Professional Documents
Culture Documents
Chapter 7
Chapter 7
7-1 The mean of 5 measurements, x is a better estimate of the true value µ than any single
individual results.
(b) CL = 90 %
(c) CL = 68 %
(d) CL = 95.4 %
7-3 (a) As the sample size, N, increases the confidence interval decreases in proportion to
N.
(b) As the desired confidence level increases the confidence interval increases.
(c) As the standard deviation, s, increases the confidence interval increases in direct
proportion.
3.5 6.25
3.1 9.61
3.1 9.61
3.3 10.89
2.5 12.25
2
48.61 − (15.5) / 5
standard deviation: s = = 0.37
5 −1
σ, we will use the t statistic for confidence intervals. From Table 7-3, at 95% confidence
CI for µ = 3.1 ±
(2.78)(0.37 ) = 3.1 ± 0.46 = 3.1 ± 0.5
5
Results for Sets A through F, obtained in a similar way, are given in the following table.
A B C D E F
The 95% confidence interval establishes the range about the mean that we can be 95%
confident that the true value lies within (in the absence of systematic errors).
7-5 If s is a good estimate of σ then we can use z at 95% confidence equal 1.96 and for set A,
CI for µ = 3.1 ±
(1.96)(0.20) = ±0.18 Similarly for sets B-F
5
A B C D E F
2 .5 − 3 .1
7-6 For set A: Q = = 0.6 and Qcrit for 5 observations at 95% confidence = 0.710.
3 .5 − 2 .5
Since Q < Qcrit the outlier value 2.5 cannot be rejected with 95% confidence.
Fundamentals of Analytical Chemistry: 8th ed. Chapter 7
A B C D E F
1.28 × 2.4
(b) 80% CL = 18.5 ± = 18.5 ± 2.2 µg Fe/mL
2
1.96 × 2.4
95% CL = 18.5 ± = 18.5 ± 3.3 µg Fe/mL
2
1.28 × 2.4
(c) 80% CL = 18.5 ± = 18.5 ± 1.5 µg Fe/mL
4
1.96 × 2.4
95% CL = 18.5 ± = 18.5 ± 2.4 µg Fe/mL
4
1.64 × 0.32
(b) 90% CL = 8.53 ± = 8.53 ± 0.26 µg Cu/mL
4
2.58 × 0.32
99% CL = 8.53 ± = 8.53 ± 0.41 µg Cu/mL
4
1.64 × 0.32
(c) 90% CL = 8.53 ± = 8.53 ± 0.13 µg Cu/mL
16
2.58 × 0.32
99% CL = 8.53 ± = 8.53 ± 0.21 µg Cu/mL
16
Fundamentals of Analytical Chemistry: 8th ed. Chapter 7
1.96 × 2.4
7-9 (a) 1.5 = For a 95% CL, N = 9.8 ≅ 10
N
2.58 × 2.4
(b) 1.5 = For a 99% CL, N = 17
N
1.96 × 0.32
7-10 (a) 0.2 = For a 95% CL, N = 9.8 ≅ 10
N
2.58 × 0.32
(b) 0.2 = For a 99% CL, N = 17
N
4.30 × 0.06
(a) 95% CL = 3.22 ± = 3.22 ± 0.15 meq Ca/L
3
1.96 × 0.056
(b) 95% CL = 3.22 ± = 3.22 ± 0.06 meq Ca/L
3
2.92 × 0.25
(a) 90% CL = 7.24 ± = 7.24 ± 0.42 % lindane
3
1.64 × 0.28
(b) 90% CL = 7.24 ± = 7.24 ± 0.27 % lindane
3
2.58 × 0.40
7-13 (a) 0.3 = For a 99% CL, N = 11.8 ≅ 12
N
1.96 × 0.40
(b) 0.3 = For a 95% CL, N = 6.8 ≅ 7
N
1.64 × 0.40
(c) 0.2 = For a 90% CL, N = 10.8 ≅ 11
N
Fundamentals of Analytical Chemistry: 8th ed. Chapter 7
7-14 The null hypothesis is that µ = 30.15% and the alternative hypothesis is that µ ≠ 30.15%.
