UserenameNot
Badges: 10
Rep:
?
#1
Report Thread starter 1 year ago
#1
So for my NEA (Geog) I did the Mann Whitney U for one of my hypothesis. My Literature review for the hypothesis supported my hypothesis, and my test supports it too. However I do not think that it is right.. I haven't mentioned the hypothesis just in case of plagiarism or something else. Also in my Literature Review it said that although mostly my hypothesis is right, not all tree (my topic is carbon) behave the same so I have decided to use this to say that i will be going against my hypothesis and rejecting my hypothesis. Is this okay to do, or should i just accept my hypothesis?
0
reply
kelefi
Badges: 19
Rep:
?
#2
Report 1 year ago
#2
In these cases you could consider saying something along the lines of 'although the statistical test suggests that there is not sufficient evidence against the hypotheses (whatever that may be), due to certain factors such as blah blah blah a repetition of the experiment should be considered with closer monitoring of confounding and lurking variables which could have influenced the test'

EDIT: I forgot to mention, no statistical test is absolute!!! All tests (no matter how complicated) are subject to assumptions and approximations. Just because a test tells you one thing doesn't mean that it's true!
1
reply
Trinculo
Badges: 21
Rep:
?
#3
Report 1 year ago
#3
It depends if you're doing science or social science type geography.

If it's social science, like human geography - well, you can dial back any research because it's by its very nature not necessarily replicable.

However, if it's sciencey geography, like counting rocks or rivers or something, then you're in dicey territory. As supervisor, I would be forced to ask - "if you were going to reject the results of your test, why did you do it in the first place?" Why do hypothesis testing if you have already decided what you're going to find?

Take the two examples:

1. Public transport is not busier on weekday lunchtimes than Saturday evenings in 2018.

2. Average temperatures in Spanish cities are not correlated to population density

You can play around with H1 because it's about human behaviour experiments and you can state all sorts of limitations - however, you would have to answer why you chose the hypothesis and variables if you thought them to be flawed. You could repeat the experiment.

H2 is about measurements and records. Why did you choose this hypothesis in the first place, if you were going to reject a test result that you don't like. If you repeat your test 100 times, you'll get the same result.
1
reply
UserenameNot
Badges: 10
Rep:
?
#4
Report Thread starter 1 year ago
#4
(Original post by kelefi)
In these cases you could consider saying something along the lines of 'although the statistical test suggests that there is not sufficient evidence against the hypotheses (whatever that may be), due to certain factors such as blah blah blah a repetition of the experiment should be considered with closer monitoring of confounding and lurking variables which could have influenced the test'

EDIT: I forgot to mention, no statistical test is absolute!!! All tests (no matter how complicated) are subject to assumptions and approximations. Just because a test tells you one thing doesn't mean that it's true!
Hi!

Thanks for the reply! Yeah that's basically what ive written but like the other reply below I just was'nt sure of what to do.. thanks anyways!
0
reply
UserenameNot
Badges: 10
Rep:
?
#5
Report Thread starter 1 year ago
#5
(Original post by Trinculo)
It depends if you're doing science or social science type geography.

If it's social science, like human geography - well, you can dial back any research because it's by its very nature not necessarily replicable.

However, if it's sciencey geography, like counting rocks or rivers or something, then you're in dicey territory. As supervisor, I would be forced to ask - "if you were going to reject the results of your test, why did you do it in the first place?" Why do hypothesis testing if you have already decided what you're going to find?

Take the two examples:

1. Public transport is not busier on weekday lunchtimes than Saturday evenings in 2018.

2. Average temperatures in Spanish cities are not correlated to population density

You can play around with H1 because it's about human behaviour experiments and you can state all sorts of limitations - however, you would have to answer why you chose the hypothesis and variables if you thought them to be flawed. You could repeat the experiment.

