are paying acute attention to the shift in popular opinion. The quick collapse of the Republican resistance to Obama’s tax hike happened mostly because Republican senators and bigwigs saw how people are viewing the impasse.

    I don’t watch talk shows of the Bangladesh so I cannot say whether this two polls have caused significant ripples in the talk universe. But I was surprised to see a remarkable silence in the column and bloggosphere. Few people are opining about this most important information. We only have two respectable polls annually and this is the only good information about public opinion in a sea of dross. Why aren’t more polls taking place? I think there is tremendous hunger for good information among the people.

    PS. As far as I can recall, in previous years also Awami League seemed to be better off in DS polls than PO polls. Is there any interesting pattern there?

  3. jrahman said, on January 9, 2013 at 12:19 pm

    Rumi bhai, I am pretty sure both the editors are trying to put their own spin, and I am pretty sure they are doing this knowingly. As I wrote in the previous post, they have their agenda.

    You may be right about general public being ‘too naive’. But I have a different explanation. Back in January 2007, when the anti-BNP feeling was at its highest, I noticed something curious in Dhaka — people in cha-paan-cigarette stalls or barbershops or public transport blamed everyone from the Prince of Hawa Bhaban to local BNP ward commissioner, but still said “Begum Zia is a good person, but she was being deceived/misled by wrong people”. In contrast, the highly educated bhadraloks were generally very disparaging of both the leaders.

    I have no way of proving this, but i think this is how the public actually views both leaders. I think they are both genuinely popular.

    As for why 44% support BNP and yet 58% are unhappy with it (and similar results with AL) — the answer is simple. The two questions are different. One asks you to choose between AL and BNP. The other asks for your opinion of BNP. One could say they are not happy with BNP, but still choose BNP over AL because they are even less happy with AL. In fact, some hard core BNP supporters may be unhappy with the party for its recent shift of India policy. However, this doesn’t mean they will vote AL.

    If there is any naivete among anyone about the polls, it is really the chattering classes. As Shafiq says, bloggers are silent on this. As are fb-ers. And even the talk show stars are relatively quiet. In fact, if I were to be cynical, I would say that the chattering classes hate these polls because they prefer to believe they know-it-all and can tell what the public thinks better than the public.

    Shafiq, DS poll showed better result for AL in 2012, but in earlier polls PA showed better result for AL. There are too few polls to say whether there’s any systematic bias between the two.

    Finally, I don’t know whether the “issues needing attention” question was open ended in this poll. In 2009 it was an open ended question. Even then, war crimes trial was mentioned by a very small number of respondents. The trial will swing the election is one of those things some pundits love to claim, even though there is no evidence for it.

  4. Nabin said, on January 19, 2013 at 12:00 pm

    For sure, the media has its own agenda; So I choose not to pay more focus on either of those statistics produced by DS-CSR or PA-ORG ! Given Joyti’s last statistics “…Issues That Need Highest Attention….” , here is my analysis :

    1. Corruption itself becomes a well established institution in Bangladesh, so I believe that both BNP and AL’s agenda will hardly matter to eradicate such institution (unless they have new policy with `big-push’, at this stage none of them has ‘guts’ to do that!).

    2. Inflation has always been one of the key issues in Bangladesh and these two parties hardly take any effective measure to control it.

    Therefore, if the both parties focus only on these two issues (given their previous and current histories), I do not think any of them has clear cut winning position, it is just 40-40 politics as the statistics told us two years back. It is not because the general voters like those parties but they have no other alternatives at this point of time (to me it sounds like a curse of the famous median voter theorem based on two-party election !!).

    3. War crime trial does matter only if BNP chooses a wrong strategy in supporting Jamat, to me so far BNP has maintained a distance on this issue with Jamat; as long as BNP maintains its current position, this trial is hardly matter to election. Indeed, BNP’s current position might help them in getting swing votes. On the other hand, AL can do little better by ensuring that the war crime trial is finished smoothly.

    4. Keeping in mind that there are some benefits of being an incumbent and it goes to AL; such as load-shedding is relatively low (despite its debatable short time solutions!), Education sector is performing relatively better etc. This might affect the swing voters if it does then again we end up in 40-40 politics.

    Given the above analysis, to me the decisive issue for this coming election will be organising a “fair-election”. As long as the incumbent party brings the confidence to every voters (whatever means they adopt ie alternate form of care taker government or something else that is acceptable to our opposition parties as well as general voters) that there will be a fair election (no matter what it costs to incumbent party) this can change the whole ball game and can bring the winning ticket to them. It may sound strange to all of you but this has been and will always be the decisive issue in Bangladeshi democracy !

Comments are closed.

%d bloggers like this: