290 likes | 664 Vues
5. Ukuran Sebaran ( keragaman ). Kuswanto 2012. Ukuran keragaman. Dari tiga ukuran pemusatan, belum dapat memberikan deskripsi yang lengkap bagi suatu data. Perlu juga diketahui seberapa jauh pengamatan-pengamatan tersebut menyebar dari rata-ratanya.
E N D
5. UkuranSebaran (keragaman) Kuswanto2012
Ukuran keragaman • Dari tiga ukuran pemusatan, belum dapat memberikan deskripsi yang lengkap bagi suatu data. • Perlu juga diketahui seberapa jauh pengamatan-pengamatan tersebut menyebar dari rata-ratanya. • Ada kemungkinan diperoleh rata-rata dan median yang sama, namun berbeda keragamannya. • Beberapa ukuran keragaman yang sering kita temui adalah range (rentang=kisaran=wilayah), simpangan (deviasi), varian (ragam), simpangan baku (standar deviasi) dan koefisien keragaman.
f f X X Measures of Dispersion and Variability These are measurements of how spread the data is around the center of the distribution
2. DEVIATION DEVIASI = SIMPANGAN You could express dispersion in terms of deviation from the mean, however, a sum of deviations from the mean will always = 0. i.e. (Xi- X) = 0 So, take an absolute value to avoid this Problem –the more numbers in the data set, the higher the SS
2 2 3 4 5 = X1 = X2 = X3 = X4 = X5 • Range Kisaran = Rentang • difference between lowest and highest numbers Place numbers in order of magnitude, then range = Xn - X1. Range = 5 - 2 = 3 Problem - no information about how clustered the data is
Sample SS = (Xi - X)2 = Sample mean deviation = | Xi- X | n 3. Mean Deviation = Simpangan Rerata Essentially the average deviation from the mean 4. Variance = Ragam Another way to get around the problem of zero sums is to square the deviations. Known as sum of squares or SS Xi2 - (Xi)2/n SS is much more common than mean deviation
Example Sample SS = (Xi - X)2 2 2 3 4 5 = X1 = X2 = X3 = X4 = X5 X = 3.2 = 1.44 + 1.44 + 0.04 + 0.64 + 3.24 = 6.8 SS = (2 - 3.2)2 + (2 - 3.2)2 + (3 - 3.2)2 + (4 - 3.2)2 + (5 -3.2)2 Problem –the more numbers in the data set, the higher the SS
Population Variance (2 ): This is just SS 2 = (Xi - )2 N N Our best estimate of 2is sample variance (s2): S2 = (Xi - X)2 n - 1 Note : divide by n-1 known as degrees of freedom The mean SS is known as the variance Xi2 - (Xi)2/n = n - 1 Problem - units end up squared
= (Xi - )2 N = 2 s = s2 s = (Xi - X )2 n - 1 5. Standard Deviation (Standar Deviasi) => square root of variance For a population: For a sample:
Example 2 2 3 4 5 = X1 = X2 = X3 = X4 = X5 X = 3.2 s = (2 - 3.2)2 + (2 - 3.2)2 + (3 - 3.2)2 + (4 - 3.2)2 + (5 -3.2)2 5 - 1 = 1.44 + 1.44 + 0.04 + 0.64 + 3.24 4 = 1.304 s = (Xi - X )2 n - 1
s CV = X 6. Coefficient of Variation = Koefisien Keragaman = KK (V or sometimes CV): Variance (s2) and standard deviation (s) have magnitudes that are dependent on the magnitudes of the data. The coefficient of variation is a relative measure, so variability of different sets of data may be compared (stdev relative to the mean) Note that there are no units – emphasizes that it is a relative measure X 100% Sometimes expressed as a %
Example: 2 2 3 4 5 = X1 = X2 = X3 = X4 = X5 X = 3.2 g s CV = X 1.304 g CV = 3.2 g CV = 0.4075 or CV = 40.75% (X 100%) s = 1.304 g Attention there is not any UNIT, or %
68.27% 95.44% f 99.73% 3 2 2 3 X 8. The Normal Distribution (Distribusi Normal) : There is an equation which describes the height of the normal curve in relation to its standard dev ()
Normal distribution with σ=1, with varying means μ= 1 μ= 2 μ= 0 ƒ 4 5 -3 -2 -1 0 1 2 3 If you get difficulties to keep this term, read statistics books
σ = 1.5 σ = 2 Normal distribution with μ= 0, with varying standard deviations σ = 1 ƒ -5 -4 -3 -2 -1 0 1 2 3 4 5
