You are on page 1of 5

Expanded Tables for Multiple Comparison Procedures in

the Analysis of Ranked Data


G. J. NEWELL and J. D. MacFARLANE

Table 1-Critical absolute rank sum differences for “all treatments” com-
ABSTRACT parisons at 5% level of significancea,b
Multiple comparison procedures involving Friedman rank sums for Number of samples
the analysis of ranked data have recently been proposed as alternatives
Panelists 3 4 5 6 7 8 9 10 11 12
to tests developed by Kramer. Expanded tables for these multiple
comparison procedures (involving the “all treatments” and “treat- 3 6 8 11 13 15 18 20 23 25 28
4 7 10 13 15 18 21 24 27 30 33
ments versus control” comparisons) were presented to accomodate 37
5 8 11 14 17 21 24 27 30 34
the range of panelists and samples usually encountered in sensory 6 9 12 15 19 22 26 30 37 42
evaluation experiments involving ranked data. 7 10 13 17 20 24 28 32 9: 40 44
8 10 14 18 22 26 30 34 39 43 47
9 10 15 19 23 27 32 36 41 46 50
INTRODUCTION 10 11 15 20 24 29 34 38 43 48 53
11 11 16 21 26 30 35 40 45 51 56
TESTS developed by Kramer for assessingsamples that have 12 12 17 22 27 32 37 42 48 53 58
been ranked (Bradley and Kramer, 1957; Kramer, 1956,196O) 13 12 18 23 28 33 39 44 50 55
have been used extensively in sensory evaluation for the last 14 13 18 24 29 34 40 46 52 57 :i
15 13 19 24 30 36 42 47 53 59 66
30 years, with corrected and expanded tables available (Kahan 16 14 19 25 31 37 42 49 55 61 67
et al., 1973; Kramer, 1963). Computerized versions of this 17 14 20 26 32 38 44 50 56 63 69
methodology have been developed (McLellan et al., 1984), as 18 15 20 26 32 39 45 51 58 65 71
well as the tests discussed and tables of critical values pre- 19 15 21 27 33 40 46 53 60 66 73
20 15 21 28 34 41 47 54 68 75
sented in the standard sensory evaluation textbooks (Amerine 21 16 22 28 35 42 49 56 El 70 77
et al., 1965; Kramer and Twigg, 1970). 22 16 22 29 36 43 50 71 79
Joanes(1985) has recently suggestedthat these multiple com- 23 16 23 30 37 44 51 ii;: Ez 73 80
parison tests are invalid for their stated purpose becausethe rank 24 17 23 30 37 45 52 59 67 74 82
25 17 24 31 38 46 53 61 68 76 84
sums of different samples are not independent of one another 26 17 24 32 39 46 54 62 70 77 85
and the probability level for one of the tests is incorrect. The 27 18 25 32 40 47 55 63 71 79 87
resulting multiple comparison procedures proposed by Joanes 28 18 25 33 40 48 56 64 72 80 89
(1985) to overcome these weaknessesin Kramer’s tests utilize 18 26 33 41 49 57 65 73 82 90
ii 19 26 34 42 50 58 66 75 83 92
the distribution of the range of rank sums for each rank sum
configuration with these non-parametric multiple comparison i1 1: 9; i"j ff z1 z; ill ;9 ii i:
proceduresdiscussedin Hollander and Wolfe (1973), as well as 61 70 78 87 96
tables of critical values being available (Dunn-Rankin and Wil- ii 2 fS i: ii z3 62 71 79 89 98
35 20 28 37 45 54 63 72 81 90 99
coxon, 1966; Hollander and Wolfe, 1973; McDonald and 36 20 29 37 46 55 63 73 82 91 100
Thompson, 1967). However, these tables only include critical 37 21 29 38 46 55 64 74 92 102
values for up to 15 panelists and as such -are limiting for their 38 21 29 38 47 56 75 :: 103
effective use in sensoryevaluation experimentation. The purpose 39 21 30 39 48 57 :; 76 85 zz 105
76 86 96 106
of this paper was to expand the scope of the available tables for fY ;: i: :: ii 5": :: 77 87 97 107
use in multiple comparisons for ranked data to accomodatethe 42 22 31 40 49 59 69 78 88 98 109
range of panelists (up to 100) and samples (up to 12) usually 43 22 31 41 50 60 69 79 89 99 110
encounteredin sensory evaluations. 44 22 32 41 51 60 70 80 90 101 111
45 23 32 41 51 61 71 81 91 102 112
46 23 32 42 52 62 72 82 92 103 114
47 23 33 42 52 62 72 83 93 104 115
THEORY METHODS 48 23 33 43 53 63 73 84 94 105 116
“All treatments” comparisons 49 24 33 43 53 64 74 85 95 106 117
50 24 34 44 54 64 75 85 96 107 118 I
To test whether any of the samples are superior or inferior to any 55 25 35 46 56 67 78 90 101 112 124
of the other samples, Joanes (1985) proposed the use of a distribution- 60 26 37 48 59 70 82 94 105 117 130
free test based on Friedman’s rank sums (Hollander and Wolfe, 1973) 65 27 38 50 61 73 85 97 110 122 135
involving the distribution of the largest absolute range between rank 70 28 40 52 64 76 88 101 114 127 140
75 29 41 53 66 79 91 105 118 131 145
sums. For a sensory evaluation involving n panelists ranking k sam- 80 30 42 55 68 81 94 108 122 136 150
ples, this test is based on all (k!)” possible rank configurations being 85 31 44 57 70 84 97 111 125 140 154
equally likely, with two samples “a” and “b” considered to be sig- 90 32 45 58 72 86 100 114 129 144 159
nificantly different if: 95 33 46 60 74 88 103 118 133 148 163
100 34 47 61 76 91 105 121 136 151 167
I&- &I 2 4&n) (1) “Exact values adapted from Hollander and Wolfe (1973) are used for up to 15 panel-
ists.
where R,, Rb are the rank sums for samples “a” and “b”, respec- blnterpolation may be used for unspecified table values involving more than 50 panel-
tively; (Y = appropriate experimentwise error rate; and r(cx,k,n) = ists.
table value such that P((R,- Rt,l < r(cu,k,n)) = 1 --(Y.
While tables of critical values are available (Dunn-Rankin and Wil-
coxon, 1966; Hollander and Wolfe, 1973; McDonald and Thompson,
1967), these tables are limited becausethe true probability distribution
of the range of rank totals is generally unknown, resulting from dif- I
ficulties in obtaining a complete enumeration of the (k!)” possible
Authors Newell and MacFarlane are with Hawkesbury Agricul- rank configurations. For example, 10 panelists ranking four samples
tural College, Richmond NSW 2753, Australia
would involve enumerating (4.)t lo = 6.3 x lOI3 possible rank con-

Volume 52, No. 6, 1987-JOURNAL OF FOOD SCIENCE- 1721


ANALYSIS OF RANKED DATA. . .
Table 2-Critical absolute rank sum dhyerences for “all treatments” com- Table 3-Rankings of four samples by 20 panelists
parisons at 1% level of significant@
Samples
Number of samples Panelists A B C D
Panelists 3 4 5 6 7 8 9 IO 11 12
1 1 2 4 3
3 9 12 14 17 22 24 30 2 1 2 4 3
4 s 11 14 17 20 26 29 36 3 1 2 4 3
5 9 13 16 19 23 26 30 33 37 41 4 1 3 4 2
6 IO 14 18 21 25 29 33 37 41 45 5 1 3 4 2
7 11 15 19 23 28 32 36 40 45 49 6 1 2 4 3
8 12 16 21 25 30 34 39 43 48 53 7 1 3 4 2
9 13 17 22 27 32 36 41 46 51 56 8 1 2 4 3
IO 13 18 23 28 33 38 44 49 54 59 9 1 3 4 2
11 14 19 24 30 35 40 46 51 57 63 IO 1 2 4 3
12 15 20 26 31 37 42 48 54 66 11 1 2 4 3
13 15 21 27 32 38 44 50 56 :; 68 12 1 2 4 3
14 16 22 28 34 40 46 52 58 65 71 13 1 3 4 2
15 16 22 28 35 41 48 54 67 74 14 1 4 3
16 17 23 30 36 43 49 56 :: 70 77 15 1 3 4 2
17 17 24 31 37 44 51 58 65 72 79 16 1 3 4 2
18 18 25 31 38 45 52 60 74 81 17 1 3 4 2
19 18 25 32 39 46 54 61 Ei 76 84 18 2 3 4 1
20 19 26 33 40 48 55 63 70 78 86 19 1 2 3 4
21 19 27 34 41 49 56 64 72 80 88 20 1 2 4 3
22 20 27 35 42 50 58 66 74 82 90 Ranksums 21 49 79 51
23 20 28 35 43 51 59 67 75 84 92
24 21 28 36 44 52 60 69 77 85 94
25 21 29 37 45 53 62 70 79 87 96
26 22 29 38 46 54 63 71 80 89 98 Table 4-Rank sum analysis for 4 samples
27 22 30 38 47 55 64 73 82 91 100
Samole Ranksuma
28 22 31 39 48 56 65 74 83 92 101
29 23 31 40 48 57 66 75 85 94 103 A 21
30 23 32 40 49 58 67 77 86 95 105 B 49a
23 32 41 50 69 78 97 107 C 79
:: 24 33 42 51 6": 70 79 8"; 99 108 51a
33 24 33 42 52 61 71 80 90 100 110 a Rank sums not followed by the same letter are significantly different (P~0.01).
34 25 34 43 52 62 72 82 92 102 112
35 25 34 44 53 63 73 83 93 103 113
36 25 35 44 54 64 74 84 94 105 115
37 26 35 45 55 65 75 85 95 106 117 where q(cy,k, m) = upper (Y percentile point for the range of k inde-
38 26 36 45 55 66 76 86 97 107 118 pendent N(O,l) variables as given in Table A.10 (Hollander and Wolfe,
39 26 36 46 56 66 77 87 98 109 120 1973), it is possible to expand the scope of the available tables to
40 27 36 47 57 67 78 88 99 110 121 accomodate the range of panelists and samples usually encountered
41 27 37 47 57 68 79 90 100 112 123 in sensory evaluations.
42 27 37 48 58 69 80 91 102 113 124
43 28 38 48 59 70 81 92 103 114 126
44 28 38 49 60 70 82 93 104 115 127
45 28 39 49 60 71 82 94 105 117 128
46 28 39 50 61 72 83 95 106 118 130 RESULTS & DISCUSSION
29 39 50 62 73 108 119 131
ii 29 40 51 62 74 :z :; 109 121 133 TABLES 1 AND 2 presenttheseexpandedtablesto enable“all
49 110 122 134 treatment” multiple comparisons to be performed for sensory
50 fX t: 2 ii ;z x; ii 111 123 135
55 43 54 79 91 104 116 129 142 evaluations in which up to 100 panelists rank up to 12 samples,