This is a two-tailed test and from Table 7-1, zcrit = 1.96 for the 95% confidence level.
30.26 − 30.15
(a) z = = 2.34
0.094 / 4
(b) From Table 7-3, t for 3 degrees of freedom = 3.18 for the 95% confidence level.
30.26 − 30.15
t= = 2.59
0.085 / 4
7-15 This is a two-tailed test and from Table 7-1, zcrit = 1.96 for the 95% confidence level.
68.5 − 68.8
(a) z = = −1.54
0.004 × 68.8% / 2
Since –1.54 ≥ -1.96, systematic error is NOT indicated at 95% confidence for C.
4.882 − 4.953
(b) z = = −3.38
0.006 × 4.953% / 2
7-16 This is a two-tailed test where s → σ and from Table 7-1, zcrit = 2.58 for the 99%
confidence level.
129 − 119
For As: z= = 1.28
3+3
9 .5
3× 3
Element z Difference?
Co -3.42 Yes
La 2.45 No
Sb 0.20 No
Th -3.42 Yes
7-17 H0: µ = 5.0 ppm dissolved O2; Ha: µ < 5.0 ppm dissolved O2. This is a one-tailed test and
4.888 − 5.0
t= = -0.79
0.40 / 8
Since –0.79 ≥ -1.90, we must accept H0 and we cannot state that the dissolved O2 is less
5 .6 − 5 .1
7-18 Q= = 0.385 and Qcrit for 8 observations at 95% confidence = 0.526.
5 .6 − 4 .3
Since Q < Qcrit the outlier value 5.6 cannot be rejected with 95% confidence.
7-19 H0: µ = 1.0 ppb: Ha: µ < 1.0 ppb. This would be a one-tailed test. The type I error for
this situation would be that we reject H0, when in fact, it is true, i.e. we decide the
impurity is < 1.0 ppb at some level of confidence when, in fact, it is not < 1.0 ppb. The
type II error would be that we accept H0 when, in fact, it is false, i.e. we decide the
impurity is not < 1.0 ppb at some level of confidence when, in fact, it is < 1.0 ppb.
Fundamentals of Analytical Chemistry: 8th ed. Chapter 7
7-20 H0: µ = the previous level; Ha: µ > the previous level. This would be a one-tailed test.
The type I error for this situation would be that we reject H0 when in fact, it is true, i.e.
we decide the level of the pollutant is > the previous level at some level of confidence
when, in fact, it is not. The type II error would be that we accept H0 when, in fact, it is
false, i.e. we decide the level of the pollutant = the previous level when, in fact, it is >
7-21 (a) H0: µ = 7.03 ppm; Ha: µ = < 7.03 ppm. This would be a one-tailed test. The type I
error for this situation would be that we reject H0 decide that a systematic error exists
when it does not. The type II error would be that we accept H0 decide a systematic error
This would be a two-tailed test. The type I error for this situation would be that we
decide the methods agree when they do not. The type II error would be that we decide
(c) H0: σ2AA = σ2EC; Ha: σ2AA < σ2EC. This would be a one-tailed test. The type I error
for this situation would be that we decide AA results are less precise than
electrochemistry results, when the precision is the same. The type II error would be that
we decide the precision is the same when they electrochemical results are more precise.
(d) H0: [impurity]X = [impurity]Y; Ha: [impurity]X < [impurity]Y. This would be a one-
tailed test. The type I error would be that we decide the [impurity]X < [impurity]Y when
it is not. The type II error would be that we decide the [impurity]X = [impurity]Y when it
does not.
Fundamentals of Analytical Chemistry: 8th ed. Chapter 7
spooled = 0.1199
degrees of freedom = 5 +5 – 2 = 8
26.338 − 26.254
t= = 1.11
5+5
0.1199
5× 5
0.084 − 0
t= = 12.52
0.015 / 5
Since 12.52 > 2.78, a significant difference does exist at 95% confidence.
