Why opinion polls cannot be taken seriously

Written by December 5, 2017 16:39

Will the opinion polls get it wrong vis-a-vis Gujarat assembly elections? Will it be proved again that opinion polls cannot predict modern democracy at work? Is it logical and right that human behaviour is never predictable in a mechanical way? These are the questions which assume significance in the wake of an opinion poll on December 4 predicting that the Gujarat elections could be a dead heat between Bharatiya Janata Party (BJP) and Congress with the former just scraping through to a win.

The Lokniti-CSDS poll done for ABP News indicated that both parties would get 43% of the votes. The pollsters have predicted neck and neck between Congress and BJP in Gujarat.  Significantly, these projections are a remarkable turnaround from the 30-point lead for BJP projected by a poll done by the same agencies in August.

But the question is whether the pollsters are good at dealing with probability. What needs to be done is to apply judgement to the findings as they have proved to be wrong in the past several elections especially where a close race was predicted. Remember, ABP News-Lokniti opinion poll had released an opinion poll ahead of elections in Uttar Pradesh which showed the governing Samajwadi Party (SP) and BJP running neck and neck.

But the UP results came out they contradicted the opinion poll as the BJP registered a huge victory over SP. Similar was the case when in the run-up to Bihar Assembly elections last year, ABP News had predicted that BJP was doing very well in Bihar, but the actual results proved to be the exact opposite. It may also be recalled how the opinion polls had given the BJP just 91 seats in 2012 assembly polls and the saffron party got 116 seats.

This brings up the question: why does anyone take these opinion polls seriously? So in the light of these instances, how can one believe the opinion polls of Monday predicting a BJP and Congress neck and neck in Gujarat. The ruling party of Gujarat may have sweeping performance as in the past instances.

According to information, the pollsters heavily rely on in-person surveys. What does it mean? It means that in order to increase the coverage the interviewers need to cover all the areas the voters come from. What happens is that the remote voters living mostly in poor Dalit villages are hardly interviewed.

In addition to this, if they are approached somehow then they do not speak up the truth due to societal pressure. These people from suppressed class also prefer not to speak because of fear of dominant castes. This adjustment related to the social dynamics is something the agency needs to make and they do not do it.

Moreover, the agency does their survey keeping in mind the anti-incumbency factor also. In the case of Gujarat, the agency must have kept in mind the 22 years of BJP rule in the state. These factors lead to agencies undermining the ruling party’s chances to perform better than the rivals.