HELP

spyro_2001ie

Hi There.
I need some help with statistics for my thesis. Ive gathered all the data and calculated my means. But im unsure what to do with the data. What type of t-test, what to do with the results. i.e. What dose the number mean to my hypothesis? Im awful at maths.

My thesis is due in 6 days, so i REALLY gotta get this done. Id be willing to pay, if someone could meet up with me for a couple hours to go through it (Dublin city). Im in desperate need at this stage.

cheers

epictetus2009

hi

can you give a description of the experiment you carried out.

then im sure some suggestions can be posted

c

spyro_2001ie

Without getting into too much detail. I had an experimental design with 2 orders of 2 conditions.

Order 1 > No independent Variable present> Independent Variable Present.

Order > Independent Variable Present > No independent Variable present.

I have the data entered into excel, but im unsure about whether to do a 1 tailed or two tailed, paired or un paired t-test.
Then when i get my t value, what do i do with that?
How do i prove or dis prove my hypothesis with this number? P values?

I also have a second hypothesis i need to prove, which involves proving that over time having been exposed to the Independent Variable, resulted in a higher mean effect of the Independent Variable.

Any suggestions would be very welcome.

I could explain it more specifically, but that would involve a 20'000 word post.

2Scoops

spyro_2001ie wrote: »

I have the data entered into excel, but im unsure about whether to do a 1 tailed or two tailed,

This depends on whether you have a directional hypothesis and a very strong reason to completely ignore the possibility that you are wrong. If in doubt, 2-tailed.

spyro_2001ie wrote: »

paired or un paired t-test.

Is it a pre/post design i.e. the same thing (person?) tested on the same endpoint twice? If yes, paired. If no, unpaired.

spyro_2001ie wrote: »

Then when i get my t value, what do i do with that?

In olden days you would compare it to the t-distribution. Modern software does the job for you -> p-value.

spyro_2001ie wrote: »

How do i prove or dis prove my hypothesis with this number? P values?

The p-value will tell you the probability of the difference between means being due to random variation. If it is sufficiently improbable (P<0.05 or other wise), you can say, with a certain degree of probability, that it is 'really' different in a systematic way. What that means depends on your study design. It means you can reject, with a certain degree of confidence, the null hypothesis. You cannot 'prove' the null hypothesis.