So if a group within a certain species produces a feature that is detrimental to its survival, and NS takes place and this group is killed off, then you're telling me that is not a textbook example of reducing information in a species? More features = more genetic information. Natural selection doesn't add, it only subtracts.
How the heck can you say it doesn't effectively reduce information?
The main problem here stems from your use of the word "information." It's too vague.
That said, natural selection affects the ratios of traits in a population. It's just as accurate to say it increases the frequency of some as it is to say that it decreases the frequency of others. Natural selection can take a rare mutation and fix it in, say, 30% of the population because that mutation is beneficial to the individual to an extent. That would be a population diversity enhancer. Natural selection in that case would be preserving more varieties of traits than simple chance would likely result in. Natural selection isn't just about relatively detrimental traits decreasing. It's about relatively beneficial traits increasing. And traits aren't as simple as good or bad, gone or total dominance. (Look up sickle cell and malaria for the classic putative example of this). What is accurate to say is that natural selection can only operate on traits as they already exist in the population. It doesn't add new traits in a population, but it can transform how they are represented in that popualtion. Again, mutation is the source novelty.