(c) The large sample-to-sample variability causes sTop and sBottom to be large and masks
the differences between the samples taken from the top and the bottom.
7-23 (a) A paired t test should definitely be used in this case due to the large variation in the
Cl concentrations resulting from the various contact times and various locations from
(b) H0: µd = 0; Ha: µA ≠ 0, where µd is the mean difference between the methods.
0.414 − 0
t= = 3.65
0.32 / 8
Fundamentals of Analytical Chemistry: 8th ed. Chapter 7
Since 3.65 > 1.90, a significant difference in the 2 methods exists at 90% confidence.
difference in the 2 methods exists at 95% confidence. Also, for 7 degrees of freedom at
99% confidence t = 3.50 indicating a significant difference in the 2 methods exists at 99%
confidence, as well.
(c) For this example the same conclusion is obtained at all three confidence levels.
spooled = 0.0027
degrees of freedom = 4 + 3 – 2 = 5
2.2978 − 2.3106
t= = −6.207
4+3
0.0027
4×3
The Excel TDIST(x,df,tails) function can be used to calculate the probability of getting a
99.84% confident that the nitrogen prepared in the two ways is different. There is a
7-25 (a)
Variation source SS df MS F
(b) H0: µsamp1 = µsamp2 = µsamp3; Ha: at least two of the means differ.
Fundamentals of Analytical Chemistry: 8th ed. Chapter 7
(c) From Table 7-4 the F value for 12 degrees of freedom in the denominator and 2
exceeds F tabulated we reject H0 and conclude that the phosphorous contents of the soil
7-26 (a)
Variation source SS df MS F
(b) H0: µbrand1 = µbrand2 = µbrand3 = µbrand4 = µbrand5; Ha: at least two of the means differ.
(c) The Excel FINV(prob,df1,df2) function can be used to calculate the F value for the
above problem. In this case we find FINV(0.05,4,25) = 2.76. Since F calculated exceeds
F tabulated we reject the null hypothesis and conclude that the average ascorbic acid
7-27 (a) H0: µLabA = µLabB = µLabC = µLabD = µLabE; Ha: at least two of the means differ.
(b)
A B C D E F
1 Result No. Lab A Lab B Lab C Lab D Lab E
2 1 10.3 9.5 10.1 8.6 10.6
3 2 11.4 9.9 10.0 9.3 10.5
4 3 9.8 9.6 10.4 9.2 11.1
5
6 Average = 10.50 9.67 10.17 9.03 10.73
7 Stan. Dev. = 0.81854 0.20817 0.20817 0.37859 0.32146
8 Variance = 0.67000 0.04333 0.04333 0.14333 0.10333
9
10 Grand Mean = 10.020
11 SSF = 5.57733 10.73 - 9.03 = 1.70 (a significant difference)
12 SSE = 2.00667 10.50 - 9.03 = 1.47 (a significant difference)
13 SST = 7.58400 10.17 - 9.03 = 1.14 (a significant difference)
14 9.67 - 9.03 = 0.64 (no significant difference)
15 MSF = 1.39433
16 MSE = 0.20067 10.73 - 9.67 = 1.06 (a significant difference)
17 10.50 - 9.67 = 0.83 (a significant difference)
18 F= 6.94850 10.17 - 9.67 = 0.5 (no significant difference)
19
20 LSD = 0.816 10.73 - 10.17 = 0.56 (no significant difference)
21 10.50 - 10.17 = 0.33 (no significant difference)
22
23 10.73 - 10.50 = 0.23 (no significant difference)
24
25 Spreadsheet Documentation
26 B6 = AVERAGE(B2:B4)
27 B7 = STDEV(B2:B4)
28 B8 = B7^2
29 B10 = SUM(B2:F4)/15
30 B11 = 3*((B6-B10)^2+(C6-B10)^2+(D6-B10)^2+(E6-B10)^2+(F6-B10)^2)
31 B12 = 2*SUM(B8:F8)
32 B13 = B11+B12
33 B15 = B11/4
34 B16 = B12/10
35 B18 = B15/B16
36 B20 = 2.23*SQRT(2*B16/3)
From Table 7-4 the F value for 4 degrees of freedom in the numerator and 10 degrees of
reject H0 and conclude that the laboratories differ at 95% confidence. We can also be
99% confident that the laboratories differ, but we cannot be 99.9% confident that the
laboratories differ.