:: 45 57 Ei: 82 95 108 121 135 148 with critical values available at (approximately) the 5% and 1%
:: 34 46 59 72 86 99 113 126 140 154 levels of significance, respectively. Because the underlying dis-
70 35 48 61 75 89 103 117 131 146 160 tribution of the range of rank sums is discrete, exact (Y% signif-
75 92 106 121 136 151 166
:; zv :i 8': 110 125 140 156 171 icance levels cannot be obtained, with all critical values in Tables
i! 38 53 68 83 iii 113 129 144 160 176 1 and 2 given to at least the ol% significance level. These ap-
40 54 70 85 101 116 132 149 165 181 proximate critical values have been shown (Dunn-Rankin and
41 56 71 87 103 120 136 153 169 186 Wilcoxon, 1966) to be accurate indicators of the range values
100 42 57 73 89 106 123 140 157 174 191
necessary to reach significance at the specified a% level of sig-
‘Exact values adapted from Hollander and Wolfe (1973) are used for up to 15 panel-
ists.
nificance, with only a minimal number of cases in which the
blnterpolation may be used for unspecified table values involving more than 50 panel- exact and approximate critical values differed. These cases in-
ists. volved small numbers of panelists (less than lo), with the re-
spective critical values differing by at most 1, further emphasizing
the reliability of the large sample approximation.
figurations, while 25 panelists ranking 5 samples would involve enum- As an example of this “all treatments” multiple comparison
erating (5!)= = 9.5 x 1051 possible rank configurations. Even with procedure, Table 3 presents the results of a sensory evaluation
modern computing facilities, these complete enumerations are com- in which 20 panelists ranked four products, with the corre-
putationally too difficult (and time-consuming) to be performed! As sponding six absolute rank sum differences being:
such, the presently available tables for these “all treatment” com-
parisons are based on either complete enumeration (for the more easily
enumerated cases) or mathematical approximations to the tail proba- RA-RB = 28
bilities in the appropriate distribution of the range of rank sums (for RA-Rc = 58
the more difficultly enumerated cases). These presently available ta- RA-R,, = 30
bles only allow up to 15 panelists which is inadequate for the range RB-Rc = 30
of panelists usually involved in sensory evaluation experiments. How- Rs-RD = 2
ever, by using a large sample approximation developed by Nemenyi Rc-R,, = 28
(1963) and discussed by Hollander and Wolfe (1973), where two
samples “a” and “b” are significantly different if: From Tables 1 and 2, the critical absolute difference is 21 at
the 5% level and 26 at the 1% level of significance, with Table
JR,- RbJ L q(a,k, 00).[(nk(k+ 1)/12); 4 giving the summary rank sum analysis for these 4 samples.

