Wednesday, September 9, 2015

IRI poll: Too good to be true?

Simply put, the results of the recent opinion poll undertaken by Nielsen Bangladesh for the International Republican Institute (IRI) could barely be better for the Awami League - and could hardly be worse for the opposition Bangladesh Nationalist Party. Consistently, across all questions, the AL government does well, and far better than the BNP.

The poll was taken between the 23 May and 10 June, and involved questioning  2550 randomly selected respondents above the age of 18. Here are the key results.

- 60% of people 'like' the Awami League, compared to 29% who do not (only 42% like the BNP, with a higher percentage, 46%, disliking it) 
- 67% of people 'approve' of the job being done by the prime minister. Only 26% disapproved. 
- 66 % of people  'approve' of the job being done by the government. Only 29% disapproved. 

- 62% of people think that that the country is 'heading in the right direction' (compared with 36% who think it is headed in the wrong direction. This is an increase from 56% in Sept 2014, from 35% in January 2014, and 33% in December 2013); 
- 54% think the Awami League 'has strong leadership'. Only 22% think the same of the BNP. 
- 49% think the Awami League 'supports women'. Only 24% think the same of the BNP. 
- 47% think the Awami League 'is trustworthy'. Only 26% think the same of the BNP. 
- 46% think the Awami League 'has strong leadership'. Only 24% think the same of the BNP. 
- 42% think the Awami League 'supports democratic reform'. Only 31% think the same of the BNP. 
- 42% think the Awami League 'care about people like me.' Only 31% think the same of the BNP. 
- 42% think the Awami League 'have policies that help the poor' Only 23% think the same of the BNP. 
- 41% think the Awami League 'consistently fights corruption.' Only 25% think the same of the BNP. 
- in all 19 of the policy areas* put to them, more people rated the government's performance 'good' than 'bad', and in 17 of the 19 policy areas (other than 'Fighting corruption' and 'Ending ethnic conflict'), over 50% of people rated the government's performance as 'good'. Even in the policy area of 'democratic reforms', 54% rated the government's performance as 'good' and only 30% 'bad'. 
- 82% of the people questioned rated the economic conditions of the country as 'very good' or 'somewhat good'. 
- 61% of people questioned thought their own economic circumstances have improved over the last year (compared with 49% in September 2014), and 63% think it will improve in the coming year.
Polling skepticism
But are the results simply too good to be true?

There is general widespread skepticism within Bangladesh about the accuracy of opinion polls. This is in part due to (a) the widespread lack of understanding about the science of randomized polling with people more willing to accept anecdotal information as true; (b) a general lack of confidence in the independence and professionalism of Bangladesh polling organizations; (c) the small number of polls undertaken which has prevented people from checking consistency; and (d) the absence of a recent situation where the accuracy of the polls had been put to the test - through for example free and fair election. (Whilst there were quite a number of polls published during and at the end of 2013, some months before the January 2014 elections, all of which suggested that the BNP was ahead of the AL, there was no subsequent free and fair elections at which these polls could be tested. Had there been a proper election in which the BNP took part, and won then people may have started to have more faith in the polls.)

The skepticism towards this particular poll is likely to be even greater than normal, as its results cuts across the views held by many ordinary people and independent observers - which is that, in the context of the country's, corruption, mal-administration, lack of democracy and the poor human rights situation, the Awami League government is far from popular, and were there to be free and fair elections (for good or for bad) the BNP would have a good chance of winning.

Of course the fact that the poll cuts across established notions or views, is no reason in itself to disbelief it. The purpose of polls is to replace anecdotal speculation with a more scientific analysis of views of a population.

Moreover, the poll was conducted by Nielson, Bangladesh which have done many of the earlier Democracy International polls, and was supervised by an independent polling expert. The results of the poll also reflects a trend in other polls which have suggested that (despite the views of the chattering classes) since January 2014 the Awami League has gained in national popularity over the BNP.

BNP Objections
It is not surprising of course that the BNP has tried to question the validity of the poll - but the arguments they have put forward are miserably weak.

According to press reports BNP spokesperson Asaduzzaman Ripon said at a press conference that: “On one hand, they say the people want a caretaker government to supervise the polls and most people said they want a swift election. …. But on the other hand, they say the government’s popularity is rising.'

These two views however are not contradictory. A person can think the Awami League is doing a good job, and also think that there should be caretaker government.

The spokesperson also is reported to have said: “There is a difference between the survey question and answer. The organisation conducted the survey— Nielsen Bangladesh—conducted a survey in 2012 which was published in the Daily Star and the Prothom Alo, said if the election was held that time BNP will form the government. The same organisation is now saying government's popularity is increasing.” 

