ibnlive » Chat

4 pm Nov 05, 2013

Is a ban on pre-poll surveys an undemocratic idea?

The Congress is demanding a ban on pre-poll surveys. It argues that the voters will be influenced by this exercise. It is a desperate reaction of rattled Congress or an argument with substance? Join a live web chat with Dr. Sanjay Kumar, an expert in poll survey at 4 PM on Tuesday.
26 questions answered
  • When Pre poll survey anyway not going to affect political parties, why congress demand for ban of pre poll surveys. Any reasons? Asked by: r srinivasan
  • Sanjay Kumar Only because all the surveys are showing Congress trailing behind and Congress fears that the support of the party may slide down further once people get to know that Congress is a loosing party in this election.
  • Polls should not be banned but as AAP says the raw data must be made public. What do you say ?? Asked by: Vijay Iyer
  • Sanjay Kumar I am all for greater transparency of the polls, but I fail to understand how useful it may be to put all the raw data file on the website How many people will be able to analyse the raw data files. My questions to the supports of AAP is if they had put the data file on the website if they were trailing in their opinion poll.
  • Sir if you put Unit level respondent-wise data, I don't think anyone can question your credibility.Why don't you put it and let everything be public? Asked by: shivam jain
  • Sanjay Kumar This is not the first time that CSDS has conducted the surveys. We at CSDS has a policy not to put the raw data file on the website and I possibly cannot violate this policy. Also one must understand that the surveys conducted by CSDS not only only ask question on voting, there are various other questions which we researcher at CSDS want to use for our own research. As someone who has put in effort in collecting data should have the first right of data usage for research purpose and if we put all the data files on the website we will loose that advantage. We do put on our website the details of methodology, the questionnaire and the basic findings of the entire surveys all the questionnaire. We do respond to specific question which people may have regarding finding of the surveys.
  • How accurate and representative are these poll survey results? Asked by: Shashank Sharma
  • Sanjay Kumar The surveys gives you estimate of the situation prevailing at the time when survey was conducted, it may not tell you if the results are going to be the same on the election day if election are still few weeks away from the day survey was conducted. The difficulty is after looking at the surveys we all tend to conclude that the result will be the same as shown by the pre poll survey.
  • What is the purpose of opinion polls and what do you think are the effects of opinion polls on viewers? Asked by: chandu
  • Sanjay Kumar The purpose of opinion poll should be much more than merely estimating vote share and seat share for parties. It should be used as a useful tool for assessing people's views on issues which parties should address, concern of common voters etc. It is important tool for measuring how people might vote, but sadly the opinion poll has turned into an instrument of predicting the magical number.
  • Why congress hasn't opposed during karnataka election? They arereally infavour of prepole survey during that time. Asked by: Sudhanandagowda
  • Sanjay Kumar It is for the Congress to reply ,we all know the reason at this stage.
  • While opinion polls could influence the voter's mind, blogs, editorials and articles can also influence the voter. Hence, would government ban these forms of expression as well? Asked by: rsg
  • Sanjay Kumar Who else than the government would know the answer to this question. If they think of banning opinion poll since it influence voters, then many other things also influence voting decision.
  • Do we see the return of glimpses of emergency? or will we see the beginning of an end of Congress? Asked by: Satish Kumar
  • Sanjay Kumar Possibly we cannot compare this with Emergency, this is just reflection of desperation of the Congress leaders, they don't know what to do, if they had any say in postponing or cancelling election they might have even thought of that option
  • Don't you think opinion poll forecasts can sometimes force parties to enter into a pre-poll understanding? It also benefits them in a way? Asked by: Sri
  • Sanjay Kumar There is a possibility of that happening, but how can this happen if one party does not have any trust on survey and some other has trust on survey. at least one need to have some trust on survey and then only this could be possible.
  • To what extent do opinion polls influence the undecided voter? Does the frontrunner in such polls gain an unfair advantage if a hung parliamant / assembly is forecasted (eg: UP 2012)? Asked by: Sri
  • Sanjay Kumar There is hardly any systematic research which indicated how much does opinion poll influence voters, but I personally believe it does not influence voters much. What ever influence is there it is both ways, if there is some bandwagon effect, there is also an effect of some groups mobilising in favour of party which may be seen as loosing, this may apply to some minority community may wish to defeat a party and vote for a party which may not be their first choice, but may shift their vote just to defeat the party which they do not like
  • Hi Sanjay Why congress was MUM when there was an opinion poll/pre poll survey during Karnataka assembly elections?This clearly shows their double standards.When the wind is blowing the opposite way they start trying all kinds of gimmicks..They never thought Modi would create such an impact across the country..Am I right? Asked by: Prasad KS
  • Sanjay Kumar Yes you are right.
  • Sir Why don't opinion polls put full raw data(not just selected tables)on websites.This will be befitting reply to those who oppose opinion polls. Asked by: shivam jain
  • Sanjay Kumar We have tried to be as transparent as possible, put all the methodological details on the website, i do not think there is need to put the data files on the website. If tomorrow critique say how do I believe that respondents were telling the truth how do I make them believe that they were telling a truth. It could be an unending story.
  • Why political parties sometimes talk on ban and sometimes support these surveys? is it only when they see personnel gains? Asked by: Pawan
  • Sanjay Kumar Yes there is a personal bias inherent in such a demand by political party. We know political parties have shifted their stand on this issues in the past.
  • Why are Prepoll surveys being treated as larger than life ? What is the need for analyzing them in such depth in the first place ? Asked by: Neeta
