MO/WBL/46E
L_ C. F. B.
7 th August 1980
4 AUG 1980
ACK .........
FILE ...
A DELTA PANEL EVALUATION OF SOME CANADIAN BRANDS
The GR&DC DELTA panel was required to undertake an
evaluation of the following Canadian brands,
supplied by
Imperial Tobacco Ltd:
No. Brand Tar/Nic
(Mg.)
1. Players Extra Light 9.0/0.9
2. Matinee 8.0/0.5
3. Coded 410 "Joy"
4. Peter Jackson Extra Light 7.o/o.7
5. Medallion Ultra Mild 1.0/0.1
6. Craven A Ultra Light 1.0/0.1
7. Craven A Special Mild 4.0/0.4
8. Accord Ultra Mild 3.0/0.3
9. Vista 12.0/1.2
10. Matinee Extra Mild 4.0/0.4
11. Viscount 5.0/0.4
12. Viscount No.1 Ultra Light 0.7/0.07
No delivery information available at time of test.
The cigarettes were smoked in standard fashion by
the panel
with order of presentation randomised between
panellists. Six
cigarettes were smoked at the first session and
six at the second,
one day later. The results were analyzed by the
INGRID program
to yield measures of the overall similarity
between products
(inter-cigarette correlations), the relationships
between the
assessment parameters (inter-parameter
correlations) and the 1-0
00
relative profiles of each of the cigarettes
(cigarette by _-1
parameter correlations). -41--
BATCo document for Province of British Columbia 10
November 1999


-2-
Results: Figure 1 is a representation of the data
in two-
dimensional space, following the principal
components
analysis of the mean scores by INGRID. It can be
seen that
very little discrimination between products has
been achieved,
beyond a general tracking of "strength".
Inspection of the
analysis revealed that more than 95% of the total
variance is
accounted for by the first component. Only for the
higher
delivery cigarettes (nos. 1, 2, 3, 4 and 9) is
there any evidence
of dispersion on the second component.
Furthermore, the
correlations between the assessment parameters was
unusually
high (> . 90) , suggesting that the "bandwidth" of
the products
being evaluated was too wide. It was
therefore.decided to
re-evaluate a sub-set of the original sample
supplied,
concentrating on those products with a tar
delivery of 5mg.
or less. Accordingly, a second evaluation was made
using the
following cigarettes. Note (1) that the code
numbers have
been changed and (2) that a further cigarette
(John Player Ultra
Mild K.S., ex-U.K) has been added to the ensemble.
This has no
particular significance but was included in order
to bring the
number of products to the usual minimum required
by the method.
No. Brand Tar/Nic
1. John Player Ultra Mild K.S.
2. Accord Ultra Mild 3.o/o.3
3. Craven A Special Mild 4.0/0.4
4. Medallion Ultra Mild 1.0/0.1
5. Viscount No.1 Ultra Light 0.7/0.07
6. Craven A Ultra Light 1.0/0.1
7. Matinee Extra Mild 4.0/0.4
8. Viscount 5.0/0.4
co
CD
BATCO document for Province of BritiSh Columbia
`10 November 1999


-3-
Results: Figure 2 shows the disposition of the
products and
parameters in respect of the first two principal
components. On this occasion 90% of the variance
is accounted
for by the f irst principal component and nearly
6% relates to
the second. It can be seen from Figure 2 that a
greater
dispersion between parameters and products has
indeed been
achieved now that attention has been focussed on
this lower
delivery sub-set.
It should be remembered that the plot is an
estimate of
the "best fit" for the data and should not be
relied upon for
detailed interpretation. Nevertheless it can be
seen that
cigarette 5 (Viscount No.1 Ultra Light) adopts its
expected
position to the left of the plot. (our convention
is to
label the "high" end of each assessment
parameter). The two
one mg. products 6. (Craven A Ultra Light) and 4.
(Medallion
Ultra Mild) are both found in appropriate
relationship with
Viscount No.1, with the ITL product apparently
giving a slightly
better account of itself than the Craven
cigarette.
Cigarette 3 (Craven A Special Mild) is somewhat
distant
from the other 4mg. product, No.7 (matinee Extra
Mild). The
position of cigarette 8 (Viscount) is influenced
by its
apparently dominant mouthfeel impression whilst
cigarette 2
(Accord Ultra Mild) achieves a position which
belies its
relatively modest delivery levels.
More detailed appreciations of each cigarette's
performance
can be obtained by reference to the tabular
information. First,
consider the overall similarities between the
products (Table 1).
It can be seen that, overall, and disregarding the
U.K.
product, the two most similar products are Accord
Ultra Mild
and Matinee Extra Mild (r = .93) and Viscount No.1
Ultra Light
and Craven A Ultra Light (r = .93). Not
suprisingly Medallion
Ultra Mild was perceived as being overall most
similar to
viscount No. 1 Ultra Light (r = .76) and Craven A
Ultra Light (r = .73)
I S 6Z 9 6 0 1
BATCo document for Province of British Columbia 10
November 1999