H2 is about measurements and records. Why did you choose this hypothesis in the first place, if you were going to reject a test result that you don't like. If you repeat your test 100 times, you'll get the same result.
Hi! Thanks for the reply!.. My NEA is on a science Geography topic so I guess I might be in the wrong if i reject my hypothesis... But my other data presentation shows that my hypothesis is wrong.. only the test suggests that my hypothesis is right! However the day i collected the results was on a rainy day and this could effect my resulsts it was on soil and its pH etc.. So would it be okay to justify my choice of rejecting the hypotehsis using this or should i just scrap it all togther and not use the mann whitney u? cause its fine if i do.
0
reply
Trinculo
Badges: 21
Rep:
?
#6
Report 1 year ago
#6
(Original post by UserenameNot)
Hi! Thanks for the reply!.. My NEA is on a science Geography topic so I guess I might be in the wrong if i reject my hypothesis... But my other data presentation shows that my hypothesis is wrong.. only the test suggests that my hypothesis is right! However the day i collected the results was on a rainy day and this could effect my resulsts it was on soil and its pH etc.. So would it be okay to justify my choice of rejecting the hypotehsis using this or should i just scrap it all togther and not use the mann whitney u? cause its fine if i do.
If you think about it, that's not really how science is supposed to work. You're supposed to carry out an experiment with confidence in your methods and then accept or reject your hypothesis on the basis of those results - not on the basis of the opposite of the results.

If you want to repeat the experiment, that's fine, but you shouldn't really change anything about it, and you'd have to do a write up justifying it.

Really - if the experiment suggests a certain outcome, you should accept the outcome on the basis of the data given the circumstances - I don't know what your experiment or thesis are, but take the following example:

H0: The number of cars passing through a roundabout is not positively correlated to rainfall.

In the experiment, you count the number of cars going through on sunny days and rainy days.

Let's say you do the experiment and get the result that car use increases on sunny days. This isn't what you thought. Should you present your findings along with the limitations - or should you keep doing the experiment until you get the results you like?
0
reply
UserenameNot
Badges: 10
Rep:
?
#7
Report Thread starter 1 year ago
#7
(Original post by Trinculo)
If you think about it, that's not really how science is supposed to work. You're supposed to carry out an experiment with confidence in your methods and then accept or reject your hypothesis on the basis of those results - not on the basis of the opposite of the results.

If you want to repeat the experiment, that's fine, but you shouldn't really change anything about it, and you'd have to do a write up justifying it.

Really - if the experiment suggests a certain outcome, you should accept the outcome on the basis of the data given the circumstances - I don't know what your experiment or thesis are, but take the following example:

H0: The number of cars passing through a roundabout is not positively correlated to rainfall.

In the experiment, you count the number of cars going through on sunny days and rainy days.

Let's say you do the experiment and get the result that car use increases on sunny days. This isn't what you thought. Should you present your findings along with the limitations - or should you keep doing the experiment until you get the results you like?
With your example, I should present the findings with the limitations. But for mine it is only the statistical test that says that i should accept my hypothesis all of my other data says that i sould reject my hypothesis so this is where im stuck. I think that i should reject my hypothesis on the basis of my other data, and say that i am not sure about the validity of the test. Is that a good idea?
0
reply
Trinculo
Badges: 21
Rep:
?
#8
Report 1 year ago
#8
(Original post by UserenameNot)
With your example, I should present the findings with the limitations. But for mine it is only the statistical test that says that i should accept my hypothesis all of my other data says that i sould reject my hypothesis so this is where im stuck. I think that i should reject my hypothesis on the basis of my other data, and say that i am not sure about the validity of the test. Is that a good idea?
Problem is I don't really know what your research is. If you have multiple sources, rather than just the test, then you kind of have to choose between your evidence - is the other evidence data-driven?
0
reply
X

Quick Reply

Attached files
Write a reply...
Reply
new posts
Back
to top
Latest
My Feed

See more of what you like on
The Student Room

You can personalise what you see on TSR. Tell us a little about yourself to get started.

Personalise

University open days

  • University of East Anglia
    PGCE Open day Postgraduate
    Tue, 3 Mar '20
  • University of Bradford
    Postgraduate Open day/Evening Postgraduate
    Tue, 3 Mar '20
  • Queen's University Belfast
    Postgraduate LIVE Masters & PhD Study Fair Postgraduate
    Wed, 4 Mar '20

Do you get study leave?

Yes- I like it (498)
59.71%
Yes- I don't like it (43)
5.16%
No- I want it (237)
28.42%
No- I don't want it (56)
6.71%

Watched Threads

View All
Latest
My Feed

See more of what you like on
The Student Room

You can personalise what you see on TSR. Tell us a little about yourself to get started.

Personalise