17224OURNAL OF FOOD SCIENCE-Volume 52, No. 6, 1987


Table 5-Critical rank sum differences for one-sided “treatments versus Table 6-Critical rank sum differences for one-sided “treatments versus
control” comparisons at 5% level of significancea.b control” comoarisons at 1% level of sianificanceQ
Number of samples (including control) Number of samples (including control)
Panelists 3 4 5 6 7 8 g 10 11 I2 Panelists 3 4 5 6 7 8 9 10 II 12
8 8 11 14 18 20 24 27 31 34 37 8 11 14 18 22 26 30 33 38 42 45
9 9 12 15 19 22 25 28 33 39 9 12 15 19 24 27 32 35 40 44 48
10 9 13 16 20 23 27 30 34 ii 41 10 12 16 20 25 29 33 37 42 46 50
11 10 13 17 21 24 28 31 36 39 43 11 13 17 21 26 30 35 39 45 49 53
12 10 14 18 22 25 29 33 41 45 12 13 18 22 27 31 36 41 46 51 55
13 11 14 18 22 26 30 34 ii 43 46 13 14 18 23 28 32 38 42 48 53 57
11 15 41 44 48 I4 39 60
1: 11 15 :i ;I ;3 ii ii 42 46 50 15 12 :: ;z ii 3": 41 ?5 :x i9 62
16 12 16 20 25 29 34 38 43 47 51 16 15 20 26 31 36 42 47 54 59 64
17 12 16 21 26 29 35 39 45 49 53 17 16 21 26 32 37 43 48 55 60 66
18 12 17 21 26 30 36 40 46 50 55 18 16 21 27 33 38 45 50 57 82 68
19 13 17 22 27 31 37 41 47 52 56 19 17 22 28 34 39 46 51 58 64 69
20 I3 18 22 28 32 38 42 48 53 58 20 17 23 28 35 40 47 52 60 66 71
21 13 18 23 28 33 39 43 50 54 59 21 17 23 29 36 41 48 54 61 67 73
22 14 18 24 29 33 39 44 51 55 60 22 18 24 30 36 42 49 55 63 69 75
23 14 19 24 30 34 40 45 52 57 62 23 18 24 31 37 43 50 56 64 70 76
14 19 25 30 35 41 46 53 58 63 24 19 25 31 38 44 51 57 66 72 78
2 15 20 25 31 36 42 47 54 59 64 25 19 25 32 39 45 52 59 67 73 80
26 15 20 26 32 36 43 48 55 60 65 26 19 26 32 40 46 53 60 68 75 81
27 I5 20 26 32 37 44 49 56 67 27 20 26 33 40 47 54 61 69 76 83
28 15 21 27 33 38 44 50 57 fF3 68 28 20 27 34 41 47 55 62 71 77 84
29 16 21 27 33 38 45 51 58 64 69 29 20 27 34 42 48 56 63 72 79 86
30 16 21 27 34 39 46 51 59 65 70 30 21 27 35 43 49 57 64 73 80 87
31 I6 22 34 40 47 66 71 21 58 65 74 81 89
32 16 22 ;: 35 40 48 zf :: 67 73 3: 21 fi :i 2 z: 59 66 76 83 90
17 23 29 35 41 48 54 62 68 74 33 22 29 36 45 51 77 84 91
3: 17 23 29 36 41 49 55 63 69 75 34 22 29 37 45 52 :: ii 78 85 93
35 17 23 30 36 42 50 55 64 70 76 35 22 30 38 46 53 62 69 79 87 94
17 30 37 43 50 56 85 71 77 36 23 63 70 80 88 95
i; 18 i: 30 38 43 51 57 66 72 78 37 23 i8 :: :; zt 64 71 81 89 97
18 24 31 38 44 52 58 66 73 79 38 23 31 39 48 55 65 72 82 90 98
2: 18 24 31 39 44 52 59 67 74 80 39 23 31 40 48 56 65 73 83 91 99
40 18 25 32 39 45 53 59 68 75 81 40 24 32 40 49 57 66 74 84 92 100
41 19 25 32 39 45 54 60 69 76 82 41 24 32 41 50 57 67 75 85 94 102
42 19 25 32 40 46 54 61 70 76 83 42 24 32 41 50 58 68 76 87 95 103
43 19 33 40 47 55 61 77 43 25 33 42 51 59 104
44 19 ;t 33 41 47 56 62 ;: 78 ~~ 44 25 33 42 51 59 :i :i ii i; 105
45 19 26 33 41 48 56 63 72 79 86 45 25 34 42 52 60 70 78 90 98 107
46 20 27 42 48 57 84 73 80 46 25 34 43 53 61 71 79 91 99 108
47 20 27 it 42 49 57 64 74 81 ii 47 26 34 43 53 61 72 80 91 100 109
48 35 43 49 58 65 75 82 89 48 35 44 54 62 72 81 92 101 110
49 ;: ;; 35 43 50 59 90 49 ;: 35 44 54 63 102 Ill
50 20 28 35 44 50 59 Fit 5: :3 91 50 26 35 45 55 63 :: ii :: 103 112
55 21 29 37 46 53 62 69 80 87 95 55 28 37 47 57 66 78 87 99 108 118
22 30 39 48 55 65 72 83 91 99 80 29 39 49 60 69 81 90 103 113 123
:"5 23 31 40 50 57 68 75 87 95 103 65 30 40 51 62 72 84 94 108 118 128
70 24 33 42 51 59 70 78 90 99 107 70 31 42 53 65 75 87 98 112 122 133
75 25 34 43 53 61 72 81 93 102 111 75 32 43 55 67 77 90 101 115 126 137
80 35 44 55 83 75 84 96 105 115 80 33 45 56 69 80 93 104 119 131 142
85 ;; 36 46 57 65 77 86 99 109 118 46 58 71 82 108 123 135 146
90 27 37 47 58 67 79 89 102 112 121 8: iz 47 60 73 85 ii 111 126 138 150
95 28 38 48 60 69 82 91 105 115 125 95 36 49 62 75 87 102 114 130 142 155
100 29 39 50 61 71 84 93 108 118 128 100 37 50 63 77 89 104 117 133 146 159
aCritical values for less than 8 panelists are not provided because of possible inac- 0 Critical values for less than 8 panelists are not provided because of possible inac-
curacies with large sample approximation. curacies with large sample approximation.
blnterpolation may be used for unspecified table values involving more than 50 panel- b Interpolation may be used for unspecified table values involving more than 50
ists. panelists.

