>>11967837Alrighty bro you're lucky I'm feeling sympathetic today,
Here's the deal. In an intro-level stats course, the test stat you're trying to find with your paired data is almost definitely a t-statistic for mean difference.
What that means is you've probably got in front of you a list of paired numbers (paired meaning there's a 1-to-1 relationship between the numbers, like height of father and height of son pairing each father with his son, or test score before and after class pairing scores before score with after score for the same student). This looks like two samples, but you're gonna treat it like one sample because you only care about the difference between each number and it's paired number (i.e. if a student got a score of 40 at first, then a score of 50 later, you only care that the difference is 10). So first step is to convert the two sets of data into one set of differences.If you're using a TI-84, put the first data in L1, second in L2, then put L3 =L2-L1. If you're using Excel or something similar, do the equivalent. Now you have a single list of numbers and you're probably trying to figure out what the mean is. For doing tests about a mean in this sort of situation, you use a t-statistic, which just takes that big list and gets all the important information from it into one number. For simplicity, we'll assume that you're trying to test if the mean difference is nonzero. For that, you'll calculate the t-statistic however is easiest, t-test in TI-84, then find the probability that a t-distributed variable would take a value that extreme or more given your data has mean zero (t-test in TI-84 does this part for you). Compare the probability with your significance level. That will be your answer.
