with an election approaching. We tend to take most polls with "a grain of salt'' even though there are a great many people who put a good deal of trust in them.

We do not think either of the main parties should base their futures on recently published polls especially since the samplings are not wide.

Depending on how polls are conducted, the size of the sampling, the way and the time of day the questions are asked, they can say what people want to hear.

The major political parties do their own polls for their own campaign purposes and in the case of the Penn and Schoen polls done for the United Bermuda Party these are widely based, carefully weighted and have in past elections been quite accurate. However, neither party makes their results public.

In other countries the polls conducted by a variety of organisations generally produce quite similar results. In the case of the two most recent independent polls in Bermuda some of the results are quite startlingly different. In fact, so different as to raise serious questions about the differences.

The local Research Innovations' survey showed that Premier Pamela Gordon's favourability among black Bermudians had dropped a startling 21 points in a month from 46 percent in August to 25 percent in September. We find that almost impossible to believe, especially given the fact that the PLP leader has been almost invisible and unavailable for months.

The Bermuda Omnibus Survey indicates that 48 percent of the population prefer Premier Pamela Gordon as Premier to 24 percent who would prefer Jennifer Smith as Premier. The survey also showed that 68 percent were completely satisfied or generally satisfied with Premier Gordon's government. In most countries that figure alone would be enough to return the UBP to power. But then in any overseas presidential race, a candidate would be "over the moon'' at the Research Innovations finding of an overall 80 percent favourable opinion of Opposition Leader Jennifer Smith.

The Omnibus Survey showed decided voters in favour of the UBP increasing from 52 percent in June to 58 percent at the end of September, while the PLP was down from 36 percent in June to 33 percent in September. A shift is not surprising given intense activity by the UBP but the extent of the shift is.

Research Innovations said that 42 percent of those polled would vote for the PLP and 35 percent for the UBP. Even hard core party supporters will find the differences in the two polls very hard to accept, always remembering that there is a difference between "decided voters'' and "those polled''.

The campaign has begun and figures do shift as campaigns progress, but we find the 58 percent UBP figure very high and the 33 percent PLP figure very low.

While opinion is not scientific, our informed reading of the present political climate is that the two parties are very much closer together.

It is important to remember that in both polls, the figures for those who don't know, are undecided or refused to answer are high and that the result of the election will hang on how they decide to vote, if they vote.