CASE STUDY
Fire-Safe Cigarettes
In the United States in 2008, 680 deaths resulted from
fires caused by dropped or discarded cigarettes, and over
114,000 fires were started by a lighted tobacco product.
The dollar value of the property lost in these fires is
staggering. Thirty-four percent of the victims of cigarette-
induced fires are the children of the smokers. About 40%
of cigarette-induced fire victims were sleeping. (Source:
www.firesafecigarettes.org)
Certainly, it makes sense to develop a cigarette that
extinguishes itself when left unattended (as would happen
if someone falls asleep while smoking). The state of New
York decided to require cigarettes sold in its stores to be
designed so they self-extinguished when left unattended.
Two types of cigarettes were tested to determine their
propensity to self-extinguish. For each type (A and B),
40 cigarettes were lit and allowed to burn unattended until
the cigarette either extinguished or did not. Cigarette brand
A was designed to have ultra thin concentric paper bands
manufacturers were concerned about was the amount of
affixed to the traditional cigarette paper. These bands are
nicotine in each cigarette. The following data represent the
referred to as “speed bumps” and cause extinguishing of amount of nicotine in each brand of cigarette for a random
the cigarette by restricting the flow of oxygen to the burning sample of 15 cigarettes.
ember. Cigarette brand B was a traditionally designed
model. The results of the experiment are in the following
table, where E represents an extinguished cigarette and F BRAND A
represents a full-burn cigarette. 1.40 1.36 1.42 1.30 1.40
1.09 1.06 1.06 1.12 1.12
BRAND A 1.14 1.14 1.13 1.21 1.15
E E E E E E F F F F
E E F E E E E E E E BRAND B
E E E E E F E F F E 1.24 1.36 1.32 1.32 1.36
F E E F F F E F E E 1.23 1.13 1.20 1.28 1.10
1.18 1.19 1.20 1.20 1.20
BRAND B
F F F F F F F F F F Write a report detailing the propensity of each
brand to self-extinguish and the level of nicotine in each
F F F F F F E F F F
brand. Include any relevant confidence intervals. If
F F F F F F F F F F the normal model or Student’s t-distribution cannot be
F F F F F F F F F F used to construct a confidence interval, use alternative
models (such as the Agresti–Coull model for estimating
Naturally, cigarette manufacturers were concerned proportions—see Problems 43 and 44 from Section 9.1 or
about the additional cost of manufacturing such a cigarette. the bootstrap methods of Section 9.5). Would you support
Plus, the manufacturers were concerned that consumers legislation that reduced the risk of fire from unattended
might prefer the cigarettes without “speed bumps” to cigarettes? How much extra would you be willing to pay
those with design changes. One particular measure the for such a cigarette if you were a smoker?
475