  • Sanjay Kumar I have mentioned that pre poll surveys gives estimate of the mood of the people at the time survey was conducted, one should not conclude that the results will the same as suggest in the prepoll survey. I don't know why people get so excited about this and treat this as final results. Things might change with regard to popularity of the party and the results may eventually be very different on the election/counting day.
  • Isn't asking a ban on corruption/ illegal funding in elections and influencing of voters by giving away money, gifts and alcohol the right way to go about it? Asked by: Dhawal
  • Sanjay Kumar This may be more useful than asking for a ban on election surveys.
  • Sir, CSDS may indeed be THE premier research agency with world-class / world-leading research standards but please do bear with my numerous questions & give your considered response. This is in the interest of clearing whatever doubts there are on the reliability & validity of these research endeavors. - Would you be kind enough to share with us the (draft at-least) survey instrument? - Is this survey conducted online as well as in-person? - If this is a questionnaire based survey, is it also multi-lingual? - Does the interviewer conduct a screening process to assess the language preference of the respondent before handing over the relevant survey form in preferred language? - Have any questions be posed to validate the consistency of respondent's choice? If so, what is the % of inconsitent responses? - When was pilot test conducted and when the main research? Asked by: Raghu RK
  • Sanjay Kumar The instrument is posted on our website www.lokniti.org No the survey is face to face not online survey. IF the survey is conducted in various states the questionnaire is translated in various languages, but not in this survey since all the Hindi speaking states. We use local language questionnaire keeping this in mind that the respondent will be more comfortable in local language. We did not do any pilot for this election survey. Pilot is done to test the validity of the questions, Since most of the questions have been asked in the past in our other surveys, we did not feel the compulsion of doing the pilot, but yes for many other surveys we do pilot surveys before doing the final survey
  • Banning anything is a bad idea. But I dont think, pre-poll surveys are often biased. Also, there are too many outliers and assumptions in the surveys done. Hence, in order to ensure that people do not get influenced by surveys, some regulations are mandatory. It is also a good idea to have exit polls, far more accurate than Opinion polls Asked by: Narayan
  • Sanjay Kumar In principle Exit poll should be more accurate since the people are interviewed after they have cast their vote and there is no possibility of their changing their mind. But voters can change their preference after the pre poll survey during campaign.
  • From ground report, Raj seem to be tuff fight between Congress & BJP. This is based on speaking to locals. R u sure about Raj opinion polls Asked by: Amit Sinha
  • Sanjay Kumar We are sure that the BJP was much ahead of the Congress when the survey was conducted in Raj between October 14-22.
  • Dr. Sanjay, is there any credible evidence to suggest that these poll surveys actually influence a voter’s decision?..what is the appropriate sample size that needs to be surveyed in order for it to reflect results which are closer to reality?..these surveys never mention if and what cross section of people were interviewed in any constituency / state..More often than not how is it that television anchors sell these surveys results as reality in spite of being proved wrong on various occasions?..Your thoughts on these points please. Asked by: Jai
  • Sanjay Kumar Most of these surveys has to be cross section of respondents as we have cross section of voters. At least we give details about the profile of the sample for every survey which CSDS conduct.
  • All political parties are alike. They wants ban when they are predicted to loose. They welcome when they are predicted to win. we should go on dong the show, our country is democratic. Asked by: Mirza K Baig
  • Sanjay Kumar But I think we must do these survey and shows with some responsibility. We must try to be transparent about the methodology of the survey
  • Dont you think such activities is exposing Congress even more giving the idea that it is going to all limits to stop BJP? Does it not indicate Congress has morally lost elections ? Asked by: Ibrahim
  • Sanjay Kumar Yes this is desperation on the part of Congress
  • Is it possible that a biased poll result can affect final results. As it is we have enough number of paid news, so why not paid surveys. Asked by: kpvidya1999
  • Sanjay Kumar There could be some paid surveys like paid news, but like not all news are paid news , not all survey are paid surveys
  • What is your take on the latest controversy surrounding opinion polls? Do you think opinions can be manufactured or manipulated? If yes, how? Does the Internet play in determining perceptions and answers in an opinion poll? Asked by: Akshay Iyer
  • Sanjay Kumar Internet can not be an effective instrument for doing survey of cross section of voters since large number of voters do not use internet. One need to do face to face interview for election surveys. Yes opinion polls could manufacture opinions, but not many are engage in such business, some may do that but they would not survive long and will get exposued
  • I am in favour of PRE-POLL OPINION POLLS. But Agencies should ask atleast 600 persons from each constituency (with atleast 40% women). (In case of lok-sabha atleast 750 persons per seat) and raw-data should be made public as Mr Arvind kejriwal is saying... Rest all is fine... Asked by: Jagdeep Singh
  • Sanjay Kumar So can you calculate what may be the sample size for doing the survey for Lok Sabha. Let Lok Sabha election come and Arvind Kejriwal do a survey in all 543 Lok Sabha constituencies with the sample size what you have mentioned in each constituency
  • Instead of demanding ban on Pre-poll surveys political parties should publish survey data with transparency. All parties trailing in pre-poll surveys dissent with survey & demand ban. In 2009 BJP & in 2013 congress demanding ban. Asked by: bankim
  • Sanjay Kumar Yes I agree I support greater transparency while publishing opinion poll
  • Based on records, how far does a pre-poll survey correlate with the actual results? Asked by: Sarkar
  • Sanjay Kumar Lot of them actually correlate with actual results, only few were horribly wrong like 2004 and 2009 and people keep giving example of these two, at the best they would also cite UP and Bihar where many surveys did not get the magnitude the victory right. But CSDS did get right in UP, Bihar, Assam and and list is longer than you may think.

Hosted by

Sanjay Kumar
Fellow, CSDS