For one there is no dominant real issue the election is being battled around. As the ruling party, BJP readies itself for a fourth consecutive win in its pet state, the opposition is confident of anti incumbency to change the momentum this time. Confusing reports, sharply divided views and high number of silent voters have led all political analyst and the media to predict a scrape through victory for the ruling party.

The CSDS -CNN – IBN Indian Express and Divya Bhaskar Gujarat elections survey used a combination of  Pre poll, exit poll and post poll exercises to help draw the political landscape of Gujarat and to predict its future.

The findings of Indian Express-CNN-IBN-Divya Bhaskar-CSDS Gujarat Poll are based on a combination of the usual exit poll (an interview as the voter exits the polling station premises) with post-poll survey (interview with the electors at their residence after the day of polling).
This was done to ensure that we can select our sample more carefully and that our investigators can interview the respondents at length and in privacy.
Thus for the first phase of elections, both types of surveys were used. In the post poll survey 1,954 persons on the voters' list were interviewed at 112 locations in 28 assembly constituencies between December 11 and 13.
This was supplemented by a small exit poll of 1,394 respondents in 28 polling booths. The second phase was only an exit poll in which 5,585 voters were interviewed at 128 polling booths in 32 assembly constituencies. Thus in both the phases a total of 8,933 respondents were interviewed in 60 assembly constituencies.
The sampled assembly constituencies were selected randomly. In every sampled assembly constituency, four polling stations were then randomly selected.
For the post-poll survey the respondents were randomly sampled from the electoral rolls of the sampled polling booths. In the exit poll the investigators randomly selected persons coming out of the polling booth.
The interviews were conducted in face-to face situation with the sampled respondents using a structured interview schedule.
The sample selected was fairly representative as it more or less reflected the social diversity of the population and census figures in Gujarat. It included, 45 percent women, (48 per cent in the state), 10 percent Dalits (7 percent in the state), 15 percent Adivasis (15 percent in the state) and 10 per cent Muslims (9 percent in the state).

The fieldwork of the survey was coordinated in Gujarat by Priyavadan M Patel (Vadodara) along with his colleagues, Hemixa Rao (Rajkot) and Bhanu M Parmar (Anand). Sanjay Kumar of Centre for the Study of Developing Societies (CSDS), Delhi directed the survey.
The Central team which designed, coordinated and analyzed the survey comprised Yogendra Yadav, Himanshu Bhattacharya, K.A.Q.A Hilal, Kanchan Malhotra, Praveen Rai, Sanjeer Alam, Dhananjai Joshi, Banasmita Bora, Vikas Gautam, Kinjal Sampat of CSDS and Rajeeva Karandikar of Cranes Software International Limited.

The CNN-IBN-Indian Express-Divya Bhaskar “Gujarat Pre Poll Survey” was designed and analysed by the Centre for the Study of Developing Societies, Delhi. The findings presented are based on a sample of 3893 respondents representing the views, opinions and attitudes of cross section of people in Gujarat. The survey was conducted at 240 locations spread across 60 Assembly constituencies in Gujarat. During the survey 7200 registered voters were approached for interviews out of which 3587 respondents could actually be interviewed.

The sample for the survey was drawn using a multi stage stratified random sampling technique. Out of a total of 180 assembly constituencies in the state a sample of 60 assembly constituencies were randomly drawn using the simple random technique. In every sampled assembly constituency 4 polling stations were randomly selected using the simple random technique. The respondents were then randomly sampled from the electoral rolls of the sampled polling booths.

The fieldwork of the survey was conducted in Gujarat between October 31st and November 6th, 2007. More than 120 investigators and supervisors interviewed the respondents in face-to-face situation at their place of residence using a standard structured questionnaire in Gujarati.