(c) Based on the calculated LSD value laboratories A, C and E differ from laboratory D,
but laboratory B does not. Laboratories E and A differ from laboratory B, but laboratory
7-28 (a) H0: µAnalyst1 = µAnalyst2 = µAnalyst3 = µAnalyst4; Ha: at least two of the means differ.
(b)
A B C D E F
1 Result No. A1 A2 A3 A4
2 1 10.24 10.14 10.19 10.19
3 2 10.26 10.12 10.1 10.15
4 3 10.29 10.04 10.15 10.16
5 4 10.23 10.07 10.12 10.1
6
7 Average = 10.26 10.09 10.14 10.15
8 Stan. Dev. = 0.02646 0.04573 0.03594 0.03742
9 Variance = 0.00070 0.00209 0.00129 0.00140
10
11 Grand Mean = 10.160
12 SSF = 0.05595
13 SSE = 0.01645
14 SST = 0.07240 10.26 - 10.09 = 0.17 (a significant difference)
15 10.15 - 10.09 = 0.06 (a significant difference)
16 MSF = 0.01865 10.14 - 10.09 = 0.05 (no significant difference)
17 MSE = 0.00137
18 10.26 - 10.14 = 0.12 (a significant difference)
19 F= 13.60486 10.15 - 10.14 = 0.01 (no significant difference)
20
21 LSD = 0.057 10.26 - 10.15 = 0.11 (a significant difference)
22
23
24 Spreadsheet Documentation
25 B7 = AVERAGE(B2:B5)
26 B8 = STDEV(B2:B5)
27 B9 = B8^2
28 B11 = SUM(B2:E5)/16
29 B12 = 4*((B7-B11)^2+(C7-B11)^2+(D7-B11)^2+(E7-B11)^2)
30 B13 = 3*SUM(B9:F9)
31 B14 = B12+B13
32 B16 = B12/3
33 B17 = B13/12
34 B19 = B16/B17
35 B21 = 2.19*SQRT(2*B17/4)
From Table 7-4 the F value for 3 degrees of freedom in the numerator and 12 degrees of
reject H0 and conclude that the analysts differ at 95% confidence. We can also be 99%
(c) Based on the calculated LSD value there is a significant difference between analyst 2
and analysts 1 and 4, but not analyst 3. There is a significant difference between analyst
3 and analyst 1, but not analyst 4. There is a significant difference between analyst 1 and
analyst 4.
Fundamentals of Analytical Chemistry: 8th ed. Chapter 7
7-29 (a) H0: µDes1 = µDes2 = µDes3 = µDes4; Ha: at least two of the means differ.
(b)
A B C D E
1 Result No. Des 1 Des 2 Des 3 Des 4
2 1 72 93 96 100
3 2 93 88 95.0 84
4 3 76 97 79 91
5 4 90 74 82 94
6
7 Average = 82.75 88.00 88.00 92.25
8 Stan. Dev. = 10.30776 10.03328 8.75595 6.65207
9 Variance = 106.2500 100.6667 76.6667 44.2500
10
11 Grand Mean = 87.750
12 SSF = 181.5
13 SSE = 983.5
14 SST = 1165
15
16 MSF = 60.50000
17 MSE = 81.95833
18
19 F= 0.73818
20
21
22 Spreadsheet Documentation
23 B7 = AVERAGE(B2:B5)
24 B8 = STDEV(B2:B5)
25 B9 = B8^2
26 B11 = SUM(B2:E5)/16
27 B12 = 4*((B7-B11)^2+(C7-B11)^2+(D7-B11)^2+(E7-B11)^2)
28 B13 = 3*SUM(B9:F9)
29 B14 = B12+B13
30 B16 = B12/3
31 B17 = B13/12
32 B19 = B16/B17
33 B21 = 2.19*SQRT(2*B17/4)
From Table 7-4 the F value for 3 degrees of freedom in the numerator and 12 degrees of
freedom in the denominator at 95% is 3.49. Since F calculated is less than F tabulated
we accept H0 and conclude that 4 flow cell designs give the same results at 95%
Fundamentals of Analytical Chemistry: 8th ed. Chapter 7
confidence.
