Scribd
Upload
89 views18 pages

Distribusi Data: Disusun: Sotya Fevriera

The document discusses various ways to detect the shape of a data distribution, including: - Comparing the mean, median, and mode to see if outliers are present. - Creating histograms, normal probability plots, and box plots. - Calculating measures of skewness and kurtosis to determine if the distribution is symmetric, left-skewed, or right-skewed. Outliers can be identified using box plots or by calculating z-scores. The document also discusses how to handle distributions with outliers.

Uploaded by

Monica Cahyaningtyas
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
89 views18 pages

Distribusi Data: Disusun: Sotya Fevriera

The document discusses various ways to detect the shape of a data distribution, including: - Comparing the mean, median, and mode to see if outliers are present. - Creating histograms, normal probability plots, and box plots. - Calculating measures of skewness and kurtosis to determine if the distribution is symmetric, left-skewed, or right-skewed. Outliers can be identified using box plots or by calculating z-scores. The document also discusses how to handle distributions with outliers.

Uploaded by

Monica Cahyaningtyas
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 18

Distribusi Data

Disusun:
Sotya Fevriera

Pengantar Statistika (GE 120)


Bentuk Distribusi Data (1)
• Tdk ada nilai ekstrim  simetri
• Ada nilai ekstrim bawah  menceng ke kiri
(negative/left-skewed)
• Ada nilai ekstrim atas  menceng ke kanan
(positive/right-skewed)
Left Skewed Symmetric Right Skewed

April 2019 By Sotya Fevriera – FEB UKSW 2


Bentuk Distribusi Data (2)
• Bentuk distribusi simetri merupakan salah satu

ciri-ciri data yang berdisttribusi normal.

Latihan 20:

Kenali bentuk distribusi data HDI dengan cara-

cara berikut ini!

April 2019 By Sotya Fevriera – FEB UKSW 3


Bentuk Distribusi Data (3)
Bentuk distribusi data dpt dideteksi dgn:
• Membandingkan mean, median & modus

April 2019 By Sotya Fevriera – FEB UKSW 4


Bentuk Distribusi Data (4)
Bentuk distribusi data dpt dideteksi dgn:
• Kurva Normal/Poligon
• Histogram
• Diagram Batang Daun (Stem & Leaf Plot)
• Normal P-P Plot atau Normal Q-Q Plot
▪ Dist. normal/simetri jk plot mdekati garis
lurus.
• Uji normalitas data (diajarkan pd materi
regresi)
• Diagram Kotak (Box Plots)

April 2019 By Sotya Fevriera – FEB UKSW 5


Bentuk Distribusi Data (5)
simetri
Xmin Q1 Median Q3 Xmax

d d
menceng
kanan
d1 d2
menceng
kiri
d1 d2
rectangular

d d d d

April 2019 By Sotya Fevriera – FEB UKSW 6


Bentuk Distribusi Data (6)
Btk dist. data dpt dideteksi dgn (lanjutan):
• Ukuran Kemencengan (Skewness)
▪ skewness = 0  simetri
▪ skewness > 0  menceng kanan
▪ skewness < 0  menceng kiri
▪ Distribusi tidak normal jika:
◦ |skewness| > 1
◦ |skewness/standard error skew.| > 2
▪ Formula skewness dpt dilihat dlm
excel-help.
April 2019 By Sotya Fevriera – FEB UKSW 7
Bentuk Distribusi Data (7)
Btk distribusi data dpt dideteksi dgn (lanjutan):
• Ukuran Keruncingan (kurtosis)
Mengukur konsentrasi data di tengah distribusi
dibdgkan data di tepi distribusi.
▪ kurt = 0  mesokurtic  dist. normal
▪ kurt > 0  leptokurtic  kurva runcing
 the observations cluster more and
have longer tails than those in the
normal distribution
▪ kurt < 0  platikurtic  kurva landai
 the observations cluster less and have
shorter tails
April 2019 By Sotya Fevriera – FEB UKSW 8
Bentuk Distribusi Data (8)
Bentuk distribusi data dpt dideteksi dengan

(lanjutan):

• Ukuran Keruncingan (kurtosis)

▪ Distribusi tidak normal jika:

◦ |kurtosis/standard error kurt.| > 2

▪ Formula kurt dpt dilihat dlm excel-help.

