This Article is From May 14, 2014

Lok Sabha Exit and Post Poll - Methodology

The "Exit Poll" analysis and forecasts are based on combining two separate survey methodologies: i) Polling Booth surveys ii) Post Poll Household surveys.

1.    Polling Booth Exit Polls: a sample size 63740 voters - involves randomly selecting polling booths across the country and as voters emerge from these polling booths a randomised selection of voters are questioned on which party they voted for.

2.    Post Poll Exit Polls: a sample size of  91712  voters.

The fieldwork and the tabulation for the Polling Booth Exit Polls and the Post Poll Exit Polls was carried out by Hansa Research.

The detailed methodology is set out below:

EXIT POLLS

Overview: Exit Polls were conducted at polling booths on election days. Voters were interviewed as they came out of polling booths. Interviewers were asked to interview every fifth voter and ask him to indicate the candidate/party for which he had voted.

Constituencies Covered: The Exit Polls covered 265 constituencies across 19 major states. The proportion of constituencies covered was higher in states where major political changes had happened. Nearly all Parliamentary constituencies were covered in states like AP / Telengana, Tamil Nadu, West Bengal, Bihar, Jharkhand and Delhi. Parliamentary Constituencies chosen for the Exit Polls were representative of the socio-political regions within the state.

Coverage and Sampling: In each Parliamentary Constituency (PC), 3 Assembly Constituencies (ACs) were randomly selected. Thus, a total of 795 ACs were sampled across the country. 3150 Polling Booths were randomly selected from these ACs for the purposes of the Exit Poll.

Sample Size: A total of 63740 interviews were conducted for the Exit Poll.

Interview Method: Each booth was covered for one hour by two interviewers. They then moved to another booth. All interviews were conducted using tablet computers, where the respondents chose the candidate and party symbol displayed on the screen, without exposing their choice to the interviewers.

Data Weighting: All data was weighted to reflect the constituency profile on gender and religion/caste.  

POST POLLS

Overview: Post Polls were conducted with voters randomly selected from the electoral rolls after the elections. The interviews were conducted 1 - 8 days after they had voted

Constituencies Covered: The Post Polls covered 330 constituencies across 19 major states. As with the Exit Polls, the proportion of constituencies covered was higher in states where major political changes had happened. Nearly all Parliamentary constituencies were covered in states like AP / Telengana, Tamil Nadu, West Bengal, Bihar, Jharkhand and Delhi. Parliamentary Constituencies chosen for the Post Polls were representative of the socio-political zones within the state.

Coverage and Sampling: In each Parliamentary Constituency (PC) 3 Assembly Constituencies (ACs) were randomly selected. Thus a total of 990 ACs were sampled across the country. 3224 Polling Booths were randomly selected from these ACs for the purposes of the Post Poll. 

Sample Size: A total of approx. 92,000 interviews are being conducted for the Post Poll.

Interview Method: Voters were randomly selected from the voters' lists of the selected booths
All interviews were conducted in homes using tablet computers, where the respondents chose the candidate and party symbol displayed on the screen, without exposing their choice to the interviewers.

Data Weighting: All data was weighted to reflect the constituency profile on gender and religion/caste.  

OVERALL SUMMARY OF ALL OPINION POLLS

  • Pre Election Polls Sample Size (in 2 phases): 90,339
  • Pre Election update Polls sample Size: 14722
  • Pre Election Telephonic Re-contact Sample Size : 18324 over the duration of polling
  • Post Election Exit Polls Sample Size: 63740
  • Post Election Post Poll: 91712
  • Total Interviews over the polling schedule: 278837

Some Salient Features:

  • For the first time, a large number of female interviewers were used - to get better responses from women - especially among sensitive ethnic / religious communities and rural regions
  • Computer Aided Personal Interviews conducted with GPRS enabled tablets to ensure the strict norms of quality control.
o    All interviews (except the exit poll) were conducted in homes and not at street corners
o    Through better monitoring of data collectors
o    Better quality of questionnaire administration by controlling the order in which questions are asked through a pre-programmed script and not leaving it to the discretion of the interviewers
o    Making data collection process less intrusive by replacing the simulated ballot box (for recording the choice of the voter) with a much less conspicuous hand-held device
o    Respondent invited to show his / her preference for a political party through a simulated ballot paper (tablet screen displaying all symbols and respondent can cast his / her vote by touching the screen)        

.