(c) No differences.
7-30 (a) H0: µColorimetry = µEDTA Tit = µAA; Ha: at least two of the means differ.
(b)
A B C D E F G
1 Result No. Color. EDTA AA
2 1 3.92 2.99 4.40
3 2 3.28 2.87 4.92
4 3 4.18 2.17 3.51
5 4 3.53 3.40 3.97
6 5 3.35 3.92 4.59
7
8 Average = 3.65 3.07 4.28
9 Stan. Dev. = 0.38571 0.64958 0.54979
10 Variance = 0.1488 0.42195 0.3023
11
12 Grand Mean = 3.667
13 SSF = 3.649773
14 SSE = 3.49196
15 SST = 7.141733
16
17 MSF = 1.82489 4.28 - 3.07 = 1.21 (a significant difference)
18 MSE = 0.29100 3.65 - 3.07 = 0.58 (no significant difference)
19
20 F= 6.27116 4.28 - 3.65 = 0.63 (no significant difference)
21
22 LSD = 0.747
23
24
25 Spreadsheet Documentation
26 B8 = AVERAGE(B2:B6)
27 B9 = STDEV(B2:B6)
28 B10 = B9^2
29 B12 = SUM(B2:D6)/15
30 B13 = 5*((B8-B12)^2+(C8-B12)^2+(D8-B12)^2)
31 B14 = 4*SUM(B10:D10)
32 B15 = B13+B14
33 B17 = B13/2
34 B18 = B14/12
35 B20 = B17/B18
36 B22 = 2.19*SQRT(2*B18/5)
Fundamentals of Analytical Chemistry: 8th ed. Chapter 7
From Table 7-4 the F value for 2 degrees of freedom in the numerator and 12 degrees of
freedom in the denominator at 95% is 3.89. Since F calculated is greater than F tabulated
we reject H0 and conclude that the 3 methods give different results at 95% confidence.
(c) Based on the calculated LSD value there is a significant difference between the
atomic absorption method and the EDTA titration. There is no significant difference
between the EDTA titration method and the colorimetry method and there is no
significant difference between the atomic absorption method and the colorimetry method.
41.27 − 41.61
7-31 (a) Q = = 0.596 and Qcrit for 4 observations at 95% confidence = 0.829.
41.84 − 41.27
Since Q < Qcrit the outlier value 41.27 cannot be rejected with 95% confidence.
7.388 − 7.295
(b) Q = = 0.894 and Qcrit for 4 observations at 95% confidence = 0.829.
7.388 − 7.284
Since Q > Qcrit the outlier value 7.388 can be rejected with 95% confidence.
85.10 − 84.70
7-32 (a) Q = = 0.833 and Qcrit for 3 observations at 95% confidence = 0.970.
85.10 − 84.62
Since Q < Qcrit the outlier value 85.10 cannot be rejected with 95% confidence.
85.10 − 84.70
(b) Q = = 0.833 and Qcrit for 4 observations at 95% confidence = 0.829.
85.10 − 84.62
Since Q > Qcrit the outlier value 85.10 can be rejected with 95% confidence.
4.60 − 4.50
7-33 Q= = 0.5 and Qcrit for 5 observations at 95% confidence = 0.710.
4.60 − 4.40
Since Q < Qcrit the outlier value 4.60 ppm cannot be rejected with 95% confidence.