You have to understand that I do not know what the coursework is, and that with these programs you can do about a hundred things that all involve testing for some level of significance. I could not access the Google drive link.
Though from what you have said so far. This is what I think is going on. You have some test done in a laboratory setting. There are 2 groups. One is control and the other is treatment. Ideally, treatment and control are randomly selected. The treatment group maybe got a certain medicine or something else is done to them. The hypothesis is that the treatment has an effect on one or more variables. To learn whether the treatment had any effect, you want to test whether some dependent variable is significantly different in treatment compared to control.
Generally this is done as a superior alternative to a simple before and after study. The main idea behind it is that things change over time, and you would not be able to say whether any change comes from the treatment or not without a control group. With a control group you have what is called the counterfactual. What matters here is not just that things changed over time, but that it changed differently from the group who did not get the treatment. The underlying idea is that both control and treatment would have changed similarly had the treatment not been there, thus giving you sort of a realtime benchmark to test your results.
If this is the case it is actually quite simple. What you do with statistics is just a refinement of saying "it is bigger" or "it is smaller". Instead, you can say, "it is bigger, and I am 95% sure that is not due to coincidence". That is what significance is (in this case, you call that 5% significance). To do this you make the crucial assumption that populations are normally distributed. This means that most people are about average , and less people are exceptionally low or high. If you know the distribution, you can make the calculation.
Is this correct? This is the situation you are looking at?
