T. P. Fernandes, N. A. Santos, and G. V. Paramei, "Cambridge Colour Test: reproducibility in normal trichromats," J. Opt. Soc. Am. A 37, A70-A80 (2020)
This study evaluated reproducibility of the Trivector subtest of the Cambridge Colour Test. Data for normal trichromats were obtained in Brazil ($ N = 111 $) at T0, six months (T1), and 12 months later (T2), and in the United Kingdom (${ N}={79}$), with the test directly followed by a retest. Coefficients of repeatability—Bland–Altman indices—for Protan, Deutan, and Tritan vectors were similar for both datasets. Intraclass correlation coefficients (ICCs)—measures of reliability—were low or moderate for these relatively homogeneous datasets; for a heterogeneous dataset, comprising color-normal and abnormal observers, ICCs were 0.80–0.98, indicating the high discriminative accuracy of the Trivector subtest.
You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
You do not have subscription access to this journal. Figure files are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
You do not have subscription access to this journal. Article tables are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
You do not have subscription access to this journal. Equations are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
Descriptive Statistics of Trivector Measures ( Units) for the Brazil and UK Samples of Normal Trichromats (NTs) and the UK Extended Sample Including Color-Vision Deficient Observers (CVDs)a
Brazil () NTs
UK () NTs
UK () NTs & CVDs
Test sess.
T0
T1
T2
Test
Retest
Test
Retest
Protan
M
48.6
48.8
50.5
50.3
52.7
80.1
77.3
SD
11.4
12.0
120.
14.8
17.2
131.9
120.1
Var
130
144
141
219
296
17396
14420
CV
0.24
0.25
0.23
0.29
0.33
1.64
1.55
Med
48
46
49
49
51
54
53
sIQR
7
7
8.5
9.5
9.5
11.5
13.5
Skw
0.71
1.39
0.27
0.93
1.10
5.84
6.53
Krt
1.23
3.42
0.60
2.40
1.62
35.86
48.17
Deutan
M
50
53.5
54.7
50.8
52.7
75.7
70.4
SD
13.3
13.5
15.1
14.3
17.6
103.4
68.3
Var
178
180
228
204
310
10703
4662
CV
0.27
0.25
0.28
0.28
0.33
1.36
0.97
Med
48
51
53
50
49
53
54
sIQR
6.5
9.5
8.5
9.5
10.5
12
13.5
Skw
1.30
0.92
1.59
0.64
0.92
5.13
3.97
Krt
3.29
1.74
6.92
0.67
0.84
27.06
16.46
Tritan
M
58.1
60.2
62.6
72
67.7
94.3
86.1
SD
20.5
17.3
21.9
24.6
22.4
54.7
46.5
Var
420
299
479
607
502
2992
2169
CV
0.35
0.29
0.35
0.34
0.33
0.61
0.54
Med
56
58
60
66
65
77
73
sIQR
16
12
12
17
12
28
24
Skw
0.32
0.39
1.18
0.68
0.95
2.39
1.94
Krt
0.84
0.35
1.80
0.21
0.76
8.68
5.41
Mean (M), standard deviation (SD), variance (Var), coefficient of variation (CV), median (Med), semi-interquartile range (sIQR), skewness (Skw), and kurtosis (Krt).
Table 2.