“Treatments versus control” comparisons four samples (up to 5 panelists). Again, these tables are very
To test whether a specified sample (called the “control”) is limiting for use in sensory evaluation experiments. However,
superior to any of the remaining samples, Joanes (1985) pro- by using a large sample approximation developed by Nemenyi
posed the use of a distribution-free test based on Friedman’s (1963) and discussed by Hollander and Wolfe (1973), where
rank sums (Hollander and Wolfe, 1973) involving the distri- the control is significantly better than sample “a” if:
bution of the largest range between rank sums containing the
control. For this one-sided test, the control sample is consid- R,-R control2 (m(cw,k- 1,1/2). (mk(k+ 1)/b): (4)
ered to be significantly better than sample “a” if:
where m(a,k- 1,*/2) = upper (Y percentile point of the max-
R,- Rcontrol2 r* (o,k- Ln) imum of (k- 1) N(O,l) variables with common correlation r/2
as given in Table A.13 (Hollander and Wolfe, 1973), it is
where r * (cw,k- 1,n) = table value such that possible to expand the scope of the available tables to accom-
P Pa - kmtrol < r *(ol,k- 1,n)) = 1 -CL modate the usual range of panelists and samples encountered
in sensory evaluations. Table 5 and 6 present these expanded
Tables of critical values for this test are available (Hollander tables to enable one-sided “treatments versus control” multi-
and Wolfe, 1973) but are very limited becauseof the difficul- ple comparisons to be performed for sensory evaluations in
ties associatedwith the complete enumeration of the rank con- which up to 100 panelists rank up to 11 samples against the
figurations discussedin the previous section, resulting in tables control sample, with critical values available at (approxi-
only being available for three samples (up to 18 panelists) and mately) the 5% and 1% levels of signficance, respectively.

