In a week, people of Bangladesh will be voting in a historic election. Questions abound in the social circles:
Undoubtedly, people are curious, concerned, and conscious.
And those of us, who are in the business of studying people, society, and the economy, are having a hard time making sense of the data -- qualitative, quantitative, and observational.
What is known?
We are sure that this will not be an easy path to victory for BNP and it is not even decided whether the party will win.
Post August 5, 2024, the party and its leadership might have spontaneously assumed that they are in power. Their language and approach became the one of a government in waiting.
Therefore, they rejected the first wave of polls done by Innovision Consulting in September 2024 and BIGD later in October 2024 that showed about a 5 percentage-point margin between BNP and Jamaat on the total voter distribution.
Innovision’s poll showed a distinct difference in perception between online and offline respondents; online sampling was leaning towards Jamaat.
Subsequently, when both the next rounds of Innovision and BIGD polls picked the same preference, both Jamaat and BNP began to pay attention but in different ways.
Innovision ’s nationwide survey in February 2025 also showed how Jamaat was slowly winning the share of student voters, Gen Z voters (aged 18-28) while BNP controlled millennials (28-45) and the Gen X voters (45-60).
The divisional divides were also clear, Jamaat was surging in North-West and the South-West. BNP was holding base in Central, North-East, and South-East.
Most importantly the Innovision and BIGD surveys showed a large percentage of the voters being undecided and unrevealed and the low percentage of voters preferring the student led party- NCP.
This time, the BNP and Jamaat vote shares remained static from February 2025 -- BNP 41% and Jamaat hovering around 30-31%. On the total sample, about one-third remained undecided and total unrevealed (including undecided) reduced from around 53% to 45%.
Surprisingly, the higher proportion of the revealed share was tilting towards Awami League. They gained about 4 percentage points in the Innovision survey in September 2025 in contrast to February 2025.
Later, an IRI survey released in November 2025 showed that Jamaat is at around 29%, BNP is at around 35%.
Another poll done by NARRATIV in the period November to December 2025 and released in January 2026 posted BNP and Jamaat neck to neck at 35% with Jamaat behind by just about 1%.
What is common in these polls?
That they were still about one-third undecided and unrevealed and that BNP and Jamaat are both hovering at around one third and the gap between them is at around 3-6%.
All of these suggest a hugely contested poll.
Set on this background, Innovision undertook a panel survey on the respondents sampled in the February 2025 and September 2025 samples.
From over 20,000 samples from across the country, a panel of 10,000 samples was created with an intended attrition rate of 55%. Eventually, 5,147 samples were successfully sampled from around 15,000 samples, with a much higher attrition rate (around 67%).
Failure to reach female respondents increased the attrition rate. But Innovision’s report explains that there are no systematic biases that could be found in those who could not be reached as majority of the attrition came from those who did not pick the phone.
After systematic analysis among the non-responsive samples, there is no evidence that a specific strata or voting segment was left out. Nevertheless, given the attrition (samples that could not be reached) read the following with a bit of statistical caution.
Unlike a cross-sectional survey that is done on a type of sample or class of sample, a panel survey is repeated on the same sample. The panel survey is thus able to show the difference of a sample’s response over time.
Innovision’s panel survey thus provides a very important nuance that the other surveys are not able to project. The panel survey explains the shifts in voters preferences as we are able to compare the responses of the samples from the previous round of polling.
In this round, around 75% reported that they have decided who they will vote for (the highest since the polling started) and among the decided, the highest proportion (88%) revealed their preference.
The result showed a widening gap between BNP and Jamaat (at 22.1%) with 52.8% respondents reporting they will vote for BNP and 31% reporting they will vote for Jamaat.
The voting preferences by age groups, gender, location, and occupation remained generally similar in terms of direction but broke more in favour of BNP.
The results made us wonder as to what might be happening.
The general hypothesis from reading the campaigns of the two parties is that the election is closely contested. Then what is causing a shift in the observed pattern in the panel survey.
The bigger question is: Will it hold?
We dug through the data and found an interesting pattern that would explain this election.
The survey shows that of the BNP votes (52.8%), 26.6% came from previously undecided and unrevealed. In contrast, of the Jamaat votes (31%), 14.1% came from the unrevealed.
Also, of the Awami League votes, BNP and alliance was getting 32.9%, Jamaat was getting 13.2% and 52.3% remained undecided and unrevealed.
The result also showed a shift from Jamaat to BNP and BNP to Jamaat and from NCP to Jamaat and BNP (leaning more towards Jamaat). The results showed, Jamaat and NCP voters are more likely to vote yes to the referendum and the BNP and Awami League voters are less likely to vote Yes for the referendum. About 59% are likely to vote Yes.
Put all these together and you will see a pattern.
What it all means
This shows that the nation is in flux. They are voting for a change for sure. That Jamaat has secured more than one third of the votes already is a signal towards change.
The findings show that there is a firewall that is protecting BNP for now -- a large share of voters are having second thoughts about their party preference which is giving BNP a lead.
In fact, when asked about their second preference to form the government, Jamaat is ahead of BNP by about 13 percentage points.
Also note, 47% think Tarique Rahman will be the next prime minister but around 52% will vote for BNP. Even some fraction of the BNP voters are not sure about BNP winning the election.
Also, factor this, Jamaat is directly contesting in 224 constituencies which is far less than BNP is directly contesting in. When asked whether the voters know about the candidates, more voters reported positively in favour of the BNP candidate than the Jamaat candidate.
If you factor these points and look at Jamaat contesting in a heavily concentrated manner with targeted campaigns, and if you consider the effect of first time Jamaat voters having last minute doubts, and if you consider the fact that Jamaat has secured one-third already, the last part of the undecided and unrevealed, along with a good part of those who revealed, might switch to Jamaat.
With a week to the election, it is safe to say that the competition is far from over.
An eventual margin of national vote share is unlikely to go beyond 12% and the seat share might be even closer. Jamaat might win a lot of seats competitively.
As such, be ready for a nail-biting finish.
Md Rubaiyath Sarwar is Managing Director, Innovision Consulting.