Mean, median and mode 9. Symmetry and Kurtosis Symmetry means that the population is equally distributed around the mean i.e. the curve to the right side of the mean is a mirror image of the curve to the left side ƒ
ƒ ƒ Symmetry Data may be positively skewed (skewed to the right) Or negatively skewed (skewed to the left) So direction of skew refers to the direction of longer tail
mode ƒ median mean Symmetry
ƒ ƒ Kurtosis refers to how flat or peaked a curve is (sometimes referred to as peakedness or tailedness) The normal curve is known as mesokurtic A more peaked curve is known as leptokurtic A flatter curve is known as platykurtic
Latihandandiskusi • Banyaknyabuahpisang yang tersengathamadari 16 tanamanadalah 4, 9, 0, 1, 3, 24, 12, 3, 30, 12, 7, 13, 18, 4, 5, dan 15. Denganmenganggap data tersebutsebagaicontoh, hitunglahvarian, simpanganbakudankoefisienkeragamannya. Statistikmana yang paling tepatuntukmenggambarkankeragaman data tersebut? • To study how first-grade students utilize their time when assigned to a math task, researcher observes 24 students and records their time off task out of 20 minutes. Times off task (minutes) : 4, 0, 2, 2, 4, 1, 4, 6, 9, 7, 2, 7, 5, 4,13, 7, 7, 10, 10, 0, 5, 3, 9 and 8. For this data set, find : • Mean and standard deviation, median and range • Display the data in the histogram plot, dot diagram and also stem-and-leaf diagram • Determine the intervals x ± s, x ± 2s, x ± 3s • Find the proportion of the meausurements that lie in each of this intervals. • Compare your finding with empirical guideline of bell-shaped distribution
3. The data below were obtained from the detailed record of purchases over several month. The usage vegetables (in weeks) for a household taken from consumer panel were (gram) : • 84 58 62 65 75 76 56 87 68 77 87 55 65 66 76 78 74 81 83 78 75 74 60 50 86 80 81 78 74 87 a. Plot a histogram of the data! b. Find the relative frequency of the usage time that did not exceed 80. c. Calculate the mean, variance and the standard deviation d. Calculate the median and quartiles. 4. The mean of corn weight is 278 g by ear and deviation standard is 9,64 g, and than we have 10 ears. If they are gotten from ten different fields, mean of plant height is Rp. 1200,- and its deviation standard is Rp 90,-, which one have more homogenous, the weight of corn ear or the plant height? Explain your answer! Verify your results by direct calculation with the other data.
5. The employment’s salary at seed company, abbreviated, as follows : 18, 15, 21, 19, 13, 15, 14, 23, 18 and 16 rupiah. If these abbreviation is real salary divide Rp. 100.000,-, find the mean, variance and deviation standard of them. 6. Computer-aided statistical calculations. Calculation of the descriptive statistic such as x and s are increasingly tedious with large data sets. Modern computers have come a long way in alleviating the drudgery of hand calculation. Microsoft Exel, Minitab or SPSS are three of computing packages those are easy accessible to student because its commands are in simple English. Find these programs and install its at your computers. Bellow main and sub menu of Microsoft Exel, Minitab and SPSS program. Use these software to find x, s, s2, and coefisien of variation (CV) for data set in exercise b. Histogram and another illustration can also be created.
7. Some properties of the standard deviation • if a fixed number c is added to all measurements in a data set, will the deviations (xi -x) remain changed? And consequently, will s² and s remain changed, too? Take data sample. • If all measurements in a data set are multiplied by a fixed number d, the deviation (xi -x) get multiplied by d. Is it right? What about the s² and s? Take data sample. • Apply your computer software to explain your data sample. Verify your results by other data.