Information Quality and Intentionality

My friend Kelly had some questions for me after my proposal last month and I’m finally getting around to thinking about some of them. One question that she had was about how a lot of low quality information (e.g. press releases, advertisements etc.) is not accidentally of low quality, but is rather intentionally biased to get a particular side across. Should a measure of information quality address the intentionality of the communication? Is it worse if something is misleading than mistaken?

Whether something is misleading or mistaken has to do with the intentionality of the communicator, however what one perceives in the end is still the same: lower quality information. I think it would be difficult to show that someone had the intention to mislead because that information is shared only by the creator of the information, or at best the institution. Based on just the end product there’s no way to know the intentionality. If we could tell if a communicator was intentionally misleading, we would be able to factor this in to their reputation score. However, there are some cues that can raise suspicions about intentionality such as the relationship of the communicator to advertisers, the political leanings of the communicator, and the funding source for the production of the information. But these aren’t smoking guns; just because a cell phone maker pays for a study on the dangers of cell phone use, it doesn’t necessarily mean that the results are biased. But it does give us pause to think about the intentions of the producer of the information.

So back to the original question: should information quality address intentional bias? Yes, I think it should, but since the true intentions of a communicator are hidden, we have to rely on the cues that I listed above. The more intentionally biased a source is, the more this should in turn affect their credibility rating; in fact this could be thought of as another facet of source annotation for the system that I’m building.

Comments are closed.