Volume 52, No. 6, 1987-JOURNAL OF FOOD SCIENCE- 1723


ANALYSIS OF RANKED DATA. . .
Table 7-Critical rank sum differences for two-sided “treatments versus Table 8-Critical rank sum differences for two-sided “treatments versus
control” comparisons at 5% level of significancea,b control” comparisons at 1% level of significanceaeb
Number of samples (including control) Number of samples (including control)
Panelists 3 4 5 6 7 8 9 10 11 12 Panelists 3 4 5 6 7 8 9 10 11 12
8 9 13 16 19 23 26 30 33 37 40 8 12 I6 19 23 27 31 35 39 44 48
9 10 13 I7 20 24 28 31 35 39 42 9 12 16 21 25 29 33 38 42 46 51
10 10 14 18 21 25 29 33 37 41 45 10 I3 17 22 26 31 35 40 44 49 53
11 11 15 19 23 27 30 35 39 43 47 11 14 18 23 27 32 37 41 46 51 56
12 11 15 19 24 28 32 36 40 45 49 12 14 19 24 29 33 38 43 48 53 58
13 12 16 20 24 29 33 38 42 46 51 13 15 20 25 30 35 40 45 50 55 61
14 12 17 21 25 30 34 39 44 48 53 14 I5 20 26 31 36 41 47 52 58 63
15 13 17 22 26 31 36 40 45 50 55 15 16 21 26 32 37 43 48 54 60 65
16 13 18 22 27 32 37 42 47 16 16 22 27 33 39 44 50 58 61 67
17 13 18 23 28 33 38 43 48 z3 z: 17 17 22 28 34 40 45 E 69
18 14 19 24 29 34 39 44 49 55 60 18 17 23 29 35 41 47 71
19 14 19 24 29 35 40 45 51 19 18 24 30 36 42 48 54 61 67 73
20 14 20 25 30 36 41 46 52 Ei :; 20 18 24 30 37 43 49 56 62 69 75
21 15 20 26 31 36 42 48 53 59 65 21 19 25 31 38 44 51 57 64 70 77
22 15 21 26 32 37 43 49 55 60 66 22 19 26 32 38 45 52 58 65 72 79
23 I5 21 27 32 38 44 50 56 62 68 23 19 26 33 39 46 53 60 67 74 80
24 16 22 27 33 39 45 51 57 63 69 24 20 27 33 40 47 54 61 68 75 82
25 16 22 28 34 40 46 52 58 64 70 25 20 27 34 41 48 55 62 69 77 84
26 16 22 28 34 41 47 53 59 66 72 26 21 28 35 42 49 56 63 71 78 86
27 17 23 29 35 41 47 54 60 67 73 27 21 28 35 43 50 57 65 72 80 87
28 17 23 29 36 42 48 55 61 68 74 28 21 29 36 43 51 58 86 73 81 89
29 17 24 30 38 43 49 56 63 69 76 22 29 37 44 52 59 67 75 83 90
30 18 24 30 37 44 50 57 64 70 77 Ei 22 30 37 45 53 60 68 76 84 92
31 18 24 31 37 44 51 58 65 72 78 31 22 30 38 46 53 61 69 77 85 93
32 18 25 31 38 45 52 59 66 73 80 32 23 31 38 46 54 62 70 78 87 95
33 18 25 32 39 46 52 59 74 81 33 23 31 39 47 55 63 71 80 88 96
34 19 26 32 39 46 53 60 Eli 75 34 24 32 40 48 56 64 73 81 89 98
35 19 26 33 40 47 54 61 69 76 :: 35 24 32 40 48 57 65 74 82 91 99
36 19 26 33 40 48 55 62 70 77 84 36 24 32 41 49 58 66 75 83 92 101
37 20 27 34 41 48 55 63 71 78 85 37 24 33 41 50 58 67 76 84 93 102
38 20 27 34 41 49 56 64 72 79 87 38 25 33 42 50 59 88 77 85 94 103
39 20 27 35 42 50 57 65 72 80 88 39 25 34 42 51 60 69 78 87 96 105
40 20 28 35 43 50 58 65 73 81 89 40 25 34 43 52 61 70 79 88 97 106
41 21 28 35 43 51 58 66 74 82 90 41 26 35 43 52 61 70 80 89 98 107
42 21 28 36 44 51 59 67 75 83 91 42 26 35 44 53 62 71 81 90 99 109
43 21 29 36 44 52 60 68 76 84 92 43 26 35 44 54 63 72 82 91 100 110
44 21 29 37 45 53 60 69 77 85 93 44 27 36 45 54 64 73 82 92 102 111
45 21 29 37 45 53 61 69 78 86 94 45 27 36 45 55 64 74 83 93 103 112
46 22 30 38 46 54 62 70 79 87 95 46 27 37 46 55 65 75 84 94 104 114
47 22 30 38 46 54 62 71 79 88 96 47 28 37 46 56 66 75 85 95 105 115
48 38 47 55 63 72 80 48 28 37 47 57 66 76 86 96 106 116
49 39 47 55 64 72 81 ii El 49 28 38 47 57 67 77 87 97 107 117
50 23 31 39 47 56 64 73 82 91 99 50 28 38 48 58 68 78 88 98 108 118
55 24 32 41 50 59 68 77 86 95 104 55 30 40 50 61 71 81 92 103 114 124
60 25 34 43 52 61 71 80 90 99 109 60 31 42 52 63 74 85 96 107 119 130
65 28 35 44 54 64 73 83 93 103 113 65 32 43 55 66 77 88 100 112 123 135
27 36 46 56 66 76 97 107 117 104 116 128 140
;t 28 38 48 58 68 79 100 111 121 5: ;z 1; E!z ;: ii El 107 120 133 145
80 28 39 49 60 71 81 92 104 115 125 80 36 48 60 73 85 98 111 124 137 150
85 29 40 51 62 73 84 95 107 118 129 85 37 50 62 75 88 101 114 128 141 154
90 30 41 52 63 75 86 98 110 122 133 90 38 51 64 77 91 104 II8 131 145 159
95 31 42 54 65 77 89 101 113 125 137 95 39 52 66 79 93 107 121 135 149 163
100 32 43 55 67 79 91 103 116 128 140 100 40 54 68 81 96 110 124 138 153 167
a Critical values for less than 8 panelists are not provided because of possible inac- a -Critical values for less than 8 panelists are not provided because of possible
curacies with large sample approximation. inaccuracies with large sample approximation.
b Interpolation may be used for unspecified table values involving more than 50 b-Interpolation may be used for unspecified table values involving more than 50
wnelists. Danelists.

