To Device Three Research Questions that Could be Examined, and Tested by Applying the Techniques Chosen from the Data Which is Provided by University Statistics Department

Background and Objectives: The questionnaire for the healthy life survey carried out in Berkshire, Buckinghamshire, Northampton shire and Oxford shire health authorities this survey was designed to study the adult population in 4 health authorities. The data is provided from this study in the form of SPSS file. I have to device three research questions or theories from the data provided and these questions should be examined and tested by applying the three statistical techniques a parametric test a non-parametric test and a kind of confidence interval.


Introduction
The questionnaire for the healthy life survey carried out in Berkshire, Buckinghamshire, North amptonshire and Oxford shire health authorities this survey was designed to study the adult population in 4 health authorities. The data is provided from this study in the form of SPSS file. I have to device three research questions or theories from the data provided and these questions should be examined and tested by applying the three statistical techniques a parametric test a non-parametric test and a kind of confidence interval. The Two sample T Test is used for the parametric test the Chi Square Test is used for the non-parametric test to find out the association between the variables.

The Chosen Research Questions are as under
a) Diabetes mellitus in (sex) males and females its interpretation with Chi Square Test. b) Comparison of weight of (sex) males and females with Two Sample T Test.

c)
To find out the confidence interval with BMI and (sex) males and females. Further we used exploratory analysis to get more information  A. Null Hypothesis (assumed to be true).

Diabetes Mellitus in Males and
There is no any association between gender (sex=male & female) and diabetes mellitus.

B. Alternative Hypothesis
There is association between gender (sex=male & female) and diabetes mellitus.
I think the chi squared test is appropriate for this issue it is done by using descriptive analysis using cross tabs and by chi square test.  (Table 4) The test statistics shows that number of valid cases are 2100 and the ratio is 6.516 and the linear-by-linear association is 6.477, Pearson Chi-square is 6.480 and continuity correction is 5.508. The degree of freedom is =1 overall. P value = < 0.011 overall and .019 for continuity correction.
As p value = <0.05 so it rejects the null hypothesis.
Therefore, it can be assumed that the alternative hypothesis is true as the p value = < 0.001 so it means that there is association between gender (sex=male & female) and diabetes mellitus and this association is proven statistically above. There is statistically significant evidence which demonstrates that the male diagnosed more diabetic as compare to female and there is association between gender (sex=male & female) and diabetes mellitus and this is proven statistically above.   b=0 cells (.0%) have expected count less than 5. The minimum expected count is 11.59.

Comparison of Weight of Males and Females with Two Sample T Test (PARAMETERIC TEST)
Here I want to compare the weight of the male and the female (sex) through two sample T test.
We used descriptive analysis to find out the results Table   5 shows total numbers of people are 2115 which are the valid numbers of males and females and the mean for weight in kg is 69.97. The standard deviation for sex is 0.497 and the standard deviation for weight in kg is 19.056. Then we used histogram to compare the weight of male and female as (Figure 2). The x axis shows the weight in kg and the y axis shows the frequency both for male and for the female and it shows that males are heavier than females. I am using the two sample T test a parametric test for the comparison of weights of the (sex) males and the females and our null and alternative hypothesis are as:   Table 7 shows that the t value of the variances assumed is 19.152 and the degree of freedom for this is 2113 and the mean difference

Levene's Test forEquality of Variances t-test for Equality of Means
P value = 0.000 for both variances assumed and not assumed.
As p value = <0.05 so it rejects the null hypothesis.
Therefore, it can be assumed that alternative hypothesis is true and male and female have different weight and the mean weight of male is 78.31 and the mean weight of female is 64.62 in kgs. There is statistically significant evidence which demonstrates that the Males and females do not have the same or equal weight and males are heavier than the females. Medically males have more food intake as naturally required by the body and females have less food intake and less hunger naturally, males have faster metabolism than females they eat more than females and consume and store more calories than females. So, males are heavier than the females naturally.
The height of the males is more than females and the bones of the males are much heavier than females due to more calcium deposition in bones of males than females. The structure of the females is smaller when comparing with male structure.

Males and Females
To find out the confidence interval with one interval i.e. BMI we use the descriptive statistics and then exploratory analysis to find out 95% confidence interval of BMI of unknown Table 8 shows the BMI of 2138 people in which 71 are missing and the valid numbers of people are 2067. 95% confidence interval mean of male and the female are slightly different statistically as proven above.