In an article written in Frontline, prior to the polls, YY and P.M.Patel wrote:
In response to a direct question, as many as 19 per cent of all respondents say Godhra and post-Godhra developments influence their decision about voting. This figure alone is not enough to conclude that the swing-back towards the BJP is due to communal violence. A detailed analysis of the exact nature of the impact of the recent communal violence on voting decisions does not lead to clear conclusions, for many of these voters could not specify the exact nature of this effect. It would be too hasty and perhaps unjust to think that all these ordinary voters approved of the massacre of Muslims and jumped on to the BJP bandwagon. But it is clear that the BJP has succeeded to keep the Godhra incident as an unrelieved experience on an unresolved tension for a large number of Hindus and use it to its advantage. For many respondents Godhra and post-Godhra violence is something that can overpower all other considerations.
They further wrote:
On the face of it, an average Gujarati does not approve of the post-Godhra carnage ("ramkhan" or "hullad"). When asked directly to chose between three possible responses to this violence, two involving degrees of approval and one expressing complete disapproval, an overwhelming majority of 72 per cent said it was "completely unjustified". This is as true of the Hindu respondents as of those who intend voting for the BJP. But this consensus breaks down as soon as we pick up the role of the State during this violence that invited widespread censure and condemnation. On balance more people disapprove the State government's handling of the riots than approve of it, but this is a contentious arena where partisan differences begin to play a role. Similarly, the government's cussed handling of the post-riot relief work also invites popular censure. A majority of the respondents, including a majority of Hindus, believe the government did not do enough or did nothing..
The survey attempted to probe deeper than these direct questions could and went into the popular psyche on the communal issue. A series of statements were read out to the respondents and they were asked if they agreed with them. The answers do not reveal as communally polarised a situation as is sometimes believed or projected. But it does bring out deep prejudices, social distance and hostility. On balance, most of the soft communal statements won approval from a majority of Hindu respondents. A majority agreed that one cannot have the same empathy for followers of other religions as one has for one's own, and that democracy means the rule of the majority community, that religious conversions and inter-religious marriages should be banned by law. The majority of Hindu respondents are not directly opposed to democracy but hold a strongly majoritarian understanding of what democracy should mean. But all the responses do not fall in this line. A general statement about all religions being the same also gets approval, while the demolition of Babri Masjid invites more disapproval than approval. Besides, it should be noted that the approval for communal statements is seriously contested by a large number of Gujarati Hindus.
It would be hard to provide clear survey evidence that the BJP's apparent comeback is directly related to the anti-Muslim carnage and the communal polarisation. But the survey does provide some clear pointers. If all the communalism-related questions are put together to form a measure of intensity of communal feelings, we see a direct association between this measure and the vote for the BJP
In an article in Frontline, explaining the post-poll findings in 2002, YY wrote:
Violence often serves to redraw boundaries of identity and affiliation. This is what seems to have happened with Gujarat. An analysis of the 65 constituencies that saw significant anti-Muslim violence in early-2002 brings out the dark shadow on this verdict of the widespread massacres.
Although the overall number of seats held by the BJP and the Congress has remained about the same, as many as 76 seats have changed hands between this and the previous round of Assembly elections. The BJP has lost 29 seats to the Congress, but has more than made up for it by snatching 35 seats. But it is important to note that 22 per cent of all voters mentioned the Godhra carnage or the post-Godhra violence as the decisive consideration. A quarter of the BJP's voters and one-sixth of the Congress' voters mentioned either of these considerations. Other secular considerations weighed more heavily for the Congress voters than those of the BJP, but there were no sharp divergences here.
What followed was a contradiction of YY’s initial finding that the violence-affected constituencies tended to return the BJP candidates. YY asked:
Did the riots play a role in swinging voters back to the BJP even in areas that did not see any riots? The answer, according to the CSDS post-poll survey, is a clear `yes'. The proportion of Hindu respondents who say that Godhra mattered a `great deal' in their voting decision is about the same in the riot-affected regions as in the rest of the State. If anything, respondents in the riot-affected areas were more circumspect in discussing this and keener than the others to deny any connection with Godhra. The Sangh Parivar has succeeded in turning Godhra from a local event to a generalised icon; the proximity of Godhra to a voter has had nothing to do with the geographical distance.
When asked to agree or disagree with a series of statements concerning Godhra and its aftermath, a majority of Hindu respondents endorsed the post-Godhra riots while trying to distance themselves from their consequences. A clear majority of 55 per cent of Hindu respondents (73 per cent of those who had any opinion on the matter) agreed with the suggestion that the post-Godhra riots were "necessary to teach a lesson to anti-national elements" (read Muslims). While this category includes a larger proportion of BJP voters, it is worth noting that 47 per cent of the Congress voters (69 per cent of Congress voters who have any opinion on this question) also agreed with this communal statement. At the same time, it is also true that an ordinary Hindu does not want to look back now. Two-thirds of them agree that both Hindus and Muslims should now forget Godhra and its aftermath. Oddly for a people who are prepared to endorse the massacre of Muslims, nearly everyone agrees that those found guilty of violence in the riots must be punished.
On a simple reading these are contradictory answers. But these contradictions bring out the state of the mind of an average Gujarati Hindu. A Congress supporter could use these as the justification of the soft-Hindutva line that was deliberately adopted by the Gujarat Congress during the election campaign.
Does the truth lie behind these contradictory findings? I doubt.
(Those interested in knowing the methodology adopted by CSDS in these surveys may read this report which appeared in The Hindu in 2003)