Similarly, to test whether the control sample is inferior to However, by using a large sample approximation developed
any of the remaining samples, R, - Rcontrolshould be replaced by Nemenyi (1963) and discussed by Hollander and Wolfe
by Rcontrol - R, in the above procedure. (1973), where the control is significantly different to sample
Alternatively, to test whether the control is different from “a” if:
any of the remaining samples, a two-sided test is available
(Hollander and Wolfe, 1973) involving the distribution of the
largest absolute range between rank sums containing the con- IRa - Rcontrol I L Irnl (o,k-l,%) . (nk(k+1)/6)$ (6)
trol. In this test, the control sample is considered to be signif-
where Iml(ol,k- 1,1/2)= upper OLpercentile point of maximum
icantly different to sample “a” if: absolute value of (k- l)N(O,l) variables with common cor-
1% - Lntroll 2 r** k-4 - l,n) relation 1/2as given in Table A.14 (Hollander and Wolfe, 1973),
(5) it is possible to expand the scope of the available tables to
where r** (a,k- 1,n) = table value such that accomodate the usual range of panelists and samples encoun-
tered in sensory evaluation. Table 7 and 8 present these ex-
P(IR,-R,,,,,,r 1 < r** (a,k-1,n)) = 1-o. panded tables to enable two-sided “treatments versus control”
multiple comparisons to be performed for sensory evaluations
Tables of critical values (Hollander and Wolfe, 1973) only in which up to 100 panelists rank up to 11 samples against the
cover the same range of panelists and samples available for control sample, with critical values available at (approxi-
the previously discussed one-sided tests and, as such, are too mately) the 5% and 1% levels of significance respectively.
restrictive for effective use in sensory evaluation experiments. As an example of these “treatments versus control” mul-

1724-JOURNAL OF FOOD SCIENCE-Volume 52, No. 6, 1987


tiple comparison procedures, the data from Table 3 can be cumstances of the usual range of panelists and samples en-
reanalyzed, with sample B treated as the control sample and countered in sensory evaluation experimentation. A BASIC
the corresponding three rank sum differences involving the computer program for the analysis of the “all treatments” and
control being: “treatments versus control” multiple comparisons procedures
is available from the authors upon request.
RA- Rcontrol = - 28
Rc-Kontro~ = 30
RD -Rcontro~ = 2. REFERENCES
Amerine, M., Pangborn, R., and Roes&r, E. 1965. “Principles of Sensory
To test whether the control is superior to any of the remaining Evaluation of Food.” Academic Press, New York.
samples, Tables 5 and 6 give ttie critical rank sum differences Bradley, R and Kramer, A. 1957. Addenda to a quick, rank test for sig-
nificance of differences in multiple comparisons. Food Technol. 11: 412.
to be 18 at the 5% level and 23 at the 1% level of significance, Dunn-Rankin, P. and Wilcoxon, F. 1966. The true distributions of the
thus indicating that the control is only superior to sample C at range of rank totals in the two-way classification. Psychometrika 31:
573.
the 1% level of significance. Similarly, to test whether the Hollander, M. and Wolfe, D. 1973. “Non-parametric Statistical Methods.”
control is inferior to any of the remaining samples, these crit- J. Wiley, New York.
ical values indicate that the control is only inferior to sample Joanes, D. 1985. On a rank sum test due to Kramer. J. Food Sci. 50: 1442.
Kahan, G., Cooper! D., Papavasilioa, A., and Kramer, A. 1973. Expanded
A at the 1% level of significance. tables for deternunins sienificant differences for ranked data. Food Tech-
- L

To test whether the control is different from any of the nol. 27: 61.
Kramer, A. 1956. A quick, rank test for significance of differences in mul-
remaining samples, Tables 7 and 8 give the critical absolute tiple comparisons. Food Technol. 10: 391.
rank sum differences to be 20 at the 5% level and 24 at the Kramer, A. 1960. A rapid method for determining significance of differ-
ences from rank sums. Food Technol. 14: 576.
1% level of significance, thus indicating that the control is Kramer, A. 1963. Revised tables for determining significance of differ-
significantly different from samples A and C at the 1% level ences. Food Technol. 17: 124.
of level of significance. Kramer, A. and Twigg, B. 1970. “Quality Control for the Food Industry.”
AVI Publishing Company, Westport, CT.
McDonald, B. and Thompson., W. 1967. Rank sum multiple comparisons
in one- and two-way classifications. Biometrika 54: 467.
McLellan, M., Way, R., and Lamb, C. 1984. An interactive computerized
CONCLUSION method for rank analysis. Hort. Sci. 19: 634.
Nemenyi, P. 1963. Distribution-free multiple comparisons. Unpublished
THE USE of the expanded tables of critical values presented Ph.D. dissertation, Princeton University.
MS received l/20/87; revised 3130187; accepted 4127187.
in this paper for multiple comparisons using ranked data have
been shown to be more complete and appropriate to the cir-

