The objective of the BT-Gallup survey is to identify the hottest states for business, based on both perceptual and factual information. Unlike our previous two surveys (the first in 1997 and the second in 1999), this year's ranks are based on two components: A perceptual score obtained from the opinion polls and a factual score compiled from various data sources. As this is the first time that a factual dimension has been added to the rankings, we have given an ad-hoc 70:30 weightage to perceptual and factual scores to arrive at the final rankings.

The perceptual scores were derived through a four-step process. One, from the list of states, the respondents mentioned the states they were very or somewhat familiar with. Two, they were then exposed to a battery of parameters from which they were to select eight that they thought were the most important. Three, the eight parameters had to be ranked in order of their importance. The most important parameter was given the first rank, the second most important the second rank, and so on. Finally, for each of the eight parameters selected, the respondent had to mention the best state/ other good states, and worst state/other bad states.

To arrive at the factual score, a large number of metrics were drawn up along four dimensions: Government support, physical infrastructure, social infrastructure, and labour. Each of the four dimensions were further broken up into parameters, all of which were accorded equal weights.

The Respondents

The target respondents were two: CEOs and policymakers. The CEOs or industrialists were identified from BT 500 and, as in the past, the questionnaires were mailed or faxed to them. While selecting the respondents, care was taken to ensure representation across industries. Policymakers were a new category of respondents introduced in the survey this year, and they included senior bureaucrats involved in urban planning activities. Only officials in the ranks of Principal Secretary, Joint Secretary, Deputy Secretary, or Collectors in the secretariat were interviewed.

The survey was conducted in 13 cities across the country (See The Respondent Universe). For the opinion poll, a structured quantitative research approach was followed to arrive at the rankings of the states. The questionnaire used in the last round was the starting point, but it was fine-tuned and some questions were added or deleted based on their relevance.

The Analysis

Once the important parameters (as provided by the respondents) had been identified, a nett score for every state under each of the parameters was derived. The nett score was computed as follows:

Nett score = [percentage saying a parameter is important] x [(number of best state mentions x 2 + number of second best state mentions x 1 + number of worst state mentions x -2 + number of second worst state mentions x -1)]

The nett scores for each state on all parameters were added to arrive at an interim score for that state. This was done for each category of respondents separately. The interim scores were used to rank the states within each respondent category. The interim scores from the CEO segment and the policymaker segment were combined to derive a composite score for each state on an overall basis. A 60:40 weightage was assigned to CEOs and policymakers, respectively. Based on the total composite scores, we arrived at final ranks for each state at an overall level.

To calculate the factual score for each state, data was classified into three categories: High, moderate, and low, with scores of three, two, and one, respectively. All the parameters were assigned equal weightage. A composite score was calculated for each state by summing the scores across all the factual parameters. States were ranked based on the composite scores. A weight of 70:30 was assigned to perceptual and factual data respectively. These weights were applied to the respective composite scores and the final ranks were determined.