Bland–Altman Parameters of the Trivector Test–Retest Measures ( Units) for Samples of NTs in Brazil and the UKa
Brazil ()
UK ()
Vector
Statistic
T0-T1
T1-T2
T0-T2
Test–Retest
Protan
()
48.71 (8.96)
49.66 (8.63)
49.56 (8.90)
51.53 (13.27)
()
−0.20 (15.10)
−1.70 (16.54)
−1.90 (15.00)
−2.38 (18.01)
95% CI
(−3.04)–(2.64)
(−4.82)–(1.40)
(−4.72)–(0.92)
(−6.41)–(1.65)
Upper LoA
29.40
30.72
27.50
32.92
95% CI Upper LoA
23.57–35.24
24.33–37.10
21.71–33.30
25.93–39.90
Lower LoA
−29.80
−34.12
−31.31
−37.68
95% CI Lower LoA
(−35.63)–(−23.97)
(−40.51)–(−27.73)
(−37.10)–(−25.51)
(−44.66)–(−30.69)
COR
29.60
32.42
29.40
35.30
Deutan
()
51.76 (10.23)
54.08 (10.87)
52.33 (10.01)
51.72 (13.57)
()
−3.46 (17.35)
−1.18 (18.64)
−4.64 (20.27)
−1.86 (17.15)
95% CI
(−6.72)–(−0.19)
(−4.68)–(2.32)
(−8.45)–(−0.82)
(−5.70)–(1.98)
Upper LoA
30.54
35.35
35.09
31.76
95% CI Upper LoA
23.84–37.24
28.15–42.54
27.26–42.92
25.11–38.41
Lower LoA
−37.46
−37.71
−44.37
−35.48
95% CI Lower LoA
(−44.16)–(−30.76)
(−44.90)–(−30.51)
(−52.20)–(−36.54)
(−42.13)–(−28.83)
COR
34.00
36.53
39.73
33.62
Tritan
()
59.15 (14.25)
61.39 (14.43)
60.34 (15.78)
69.86 (20.33)
()
−2.09 (25.06)
−2.39 (26.94)
−4.48 (28.34)
4.38 (23.82)
95% CI
(−6.81)–(2.62)
(−7.45)–(2.67)
(−9.80)–(0.85)
(−0.95)–(9.72)
Upper LoA
47.03
50.41
51.06
51.07
95% CI Upper LoA
37.35–56.71
40.01–60.82
40.12–62.01
41.83–60.31
Lower LoA
−51.21
−55.19
−60.02
−42.31
95% CI Lower LoA
(−60.89)–(−41.53)
(−65.59)–(−44.78)
(−70.96)–(−49.08)
(−51.55)–(−33.07)
COR
49.12
52.80
55.55
46.69
Note: ${\bar X_V}$, mean of the vector (${V}$) test–retest; ${\textit{SD}_V}$, standard deviation of ${\bar X_V}$; ${\bar X_D}$, mean of test–retest difference; ${\textit{SD}_D}$, standard deviation of ${ D}$; LoA, limit of agreement; CI, confidence interval; COR, coefficient of repeatability (highlighted in bold).
Table 3.
Trivector Test–Retest Intraclass Correlation Coefficients (ICCs) and Their 95% Confidence Intervals (in Square Brackets) for the Brazil and UK Samples of Normal Trichromats (NTs) and for the UK Extended Sample Including Color-Vision Deficient Observers (CVDs)
Brazil () NTs
UK () NTs
UK () NTs & CVDs
Vector
T0-T1
T1-T2
T0-T2
Test–Retest
Test–Retest
Protan
0.29 [(−0.03)−(0.51)]
0.08 [(−0.33)−(0.36)]
0.29 [(−0.03)–(0.51)]
0.54 [(0.28)–(0.70)]
0.98 [(0.97)–(0.99)]
(T0-T1, T0-T2, T1-T2) = 0.30 [(0.04)–(0.50)]
Deutan
0.28 [(−0.04)−(0.50)]
0.26 [(−0.07)−(0.49)]
−0.02 [(−0.47)−(0.28)]
0.60 [(0.38)–(0.74)]
0.90 [(0.87)–(0.93)]
(T0-T1, T0-T2, T1-T2) = 0.24 [(−0.03)–(0.45)]
Tritan
0.23 [(−0.12)−(0.46)]
0.13 [(−0.27)−(0.40)]
0.19 [(−0.17)−(0.44)]
0.65 [(0.46)–(0.78)]
0.80 [(0.71)–(0.86)]
(T0-T1, T0-T2, T1-T2) = 0.25 [(−0.02)–(0.46)]
Tables (3)
Table 1.
Descriptive Statistics of Trivector Measures ( Units) for the Brazil and UK Samples of Normal Trichromats (NTs) and the UK Extended Sample Including Color-Vision Deficient Observers (CVDs)a
Brazil () NTs
UK () NTs
UK () NTs & CVDs
Test sess.
T0
T1
T2
Test
Retest
Test
Retest
Protan
M
48.6
48.8
50.5
50.3
52.7
80.1
77.3
SD
11.4
12.0
120.