-4-
otherwise it is dissimilar to all the other
Canadian products
evaluated.
The relative profiles of each of the products can
be
inferred from Table 2 which shows the correlations
between
products and parameters. It should be noted that
the values
of the correlations in this example are atypically
high,
reflecting the high loading on the first principal
component.
The cigarettes achieving the highest ACCEPTABILITY
are
Accord Ultra Mild and Matinee Extra Mild (r =
.93). Their
profiles are remarkably similar with the Matinee
cigarette
achieving slightly better mechanics (MOUTHFULL and
DRAW EFFORT)
and giving a higher irritation, particularly in
the mouth
(MOUTHFEEL). Accord does better in terms of
IMPACT. It is
our view that high ACCEPTABILITY can be achieved
in a variety
of ways; given a minimum standard on mechanics,
irritation/impact,
and taste it is probably the balance between these
components
which determines the overall ACCEPTABILITY. Within
limits,
we would suggest, it is possible to trade poor
performance on
one aspect for better performance on another. in
this way,
perhaps, Accord is able to achieve parity with
Matinee Extra
Mild in terms of ACCEPTABILITY despite-somewhat
different
relative profiles.
The least ACCEPTABLE Canadian cigarettes were
Craven A
Ultra Light (r = -.91), Medallion Ultra Mild (r =
-.90) and
Viscount No.1 Ultra Light (.r = -.89). inspection
of the
profiles would suggest that Medallion Ultra Mild
is a "better"
product than its ACCEPTABILITY rating would
suggest since it
performs better than the other two on all other
parameters.
This anomaly is probably a reflection of the
expectations of
this particular (U.K) panel.
Both Matinee Extra Mild and Viscount display good
mechanics
but the ITL product provides higher impressions of
irritation (:~D
in the nose (NOSEFEEL) and throat (THROATFEEL),
has a higher 1-0
00
IMPACT, and gives greater impressions of both
TASTE and AROMA.
00
BATCO document for Province of BritiSh Columbia 10
November 1999


Cigarette 3 (Craven A Special Mild) presents a
"middle-of-the-road"
profile perfectly consistent with its location in
Figure 2.
Finally, inspection of Table 3 shows the
interrelationships
between the assessment parameters. it should be
noted that
there is an unusual polarity of association
between the parameters
which is not usually found in these evaluations.
it is to be
recalled that (1) the range of differences between
these products
is quite small, and (2) for this "bandwidth" of
deliveries, the
parameters are probably highly intercorrelated in
an objective
sense anyway.
It is usually instructive, although less so in the
present
case, to inspect the correlations with
ACCEPTABILITY in order
to determine the parameters most influential in
achieving
ACCEPTABILITY. Although the differences in the
present case
are small, it appears that a low DRAW EFFORT (r =
.96) and a
high level of TASTE (r = .94) are particularly
important in
determining ACCEPTABILITY for these specific
cigarettes on
this occasion for the GR&DC panel. It should be
emphasised
that determinations of ACCEPTABILITY are
undertaken for purely
intrinsic reasons, as demonstrated here, and as
yet have no
predictive value in relation to market
acceptability or consumer
response.
Finally, one cannot fail to observe that all the
Canadian
products entering this evaluation fared better on
almost all
parameters than the U.K. John Player Ultra Mild
X.S. which was
perceived by this panel to be the least
acceptable. cigarette.
C__J
BATCo document for Province of British Columbia 10
November 1999


