Scribd
Upload
445 views

Assignment 2 - Solved

This document contains answers provided by Ramashis Biswas to four questions. In the first answer, Ramashis calculates the mean, confidence intervals at 99%, 95%, and 90% levels, and margin of error for a dataset. Ramashis notes that the margin of error decreases as the confidence level decreases. The second answer discusses estimating percentiles from a non-normal population. The third answer addresses assuming house prices do not change over time and applying the central limit theorem. The fourth answer calculates 95% and 99% confidence intervals for the proportion of "old" houses in a sample of 302 houses.

Uploaded by

lordavenger
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
445 views

Assignment 2 - Solved

This document contains answers provided by Ramashis Biswas to four questions. In the first answer, Ramashis calculates the mean, confidence intervals at 99%, 95%, and 90% levels, and margin of error for a dataset. Ramashis notes that the margin of error decreases as the confidence level decreases. The second answer discusses estimating percentiles from a non-normal population. The third answer addresses assuming house prices do not change over time and applying the central limit theorem. The fourth answer calculates 95% and 99% confidence intervals for the proportion of "old" houses in a sample of 302 houses.

Uploaded by

lordavenger
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Ramashis Biswas | PGID: 61510121 | Section: H | Ramashis_Biswas2015@isb.

edu

Answer 1:
Mean: 163862.13

Confidence Interval 99%

95%

90%

Upper Limit 169257.41

167964.69

167304.17

Lower Limit 158466.84

159759.56

160420.08

4102.565

3442.045

Margin of Error 5395.285

The Margin of Error decreases as level of confidence decreases. This is acceptable, as it is


essentially a trade-off between precision and confidence. At 100% confidence, this would
essentially mean the entire range (not precise at all) while at 0% confidence, this would be
essentially a point estimate the mean of a single sample.
For the 99% CI, required n > (t99%, 1046*S/DMOE) 2, or n ~> 32 and hence CLT holds.

Answer 2

The percentile of the population will be an ordered statistic. In order to estimate the same, we
will need more information on the population distribution. Also, since the population is quite
obviously not normal, it would not be possible to make statements about the percentile from
the confidence intervals of the mean.

Answer 3
If we consider that the house prices do not change over time, then each years sample should
represent a simple random sample (assuming that prices have not gone up on resale either,
and houses sold in one year have an equal opportunity of being sold in the next year). Thus
the distribution of the sample averages, as per CLT will be a Normal Distribution with mean
and Variance 2/n.

Answer 4
The sample has 302 houses aged more than 30, i.e. 29% of the sample is old.

From CLT, the 95% and 99% CI for the proportion of old houses will be:

Ramashis Biswas | PGID: 61510121 | Section: H | [email protected]

95%: 29% + Z95% (p (1-p))/n = [26%, 32%]


99%: 29% + Z99% (p (1-p))/n = [25%, 32%]

This essentially means that the proportion of old houses in the population has a (95/ 99) %
chance of being in the specified intervals.
Or, in (95/ 99)% of random samples will have a proportion in the specified confidence
intervals.

You might also like