Hi Lewis,

We thought long and hard about how best to publish the survey's results. Our original plan, in fact, was to simply publish the survey data in its entirety. The problem is that the data has to be analyzed in some detail to draw some of the conclusions we highlighted. That adds complexity and the potential for misunderstanding, so we elected to simply summarize our analysis.

Let me give you an example: To determine the dominant reasons for selecting a specific database, it is necessary to exclude from that measurement those respondents who had implemented multiple databases. If we had not done so, then the results would not have been valid.

There are several examples like this, so we thought it would be simpler to just summarize the results of our analysis (which was not insubstantial).

To be clear, this was not a scientific, peer-reviewed study. But it was also not an empty marketing exercise. We think some interesting data came out of the survey, and we hope the community agrees.
Andy Astor | Homepage | 07.24.08 - 10:36 am | #

Hi Andy,

Thanks for the update. Would it be possible for the unanalyzed data, in aggregate, to be released (spreadsheet maybe)?

Say something like:

Question 1: 500 respondents: 100 A, 150 B, 75 C, 0 D, 175 E
Question 2: etc

Numbers over the total # of respondents would show that some people chose multiple answers, which in itself is valuable info.

That would allow people to analyze the data and draw their own conclusions.

Still wouldn't make it scientific but it would make it more useful.

Thanks,

LewisC
LewisC | Homepage | 07.24.08 - 11:00 am | #