He seems to be suggesting that it is inconsistent for poll results to change over time. This is of course ridiculous. It is the very purpose of a poll to find out if there has been a change of opinion - so a change of opinion reflected in a change in poll numbers can obviously not be used to discredit a poll.

The chilling effect?
However, there is one aspect of the poll which does require some consideration - and this is whether the current political climate and its 'chilling' effect on freedom of speech could have skewed the results.

Anyone living in Bangladesh these days can appreciate that in the last year vocal opposition to the government increasingly brings with it risks and dangers, including for some prosecution and detention. The poll could potentially be effected by the current political environment in two ways.

First, due to concerns respondents may have had about the identify of the people who were asking the questions and the confidentiality of what they had to say, a proportion of those who agreed to respond might have been scared to give full and frank views critical of the government. This would have resulted in a more pro-government opinions being given.

Secondly, and this is linked to the first point, the sample of people interviewed could have been disproportionately pro-AL, skewing the results. The reason why this might have happened is because the response rate for the poll was 47% - that is to say that only 47% of those initially randomly chosen to be questioned agreed to respond. Whilst this is not a low percentage when compared to polls in other countries (where it can often be as low as 9%), it might be significant in the context of Bangladesh (with its current 'chilling effect' on freedom of speech) if a disproportionately high number of those who did not want to respond were people who would have been critical of the government (but did not want to take part fearing repercussions of giving their views).

Potentially this could have the effect of making the final sample of 2550 include disproportionately high number of Awami League supporters, making it unrepresentative of the country as a whole. If this would the case the results would not be accurate.

This is not some arcane point. In November 2013, a poll was undertaken by the polling company MRC Mode for the Awami League. The independent experts advising the AL, suggested that the raw results should be adjusted as the sample contained a disproportionately high number of previous-AL supporting voters. The experts knew that there was a disproportionate number as the poll had asked the respondents how they had voted in the 2009 election. After making this adjustment, the BNP had a 3 per cent lead. However, the Awami League did not do so and the party made public its results which showed that the AL was 6% ahead. So taking into account past political preference to ensure that the sample of respondents  is not skewed to one party or another can be significant.

It is not known whether the final IRI sample did contain a similarly disproportionate number of AL supporters because the IRI poll (unlike the polls for example taken by Democracy International in 2013 and 2014) did not ask any questions on this issue.

It should be noted, though, that none of the previous IRI polls (which included two that came up with more pro-opposition results) were adjusted to take into account past political preference - so IRI's methodology has been consistent.

Where does that take us?
It is not possible to know with any certainty whether the so-called 'chilling effect' may be an explanation for the poll's pro-AL results - but there are reasons to be cautious before observers come to the conclusion that this is the case.

Whilst the situation of freedom of speech in Bangladesh is certainly worse now than November 2013 and January 2014, which was when IRI did two polls that were more pro-opposition, it is not clear that the situation has got so much worse for ordinary people that this can explain such a shift in the results.

And arguably, if there was such a chilling effect, one would not expect 67% of these same respondents stating that they supported the introduction of the caretaker government - a policy supported by the opposition, not the government - which is what they did in this poll.

In the round, it is probably best to assume that this poll is a relatively accurate snapshot of the attitudes of the county. That being so, how can we explain the significant shift that has taken place in the last two years - from a position in which most people thought that the country was moving in the wrong direction to one where most people think it is moving in the right one, and where most people have a positive view of the government and its achievements? Perhaps some or all of these factors set out below are responsible

- the campaign of the opposition in the final months of 2013 and the first few months of 2015, which caused huge disruption in peoples' lives and which at times turned violent with the burning to death of civilians, had resulted in increasing negative opinions of the opposition;

- the increased stability in Bangladesh, brought about by the government bringing the opposition protests to an end, has made a positive difference to people's lives, and has resulted in more positive perception of the government;

- human rights violations on the part of the government's law enforcing authorities (disappearances and extra judicial killings etc) effect a small minority of people and mostly concern those involved in opposition politics, and are not of great concern to most people;

- actions by the government to prevent the BNP from functioning as an opposition party, by detaining them in jail or keeping them caught up in judicial processes, has meant an almost total absence of opposition protests and vocalization of their concerns which has benefited the government;

- Bangladeshis have read the writing on the wall, and recognize that the opposition has been roundly defeated by the Awami League government, which is the only game in town and so are willing to give the party and government their support;

- the decrease in criticism of the government in the electronic and other media;

- the government has continued to push a pro-development agenda, which is popular with people.


* The policy areas were: Improving education; Health care; Water and Sanitation; Infrastructure development; Water availability; Gender Equality; Keeping the peace; Inflation; Employment; Poverty Reduction; Electricity availability; Dealing with hartals; Housing; Democratic reforms; Fighting crime; Ending Sectarian conflict; Land rights; Fighting Corruption; and Ending Ethnic conflict.

No comments:

Post a Comment