Hi! Just curious, is my question okay for this site? Either way, could you all point me in the correct direction for finding an appropriate statistical test to use? stats.stackexchange.com/questions/241828/….
@Benzene I just came here to complain about your title, actually.
It hits one of my pet peeves ... funny you should wander in right behind me
The issue is that to talk about significance, you need a specific hypothesis and a test statistic. Data are just data, they don't have the property of "significance" on their own. Any number of things might be tested in any number of ways.
phrases like "test the confidence interval of our results" are so vague as to convey almost nothing to me. (To be honest, it doesn't sound like a hypothesis test at all.)
First step would be to take a step back from notions of significance or confidence intervals (or indeed jargon more generally) and talk about what you're trying to find out.
That will help you to explain the context, the ideas and the problem more.
Your final sentence should be its own question with different tags
Your question is essentially on topic here but as it stands, your question is so unclear as to be pretty much unanswerable (and may well risk straying into 'too broad' as well, but first let's try for greater clarity)
It will likely be put on hold. This should not be taken as discouraging; it's an invitation to improve your question without getting premature answers (which are frequently unhelpful)
It may take several iterations to get a good question.
@Glen_b I added more information. Does this look better? I am basically trying to determine if giving a random writer (Markov chain) more context (a higher order) results in its output more closely matching the tone of its input.
@Benzene "trying to determine if giving a random writer (Markov chain) more context (a higher order) results in its output more closely matching the tone of its input." -- that's a highly informative sentence.
It looks better than it did. I think there's more improvement to be made (I'd avoid the word correlation for example and I'd try to make it shorter if possible) but I'm going to have to leave that to others, sorry.