FIGURE I
ACCEPTABILITY
IMPACT .
i.L
,,12 6 .5 7 i
DRAW EFFORT 1 9 -
IS
.8 19
.1 TASTE
. .3 AROM
.4
NOSEFEEL/THROATFEEL
MOUTHFULL
-2 MOUTHFUL
~ 9 6 ~ Z G 6 0 1


FIGURE 2
DRAW EFFORT
MPACT
AROMA
TASTE
THROATFEEL
ACCEPTABILITY
NOSEFEEL
MOUTHFULL
MOUTHFEEL
S20 R60 1


0
0
CL
0
0
a
3
Q
T
1
0
<
(D
0
0
0
z
0
CD
3
cr
CD
TAB-- 1 INTER-CIGARETTE CORRELAT7 NS
JOHN PLAYER ULTRA MILD K.S.(U.K)
ACCORD ULTRA MILD
CRAVEN 'A' SPECIAL MILD
MEDALLION ULTRA MILD
VISCOUNT NO.1 ULTRA LIGHT
CRAVEN 'A' ULTRA LIGHT
MATINEE EXTRA MILD
HOZB601
1
2
3
4
5
6
7
im
E-4
0
u
L)
4
2
-.93
z
P4
u En
3
-.24
.90
z .1
0
H
Q
4
.77
-.83
-.90
0 E-4
z 2:
0
C-4 H
6.4
80
(n H
m ~
>
5
.95
-.92
-.43
.76
- m
4 0
- H
4
.0
6
.94
-.83
-.49
.73
.93
p
pq
W
%41
0
7
-.99
.93
.31
-.78
-.97
-.94
r-4
6
0
u
8
-.90
.72
.39
-.75
-.84
-.93
.86


>
TABLE 2 CORRELATIONS BETWEEN CIGARETTES AND
ASSESSMENT PARAMETERS
0
0
CL
0
0
c
CD
E-1 F--
z w E-
U ;~ U)
0
H E-1
R E-
UH
O z 0
w
E, U
El
E-1 H a ci
En
0
u ~
>
0
1
2
3
4
5
7 >
0
0
h MOUTHFULL -.95 .82 .46 -.76 -.95 -.96 .94 .96
DRAW EFFORT .96 -.84 -.27 .82 .90 .94 -.93 -
98
.
0 MOUTHFEEL -.89 .72 .62 -.60 -.90 -.98 .90 .91
0
NOSEFEEL -.97 .91 .43 -.8o -.99 -.96 .99 .88
THROATFEEL -.96 ..95 .36 -.75 -.99 -.94 .98 .82
Q
z
0 IMPACT -.85 .98 .16 -.78 -.90 -.78 .89 .60
'ASTE -.96 .98 .15 ".83 -.94 -.87 .98 .78
cr
AROMA -.96 .96 .02 -.71 -.89 -.83 .94 .75
(D
c0 ACCEPTABILITY -.95 .93 .14 -.90 -.89 -.91 .93
.89
Z80Z6601


VD
>
q
0
0
CL
0
3
(D
0
(D
0
0
0
cr
C)
z
0
cr
m
MOUTHFULL 1
DRAW EFFORT 2
MOUTHFEEL 3
NOSEFEEL 4
THROATFEEL 5
IMPACT 6
TASTE 7
AROMA 8
88 OZ8601
T.KBLE 3 INTER-ASSESSMENT PAPAtliz2ER CORRELATIONS
0
Ul El
co
0
0 W C4
2 3 4 5 6 7 8
-.96 .94 .97 .93 .75 .87 .82
-.89 -.93 -.89 -.74 -.89 -.86
.92 .88 .67 .79 .75
.98 .88 .95 .89
.93 .96 .93
.94 .89
.97
E-4
E-1
9
.90
-.96
.80
.92
.91
.85
.94
.91