Submitted by: Submitted by

Views: 384

Words: 974

Pages: 4

Date Submitted: 11/19/2012 05:44 AM

Report This Essay

STUDENT’S ID: 2012003, 2012015, 2012035, 2012045, 2012046

CHINESE NAME: 曹明镇, 李昇潤, Duangnapa Rojphongkasem, 王远方, 武小宁

ENGLISH NAME: MJ, Steve, Bing Bing, Alonzo, Hughes

PROFESSOR: Sean X. Chen

DATE/(Y/M/D): 2012/09/11

I hereby certify that the following work is entirely and exclusively my own.

Signature:

Score:

(1) We propose to use two types of statistical methods to estimate the total damage - ‘confidence interval’ and ‘hypothesis testing (one sample t test)’. The confidence interval method should be used to estimate the range of the total damage while the hypothesis testing should be used for determining whether the damage exceeds the insured amount.

(2) In order to facilitate our statistical analysis, we need the following information.

Information Required Response from clients

Accounting document processing procedure

(e.g. How are the multiple vouchers grouped to a single invoice? By drivers? By periods?) Vouchers are grouped into invoices on a random basis

The confirmation of the accounting firm on whether the selected samples are real or fake The validity of vouchers will be determined by the accounting firm and hence these information is not necessary

Types of fake vouchers: increasing of the reimbursement amount or creating non-existing vouchers The amount of vouchers can only be 100% true or fake

Confidence level and margin of error acceptable by insurance company, blood center and accounting firm The preferred confidence level by the insurance company is 99%. There is no specific requirement on margin of error.

What is the cut-off value to determine whether the conduct is considered civil or criminal? If the damage is higher than \$100,000, it is considered as fraud.

(3) The sample should be...