Do you think that the text analyzer (and, e.g., Gunning Fog readability index) provides true information on a readability of a text? I have tried to analyze some of my texts and am not sure about this, really...
Readability is a relative concept, using maths to try to determine difficulty, so it fails to take into account things like the schema of the individual, but it can provide an indication at a certain level. the results can be skewed simply by the length of the text- a short sentence will often have a lexical density of 100%, though it could be very simple. Many teachers say that they can use the levels as a basic way of checking whether text is appropriate for learners at certain levels. They might exclude texts wrongly, but tend to include ones that are OK. It's far from an exact science. In our case, it's something we're working on in the background and hope to produce something better over time, but time is always the problem- there are only twenty-four hours and ten fingers. BTW- any suggestions or comments, please send them to me by PM or email as we are always trying to improve things.