14.8
17.2
131.9
120.1
Var
130
144
141
219
296
17396
14420
CV
0.24
0.25
0.23
0.29
0.33
1.64
1.55
Med
48
46
49
49
51
54
53
sIQR
7
7
8.5
9.5
9.5
11.5
13.5
Skw
0.71
1.39
0.27
0.93
1.10
5.84
6.53
Krt
1.23
3.42
0.60
2.40
1.62
35.86
48.17
Deutan
M
50
53.5
54.7
50.8
52.7
75.7
70.4
SD
13.3
13.5
15.1
14.3
17.6
103.4
68.3
Var
178
180
228
204
310
10703
4662
CV
0.27
0.25
0.28
0.28
0.33
1.36
0.97
Med
48
51
53
50
49
53
54
sIQR
6.5
9.5
8.5
9.5
10.5
12
13.5
Skw
1.30
0.92
1.59
0.64
0.92
5.13
3.97
Krt
3.29
1.74
6.92
0.67
0.84
27.06
16.46
Tritan
M
58.1
60.2
62.6
72
67.7
94.3
86.1
SD
20.5
17.3
21.9
24.6
22.4
54.7
46.5
Var
420
299
479
607
502
2992
2169
CV
0.35
0.29
0.35
0.34
0.33
0.61
0.54
Med
56
58
60
66
65
77
73
sIQR
16
12
12
17
12
28
24
Skw
0.32
0.39
1.18
0.68
0.95
2.39
1.94
Krt
0.84
0.35
1.80
0.21
0.76
8.68
5.41
Mean (M), standard deviation (SD), variance (Var), coefficient of variation (CV), median (Med), semi-interquartile range (sIQR), skewness (Skw), and kurtosis (Krt).
Table 2.
Bland–Altman Parameters of the Trivector Test–Retest Measures ( Units) for Samples of NTs in Brazil and the UKa
Brazil ()
UK ()
Vector
Statistic
T0-T1
T1-T2
T0-T2
Test–Retest
Protan
()
48.71 (8.96)
49.66 (8.63)
49.56 (8.90)
51.53 (13.27)
()
−0.20 (15.10)
−1.70 (16.54)
−1.90 (15.00)
−2.38 (18.01)
95% CI
(−3.04)–(2.64)
(−4.82)–(1.40)
(−4.72)–(0.92)
(−6.41)–(1.65)
Upper LoA
29.40
30.72
27.50
32.92
95% CI Upper LoA
23.57–35.24
24.33–37.10
21.71–33.30
25.93–39.90
Lower LoA
−29.80
−34.12
−31.31
−37.68
95% CI Lower LoA
(−35.63)–(−23.97)
(−40.51)–(−27.73)
(−37.10)–(−25.51)
(−44.66)–(−30.69)
COR
29.60
32.42
29.40
35.30
Deutan
()
51.76 (10.23)
54.08 (10.87)
52.33 (10.01)
51.72 (13.57)
()
−3.46 (17.35)
−1.18 (18.64)
−4.64 (20.27)
−1.86 (17.15)
95% CI
(−6.72)–(−0.19)
(−4.68)–(2.32)
(−8.45)–(−0.82)
(−5.70)–(1.98)
Upper LoA
30.54
35.35
35.09
31.76
95% CI Upper LoA
23.84–37.24
28.15–42.54
27.26–42.92
25.11–38.41
Lower LoA
−37.46
−37.71
−44.37
−35.48
95% CI Lower LoA
(−44.16)–(−30.76)
(−44.90)–(−30.51)
(−52.20)–(−36.54)
(−42.13)–(−28.83)
COR
34.00
36.53
39.73
33.62
Tritan
()
59.15 (14.25)
61.39 (14.43)
60.34 (15.78)
69.86 (20.33)
()
−2.09 (25.06)
−2.39 (26.94)
−4.48 (28.34)
4.38 (23.82)
95% CI
(−6.81)–(2.62)
(−7.45)–(2.67)
(−9.80)–(0.85)
(−0.95)–(9.72)
Upper LoA
47.03
50.41
51.06
51.07
95% CI Upper LoA
37.35–56.71
40.01–60.82
40.12–62.01
41.83–60.31
Lower LoA
−51.21
−55.19
−60.02
−42.31
95% CI Lower LoA
(−60.89)–(−41.53)
(−65.59)–(−44.78)
(−70.96)–(−49.08)
(−51.55)–(−33.07)
COR
49.12
52.80
55.55
46.69
Note: ${\bar X_V}$, mean of the vector (${V}$) test–retest; ${\textit{SD}_V}$, standard deviation of ${\bar X_V}$; ${\bar X_D}$, mean of test–retest difference; ${\textit{SD}_D}$, standard deviation of ${ D}$; LoA, limit of agreement; CI, confidence interval; COR, coefficient of repeatability (highlighted in bold).
Table 3.
Trivector Test–Retest Intraclass Correlation Coefficients (ICCs) and Their 95% Confidence Intervals (in Square Brackets) for the Brazil and UK Samples of Normal Trichromats (NTs) and for the UK Extended Sample Including Color-Vision Deficient Observers (CVDs)