This analysis is based on a Post-Poll survey conducted in Punjab by Lokniti, Centre for the Study of Developing Societies (CSDS), Delhi, for ABP News. The survey was conducted from February 9 through February 20, 2017 among 3268 voters in 195 locations spread across 39 assembly constituencies. All constituencies and the polling stations within them were the same as those where Lokniti had conducted the first two rounds of the pre-poll tracker survey in early December 2016 and mid-January 2017. The sampling design adopted was Multi-stage random sampling. The assembly constituencies where the survey was conducted were randomly selected using the probability proportional to size method. Thereafter, five polling stations within each of the sampled constituencies were selected using the systematic random sampling method. Finally, the respondents were also randomly selected from the electoral rolls of the sampled polling stations. Before going to the field for the survey, field investigators were imparted training about the survey method and interviewing techniques at day-long training workshops held in Amritsar and Chandigarh. The field investigators conducted face-to-face interviews of the respondents in Punjabi (and Hindi in some cases) asking them a set of standardized questions. The duration of an interview was about 35-40 minutes. At some locations the non-availability of sampled respondents or difficulty in finding households necessitated replacements or substitutions. It must be noted that the post poll survey conducted by CSDS is very different from an exit poll in which voters are approached outside the polling booth on voting day. Instead voters were approached by the field investigators for an interview at their place of residence after votes had been cast in their respective constituencies, but before the results were known.
Profile of achieved sample
|Survey Sample (%)||Census 2011 (%)|
|Scheduled Caste (18+ years)||35.5||29.8|
|Women (18+ years)||40.8||48.3|
|Urban (18+ years)||29.5||38.3|