PakMediNet Discussion Forum : Biostatistics : help me plzzzzz
help me plzzzzz
I am stuck up to find out the p-values for freq of signs in one of my article. I don't know how to apply Chi square or whatever test applicable.
My topic is hyperglyceamia in stroke
OBJECTIVE
‘The objectives of this study are to;’
1. Determine frequency hyperglycemia in non-diabetic patients after acute ischemic and hemorrhagic stroke.
2. The association of hyperglycemia with severity of ischemic and hemorrhagic stroke in non-diabetic patients.
Stroke severity
The level of consciousness will graded as,
1. Alert [GCS 15/15]
2. Disoriented or somnolent [GCS 4-14/15]
3. Unconscious, weather or not localizing signs [GCS 3-4/15]
The severity of hemiparesis will be graded as
1. No paresis,
2. Mild to moderate paresis [decreased power but able to lift the limb against gravity]
3. Severe paresis [unable to lift the limb against gravity, total paralysis].6
Hyperglycemia
The flowing cut-off values for blood sugar will be used 5
1. 60-129 mg/dL = Euglycemia on admission
2. > 130 mg/dL = hyperglycemia
HYPOTHESIS
Null hypothesis: There is no association of hyperglycemia after the attack of acute ischemic and hemorrhagic stroke.
how i apply the the t-test
i measured the blood sugar and assessed the severty of stroke for three days
[Edited by DRKHUHRO on 29-09-2007 at 06:52 PM GMT]
Posted by: DRKHUHRO Posts: 4 :: 28-09-2007 :: | Reply to this Message
Re: help me plzzzzz...........
Posted by: drhashim Posts: 2 :: 28-09-2007 :: | Reply to this Message
Re: help me plzzzzz
Your data isnt clear in your description. However, my question is why you need to apply test and then p-value for your data? If you want to calculate the association between different variables, then you need a test to see if the association is significant or not. The selection of a statistical test is based on different ways and type of variables is one of them.
I suggest to make 2 groups, one having normoglycemia and other having hyperglycemia. You can then apply Chi-Square to study the comparison between these two groups. Result will be a p-value.
I hope this would he helpful.
Posted by: docosama Posts: 333 :: 29-09-2007 :: | Reply to this Message
Re: help me plzzzzzWhat is your research question?
Posted by: asiddiqui Posts: 26 :: 29-09-2007 :: | Reply to this Message
Re: Re: help me plzzzzz
thanx dear
i m here submited the little introduction of my synopsis. please tell me about the t-test how can i apply this
OBJECTIVE
‘The objectives of this study are to;’
1. Determine frequency hyperglycemia in non-diabetic patients after acute ischemic and hemorrhagic stroke.
2. The association of hyperglycemia with severity of ischemic and hemorrhagic stroke in non-diabetic patients.
OPERATIONAL DEFINITION
Stroke severity
The level of consciousness will be assessed by Glasgow Coma Scale (GCS) [attached annexure B] it is widely used and recognized measure of neurological impairment, 5 and graded as,
1. Alert [GCS 15/15]
2. Disoriented or somnolent [GCS 4-14/15]
3. Unconscious, weather or not localizing signs [GCS 3-4/15]
The severity of hemiparesis will be graded as
1. No paresis,
2. Mild to moderate paresis [decreased power but able to lift the limb against gravity]
3. Severe paresis [unable to lift the limb against gravity, total paralysis].6
Hyperglycemia
The flowing cut-off values for blood sugar will be used 5
1. 60-129 mg/dL = Euglycemia on admission
2. > 130 mg/dL = hyperglycemia
HYPOTHESIS
Null hypothesis: There is no association of hyperglycemia after the attack of acute ischemic and hemorrhagic stroke.
please send me ur email i will send u the spss data
[Edited by DRKHUHRO on 29-09-2007 at 07:29 PM GMT]
Posted by: DRKHUHRO Posts: 4 :: 29-09-2007 :: | Reply to this Message
Re: Re: Re: help me plzzzzzYour hypothesis (or null hypothesis) is still not clear. What I understand so far (which may be incorrect) is that you have two groups of patients, ischemic stroke and hemorrhagic stroke. You want to see whether there is difference in the proportion of patients with hyperglycemia in each group. If this is true, then you can’t use t-test unless you stop dichotomizing blood glucose levels into euglycemia and hyperglycemia. I am certain that your sample size is small; therefore, my advice will be to use blood glucose levels as a continuous variable, and then do a t-test. The reason is that you loose information by dichotomizing a continuous variable. If you insist on dichotomizing blood glucose levels then your can use chi-square test or Fischer’s Exact test. Having said that, as this is an observational study, you should adjust for confounding variables by either multivariate linear or logistic regression analysis or with propensity scores. You should get help from a biostatistician and should not simply use SPSS to get answers.
Posted by: rqayyum Posts: 199 :: 13-10-2007 :: | Reply to this Message