National

The Outlook-CMS Opinion Poll

Full text of the Opinion Poll for the assembly elections

Advertisement

The Outlook-CMS Opinion Poll
info_icon

The "3S" METHODOLOGY

The CMS Methodology for pre-poll surveys has been developed based on theexperience of three decades in conducting opinion studies, voting analysis andunderstanding of the political process in different parts of the country. The CMS model isbase on SSS-Size of sample, Spread of sample within a constituency and Selectionprocedures of sample respondents. One of the unique features of this model is thetransferability of votes within an alliance, and more importantly, shift in pattern ofvoting intentions.

The field work for the survey was conducted between April 19-29 April. In all about 76Assembly constituencies spread across the four states were covered.

Advertisement

In each Assembly constituency, three villages and one urban centre were selected. Ineach Assembly segment a total of 100 voters were interviewed with the help of asemi-structured interview schedule following a stratified sampling procedure.

Simultaneously, care was taken to see that opinions were gathered from a cross-sectionof voter population representing various caste/community groups. In all about 7,859carefully selected voters were personally interviewed. In order to sample 7,859 voters wehad to contact atleast 8700 voters.

In each household only one voter was interviewed. Further, an equal proportion of maleand female voters was maintained. This survey draws a sample which closely represents thecharacter of each Assembly constituency in respect of their urban/rural population, socialgroups and gender profile.

Advertisement

CMS has successfully demonstrated the relevance and usefulness of such methodologyframework through various successful poll forecasts.

All sample surveys, however, scientific are subjected to sampling as well asnon-sampling error. In view of the possibility that sample distribution as such may nottruly reflect the voter distribution, care has been taken to adjust the voting data basedon the past voting trends in the sample as well as the population. This procedure improvesthe over all accuracy of the estimate. The actual voting percentage at the state level ofthe "Outlook-CMS" survey are expected to be within a margin of 2 per cent of theestimate of major parties/alliances.

Tags

Advertisement