Free Motion for Miscellaneous Relief - District Court of Delaware - Delaware

Case 1:04-cv—01560-JFC Document 35-2 Filed 08/19/2005 Page1 0f4
EXHIBIT A

Case 1 :04-cv—01560-JFC Document 35-2 Filed 08/19/2005 Page 2 of 4
Summary of Personal Injury Claims Estimation Sampling Plan
Statistical sampling is a method of using established mathematical formulas to
validly infer characteristics of an overall population by examining representative
"samples," or subgroups, of that population. Statistical sampling methods are useful
when examining the entire population would be impossible, impracticable, or highly
inefficient and costly. Here, for example, there were approximately 150,000 claims
pending against Debtors at the time they filed their bankruptcy petition. Examining every
one of those claimants to determine key characteristics germane to the issues that Debtors
expect to raise in this action (e. g., whether or to what degree a claimant was exposed to
any of Debtors’ products that contained chrysotile asbestos, whether a claimant’s
radiographic readings and pulmonary functions tests indicate they are impaired by
asbestosis), would be costly and unduly time constuning. By defining a representative
sample of claimants and by collecting data from that sample, accurate conclusions can be
. drawn about the overall claimant population that will assist the Court in evaluating the
impact of Debtors’ various defenses.
Debtors propose to take discovery from a sample of approximately 1000 present
personal injury claimants using a standardized set of written questions. This discovery
will compile key information from this sample, including, without limitation, the
claimants’ occupational history (including whether the claimant worked in the
construction industry), medical history, claimed disease and asbestos litigation history.
A copy of Debtors’ draft questionnaire is attached as Exhibit A.
ln a stratified sampling plan, the overall sample is divided into separate sub—
samples taken from subgroups of the population. From this sample, the parties can draw
- 1 -
asassz vi/sr
08/03/05 5:12 PM

Case 1 :04-cv—01560-JFC Document 35-2 Filed 08/19/2005 Page 3 of 4
statistically valid conclusions regarding key characteristics ofthe population as a whole
as well as subgroups within that population. Using stratified sampling, valid conclusions
regarding the characteristics of subgroups ofthe claimant population, such as claimants
with specific claimed diseases, can be drawn (ag., what percentage of asbestosis
claimants were exposed to a minimum level of chrysotile asbestos necessary to cause
asbestosis).
Debtors propose to stratify the overall sample of 1000 into separate sub-samples
corresponding to each category of disease claimed (as recorded in the CCR database).
Debtors propose to sample 200 mesothelioma claimants, 200 lung cancer claimants, 200
asbestosis claimants, and 200 other cancer claimants. Within each disease category,
debtors propose to sample 100 claimants who are identified in the CCR database as
having worked in the construction industry and 100 claimants who did not.
To account for the fact that a significant number of claimants in the CCR database
failed to report any specific disease, Debtors also propose to include a sub-sample of 200
such claimants. Claimants in the "unl to the appropriate disease category once claims forms containing this information are
returned by claimant. This claim f`orm information may also be used to calibrate any
disease category information imputed to unknown/tmstated diseases in the CCR database.
Using a sample size of 200 claimants per disease category will ensure that, within
any specific sub—sample, it can be concluded with 95% confidence that any given
characteristic ofthe sampled population will be within 7% of that within the actual
population.
Debtors propose to identify claimants to be sampled using a stratified systematic
- 2 -
959582 vrxsr
as/02/05 5:12 PM

Case 1 :04-cv—01560-JFC Document 35-2 Filed 08/19/2005 Page 4 of 4
sample with random start method, further described as follows:
(1) Each of the approximately 150,000 claimants in the CCR database
will be sorted into 10 groups corresponding to the 5 disease
categories subdivided by occupational category (i.e. , occupational
history in construction or non~construction), with claimants in each
group sorted alphabetically by state of residence.
(2) For each of the 10 disease category/occupational history groups a
sampling ratio H will be determined to produce a total stratum size
sample of 100. (For example, if the stratum size was 20,000, the
sampling ratio H would be 200 because by selecting every 200m
plaintiff from the stratum, you would end up with a sample of
100). A sub—sample for each disease/occupational history category
will then be constructed by sampling every nth individual,
beginning with a start selected at random.
(3) Individual claimants are sampled in this fashion until an overall
sample of 1000 is constructed.
- 3 -
asassz vi/sr
oz/03/05 5:12 PM