Professional Documents
Culture Documents
(p^±z(p^−pon))=(0.50±2.5760.4295(1−0.4295)3940)=[0.480,0.520]=[48%,52%]\left ( \hat{p}
\pm z\left ( \sqrt{\frac{\hat{p}-p_o}{n}}\right )\right )=\left ( 0.50\pm
2.576\sqrt{\frac{0.4295(1-0.4295)}{3940}} \right )=[0.480,0.520]=[48\%, 52\%](p^±z(np^
−po))=(0.50±2.57639400.4295(1−0.4295))=[0.480,0.520]=[48%,52%]
n(1−p)=200(1−14200)=186≥30n(1-p)=200(1-\frac{14}{200})=186\geq
30n(1−p)=200(1−20014)=186≥30 (satisfied for normal distribution)
c) SE=p(1−p)n=14200(1−14200)200=0.018SE=\sqrt{\frac{p(1-p)}{n}}={\sqrt\frac{\frac{14}
{200}(1-\frac{14}{200})}{200}}=0.018SE=np(1−p)=20020014(1−20014)=0.018
d) z=p^−pop(1−p)n=0.06−1420014200(1−14200)200=−0.554z=\frac{\hat{p}-p_o}
{\sqrt{\frac{p(1-p)}{n}}}=\frac{0.06-\frac{14}{200}}{{\sqrt\frac{\frac{14}{200}(1-\frac{14}
{200})}{200}}}=-0.554z=np(1−p)p^−po=20020014(1−20014)0.06−20014=−0.554
e) P(z<−0.554)=0.29P(z<-0.554) = 0.29P(z<−0.554)=0.29
f) Since, p-value of our test statistics is greater than the significance level of say 0.05 . We have
insufficient evidence to reject null. So, we accept null that the sample proportions of Brazilian
adults with diabetes is greater than 6%.
3) a) CI for population for significance level of α=10%\alpha =10\%α=10%, its critical
value tα2=±1.663t_\frac{\alpha }{2}=\pm 1.663t2α=±1.663 (refer to t-table, two-tailed with df=
n-1 = 90-1 = 89)
c) SE=sn=5.154=0.694SE=\frac{s}{\sqrt{n}}=\frac{5.1}{\sqrt{54}}=0.694SE=ns=545.1
=0.694
d) Test statistic,
t=xˉ−μsn=17.9−16654=2.327t=\frac{\bar{x}-\mu }{\frac{s}{\sqrt{n}}}=\frac{17.9-16}
{\frac{6}{\sqrt{54}}}=2.327t=nsxˉ−μ=54617.9−16=2.327
Since p-value of F-statistic 12.3, at the last column is less than significance level of say 0.05.
Therefore, over all significance of the model is significant.
Here are the result of 105 observations for monthly rent of a property(response or dependent
variable) together with its predictor or independent variables, Size, Time and Pool . Regression
output table above shows the coefficients of each , we can now conclude regression line as,
let y^=rent\hat{y}=renty^=rent
y^=7415.7603+1.6941(size)−165.2999(time)
+1139.6987(pool)\hat{y}=7415.7603+1.6941(size)-165.2999(time)+1139.6987(pool)y^
=7415.7603+1.6941(size)−165.2999(time)+1139.6987(pool)
Conclusion: We can use this linear regression in which all dependent variables are included or
significant in the model since in the above table, (column of P-value), all of them has p-value of
less than 0.05 significance level. Thus, time, size and pool provide good linear fit for our
prediction model for rent.
As you can see, this three independent variables or factor affect the value of rent of a property.
Say, If Build-up area (sq ft) of a property is 1 unit (size=1), and time is zero minute (time=0) or
the house is just near MRT station and no pool at all.(pool=0) , the Value of rent will be
y^=7415.7603+1.6941(1)−165.2999(9)+1139.6987(0)=7417.74\hat{y}=7415.7603+1.6941(1)-
165.2999(9)+1139.6987(0)=7417.74y^
=7415.7603+1.6941(1)−165.2999(9)+1139.6987(0)=7417.74.
Imagine if we the property is away from the MRT station, say t= 10 minutes and still no pool ,
the value of rent will change to cheaper one :
y^=7415.7603+1.6941(1)−165.2999(10)+1139.6987(0)=5764.46\hat{y}=7415.7603+1.6941(1)-
165.2999(10)+1139.6987(0)=5764.46y^
=7415.7603+1.6941(1)−165.2999(10)+1139.6987(0)=5764.46
As you can see, these factors like time, size and availability of pool changes the value of rent.
__________________
Using excel, (DATA ANALYSIS) you can get the above summary for t-test, two sample
assuming unequal variance.
Or simply execute this in excel, to solve for mean in each column, No sibling and With Sibling.
Average(data)
t-statistic : t= 0.00549
with df = 38, P(T<t)= P(T<0.00549)=0.9956 (it falls in the rejection region of null )
b. Provide a 90% confidence interval estimate of the mean difference between the critical
reading SAT scores for the twins raised with no siblings and the twins raised with siblings.
(, ) (to 2 decimals)