Professional Documents
Culture Documents
DKK
n
x f
k
i
i i
Standard deviation: 73 . 104
202 , 2
559 , 154 , 24
) (
1
2
n
x f
k
i
i i
15
Histogram, Quartiles, Median and Box-plot
Consider the relative and cumulative distribution of data
Statistics EUS & Negot Chinese 29
Disponible husstandsindkomster, Danmark, 1987
i
Interval for incomes
1,000 DKK
Number of
households,
1,000
Number of
households
frequency, %
Cumulative
frequency, %
fi fi/n
1
2
3
4
5
6
7
8
0
50
100
150
200
250
300
400
- 49.9
- 99.9
- 149.9
- 199.9
- 249.9
299.9
399.9
-
146
590
414
323
325
210
139
55
6.6
26.8
18.8
14.7
14.8
9.5
6.3
2.5
6.6
33.4
52.2
66.9
81.7
91.2
97.5
100.0
Sum 2,202 100.0
Source: Statistics Denmark, Annual Statistical Review, 1994, page 220-221
Histogram
Distribution Income, Denmark, 1987
0,00
5,00
10,00
15,00
20,00
25,00
30,00
0 - 49 50 - 99 100 -
149
150 -
199
200 -
249
250 -
299
300 -
349
350 -
399
Above
400
%
Statistics EUS & Negot Chinese 30
16
Sum Function
Statistics EUS & Negot Chinese 31
How to do the interpolation
We use a formula for example given as:
Value = End value interval
" "
" "
pct percent in width Total
fractile to relative long too
interval width in value
Illustration:
Frequency %
52.2
50
33.4
100 ? 149 income (1,000 DKK)
Statistics EUS & Negot Chinese 32
17
Median: 149 , 144 851 , 5 000 , 150 000 , 50
8 . 18
) 50 2 . 52 (
000 , 150
Similarly for the other quartiles and deciles:
Lower quartile: 328 , 84 000 , 50
8 . 26
) 25 4 . 33 (
000 , 100
(Q
1
)
Upper quartile: 365 , 227 000 , 50
8 . 14
) 75 7 . 81 (
000 , 250
(Q
3
)
Lower decile: 343 , 56 000 , 50
8 . 26
) 10 4 . 33 (
000 , 100
Upper decile: 684 , 293 000 , 50
5 . 9
) 90 2 . 91 (
000 , 300
Statistics EUS & Negot Chinese 33
Inter Quartile Range (IQR): (Q
3
Q
1
) = 227,365 84,328 = 143,037
Lower inner fence: Q
1
1.5IQR = 84,328 1.5(143,037) = 130,228
Lower outer fence: Q
1
3.0IQR = 84,328 3.0(143,037) = 344,783
Upper inner fence: Q
3
+ 1,5IQR = 227,365 + 1.5(143,037) = 441,921
Upper outer fence: Q
3
+ 3.0IQR = 227,365 + 3.0(143,037) = 656,476
Box-plot
300 200 100 0 100 200 300 400 500 600
LOF = 345 LIF = 130 Q
1
=84 M=144 Q
3
=227 UIF = 442 UOF = 656
Statistics EUS & Negot Chinese 34
18
9. Descriptive Statistics an Example of Outliers
Outliers are extremes
Outliers make distributions non-normal
Outliers changes the mean, standard deviation and skewness
However, the median remains constant
Statistics EUS & Negot Chinese 35
Basic Max=34 Max=44 Max=54
Mean 15.85 16.35 16.85 17.35 Increases
Standard Error 1.00 1.29 1,69 2.13
Median 16 16 16 16 Constant!!
Modus / Mode 16 16 16 16
Standard deviation 4.46 5.79 7.56 9.52
Sample variance 19.92 33.50 57.08 90.66
Kurtosis 0.12 3.88 8.99 12.55
Skewness -0.35 1.19 2.43 3.16 Increases
Range 18 28 38 48
Minimum 6 6 6 6
Maximum 24 34 44 54
Sum 317 327 337 347
Observations 20 20 20 20
Confidence interval(95 %) 2.09 2.71 3.54 4.46 Increases
Statistics EUS & Negot Chinese 36