Sign test help? Urgent!

Generally, you run down the columns until you hit your degrees of freedom, go along the row until you hit the probability that meets your critical values (typically 5% for one-tail and 2.5 for two). Then you take that figure and compare it to your test statistic. If your test stat is more extreme, then it falls within the critical region and is therefore significant to that level (p > 0.05).

Im really sorry, that kind of went over my head..ive worked out the difference between the two sets of data, and now ive ordered the data from smallest to largest...now im completely stuck on what to do next?
On my fomula table it says number of pairs of score ect. and then numbers? This is the bit i dont understand. How to apply my data to the formula. Thanks for the help as well btw i really appriciate it.

Hmm. I thought you were referring to distribution tables. Perhaps not.

What sign test are you using?

Hmm. I thought you were referring to distribution tables. Perhaps not.

What sign test are you using?

I think ive suddenly got it, i know you cant see any of my data but would you tell me if this sounds about right....it says number of scores, and ive got 19 pieces of data....so then across the table to 19 pieces of data it says 0-4 under the table headed significant at 5% level Accept hypothesis...so does this mean that as long as all of my data is between 0 and 4 then it supports the hypothesis?

Degrees of Freedom = N-1, so you should be going to 18.

Then you look at either 5% or 2.5% depending on if your hypothesis is directional (I.e. "x will be greater/less than.." or "x will differ from" ) .

Look at that corresponding number. Lets say hypothetically it is 2.3, and you want to see if your test statistic (the number you got at the end of your equation) if more extreme than that.

Compare those two figures. If your test stat is more extreme, (i.e. >2.3 in this case) then you can say your data is significant to that level (5% being the standard for a one-tail hypothesis). Thus, you can reject the null hypothesis in favour for the alternative.

It's hard to explain without a data set to work with. What test are you using? Wilcoxon Signed Ranks?

Do you want me data? The differences or the whole lot?

Tbh, at 12.30am, the last thing I want is your data.

Lets go from the start.

What is your hypothesis?
What is the design?
What test do you want to use?