CHARACTER/ZAT/ON OF CAfffhVE-RESISTANT A. PARASITICUS. . .From page 1720

propriate amino acid source were available could sufficient of a caffeine-resistant mutant of Aspergillus parasiticus. J. Food Sci. 52:
194.
levels of precursor accumulation occur such that a significant Davis, N. D., Diener, U. L., and Eldridge, D. W. 1966. Production of af-
portion was shunted into aflatoxin production. Additional stud- latoxins B1 and G1 by Aspergillus flauus in a semisynthetic medium.
Apt. Microbial. 14: 378.
ies with BCRl that will be reported separately have indicated Dura ovlc, S.,, Durakovlc, Z., Beritic, T., Radic, B., Lalic, L. M., and Delas,
that caffeine does not directly affect aflatoxin synthesis, but F. 1985a. Biosynthesis of aflatoxins by Aspergillusparasiticus on roasted
instead acts by altering some aspect of primary metabolism. coffee beans. Periodicum Biologorum 87: 503.
Durakovic. S.. Durakovic. Z.. Lalic. L. M.. Posnisil. 0.. and Radic. B. 1985b.
However, additional studies will be needed to determine if the Influence of selected cultiiation para&t&s on thk growth df the toxi-
caffeine-dependent, amino acid-dependent nature of aflatoxin genie mold Aspergillus parasiticus on coffee beans and the biosynthesis
of aflatoxins. Microbioloev (Belmade) 22: 1.
synthesis in caffeine-resistant A. parasiticus involves a mech- Lenovich, L. M. 1981. Eff& of caffeine on aflatoxin production in cocoa
anism similar to that hypothesized above. beans. J. Food Sci. 46: 655.
Naik. M.. Modi. V. V.. and Patel. N. C. 1970. Studies on aflatoxin biosvn-
thesis in Asp&gill& flauw. I&an J. Exp. Biol. 8: 345.
Nartowicz, V. B., Buchanan, R. L., and Segall, S. 1979. Aflatoxin produc-
tion in regular and decaffeinated coffee beans. J. Food Sci. 44: 446.
REFERENCES Payne, G. A. and Hagler, W. M. Jr. 1983. Effect of specific amino acids on
growth and aflatoxin production by Aspergihs parasiticus and Asper-
Betancourt, L. E. and Frank, H. K. 1983. Bedingungen des mikrobiellen gillus fZauus. Appl. Environ. Microbial. 46: 805.
verderbs van grunem kaffee. Deut. Lebn.-Rund. 79: 404. Reddy, T. V., Viswanathan, L., and Venkitasubramian, T. A. 1979. Factors
Buchanan, R. L. and Fletcher, A. M. 1978. Methylxanthine inhibition of affecting aflatoxin production by Aspergillus parasrticus in chemically
aflatoxin production. J. Food Sci. 43: 654. defined medium. J. Gen. Microbial. 114: 409.
Buchanan, R. L., Hoover, D. G., and Jones, S. B. 1983. Caffeine inhibition Sahay, M. 1983. Effect of nitrogen sources on growth and aflatoxin pro-
of aflatoxin production: Mode of action. Appl. Environ. Microbial. 46: duction by Aspergillus arasticus (NRRL 3240). In “Proc. Symp. Myco-
1193. toxin in Food and Feed%, K S Bllerami. ed. Allied Press, Bhaealour.
Buchanan, R. L. and Houston, W. M. 1982. Production of blue-fluorescent India. p. 199.
pyrazines by Aspergillus parasiticus. J. Food Sci. 47: 779. MS received 2/10/87; revised 5/2/87; accepted 5/4/87.
Buchana?, R. L. and Lewis, D. F. 1984a. Caffeine inhibition of aflatoxin
synthesis: Probable site of action. Appl. Environ. Microbial. 47: 1216.
Buchanan, R. L. and Lewis, D. F. 1984b. Regulation of aflatoxin biosyn-
thesis: Effect of glucose on the activities of various glycolytic enzymes.
A pl. Environ. Microbial. 48: 306. Presented at the 47th Annual Meeting of the Institute of Food Technologists, Las
But K anan, R. L., Zaika, L. L., Stahl, H. G., and Mertz, S. E. 1987. Isolation Vegas, NV, June 1619, 1987.

Volume 52, No. 6, 1987-JOURNAL OF FOOD SCIENCE- 7725

You might also like