April 2019 By Sotya Fevriera – FEB UKSW 9


Bentuk Distribusi Data (9)
• Mengidentifikasi nilai ekstrim (outlier)
• Menggunakan Box Plots:

Q1 Md Q3

1,5(Q3-Q1) 1,5(Q3-Q1)
inner fences inner fences
3(Q3-Q1) 3(Q3-Q1)
outer fences outer fences

April 2019 By Sotya Fevriera – FEB UKSW 10


Bentuk Distribusi Data (10)
▪ mild outliners (di SPSS ditandai dot- ◦)

 diletakkan di antara inner & outer fences

▪ extreme outliners (di SPSS ditandai bintang- *)

 diletakkan di luar outer fences

April 2019 By Sotya Fevriera – FEB UKSW 11


Bentuk Distribusi Data (11)
• Menggunakan nilai Z (Skor Baku)
▪ Bmanfaat utk mengidentifikasi nilai ekstrim
(outliers):
X X
Zi  i
S
▪ Aturan umum:
Xi merup nilai ekstrim jk Zi  −3 atau Zi  3.
Latihan 21:
Hitung nilai Z utk data HDI!

April 2019 By Sotya Fevriera – FEB UKSW 12


Bentuk Distribusi Data (12)
• Bagaimana jika data mengandung nilai ekstrim?
 Nilai ekstrim akan bpengaruh thd rata2, mk jk
data tdk terlalu banyak, tetap gunakan
seluruh data tetapi gunakan Md atau Mo utk
mewakili ukuran pemusatan.
 Lakukan transformasi data shg distribusi data
mjd simetri (dist normal) & kmd ukuran
penyebaran dpt diwakili rata2 (cara ini tdk
diajarkan dlm materi ini).
 Gunakan metode statistika non parametrik.

April 2019 By Sotya Fevriera – FEB UKSW 13


Bentuk Distribusi Data (13)
• Bagaimana jika data mengandung nilai ekstrim?
 Jk jlh data ckp byk, buang data yg merup nilai
ekstrim (shg data mjd lbh sedikit) & kmd
ukuran penyebaran dpt diwakili rata2. Tetapi,
ada pula pendapat bahwa jika data ekstrim
tersebut bukan disebabkan karena kesalahan
pada metode pengumpulan data, maka
sebaiknya tetap dipertahankan.
April 2019 By Sotya Fevriera – FEB UKSW 14
Bentuk Distribusi Data (14)
• Dalil Chebysev
“Utk sembarang kelompok data, dgn mengabaikan btk

distribusinya, persentase nilai yang berada dlm interval

k kali standar deviasi dari rata-rata haruslah minimal:

[1 − (1/k2)]  100%; k > 1”

April 2019 By Sotya Fevriera – FEB UKSW 15


Bentuk Distribusi Data (15)
Contoh:
Data HDI dari 505 kota/kabupaten memiliki
rata-rata 71,94 dan standar deviasi 5,24.
Untuk k  2 maka menurut Dalil Chebysev,
kabupaten/kota yang memiliki HDI dalam
interval: [71,94  (2  5,24)]  [61,46 ; 82,42],
minimal: 1 − (1/22)]  100% = 75%.
Kenyataan: 485/505 = 96% (terbukti)

April 2019 By Sotya Fevriera – FEB UKSW 16


Bentuk Distribusi Data (16)
• Kaidah Empiris
Jika distribusi data mendekati simetri, maka:
 sekitar 68% data akan ada dlm interval (X  S)
 sekitar 95% data akan ada dlm interval (X  2S)
 sekitar 99% data akan ada dlm interval (X  3S)
Contoh:
Utk k = 2 mk byknya data HDI yg ada dlm
interval [71,94  (2  5,24)]  [61,46 ; 82,42]
hrsnya  95%.
Kenyataan: 485/505 = 96% (terbukti)

April 2019 By Sotya Fevriera – FEB UKSW 17


Tugas
Lihat petunjuk di PPT terpisah.

April 2019 By Sotya Fevriera – FEB UKSW 18

You might also like