Perbedaan Statistika Parametrik dan Non Parametrik

Summary

This video script offers an insightful overview of parametric and nonparametric statistics, explaining the conditions under which each type of statistical analysis is appropriate. It covers the importance of data distribution in choosing the right test, with parametric tests used for normally distributed ratio or interval scale data and nonparametric tests for non-normal distributions or ordinal and nominal scales. The speaker also introduces various statistical tests, including t-tests, Wilcoxon, Mann-Whitney, ANOVA, Kruskal-Wallis, and logistic regression, tailored to different data types and research questions. Despite technical difficulties with the camera overheating, the presenter remains dedicated to educating viewers on the versatility of statistical methods, even when data does not meet normal distribution assumptions.

Takeaways

• 📚 The video discusses the topic of statistics, specifically parametric and nonparametric statistics.
• 🔍 Statistics is divided into descriptive and inferential statistics; descriptive for presenting data, and inferential for drawing conclusions through data analysis.
• 📈 Descriptive statistics make data more understandable through tables and graphs, while inferential statistics use analysis methods.
• 📊 Inferential statistical analysis is further divided into parametric and nonparametric methods based on the data's distribution and scale.
• 📉 Parametric statistics are used for ratio and interval scales, requiring normal distribution, while nonparametric statistics are for nominal and ordinal scales without normal distribution.
• 🌡 Examples of interval scale data include body temperature and height, which have absolute zero values and no negative values.
• 🏷️ Nominal scale data are categorical without order or rank, such as religion or gender.
• 🔢 Ordinal scale data are categorized but have a rank or order, like Likert scale responses in surveys.
• 🧐 When data does not follow a normal distribution, nonparametric tests like the Wilcoxon or Mann-Whitney tests are used instead of parametric tests.
• 📝 The video references a book titled 'Dasar Metodologi Penelitian Kuantitatif, Kualitatif, dan Statistika' by Sarmanu published in 2017 as the source material.
• 🌟 The speaker encourages viewers to share references or books for further discussion on nonparametric tests that were not explained in the video.

Q & A

What is the main topic of the video script?

The main topic of the video script is statistics, specifically focusing on parametric and nonparametric statistics.

What are the two types of statistical analysis mentioned in the script?

The two types of statistical analysis mentioned are descriptive statistics and inferential statistics.

What is the difference between parametric and nonparametric statistics?

Parametric statistics are used when the data follows a normal distribution and the scale is ratio or interval. Nonparametric statistics are used when the data does not follow a normal distribution and the scale is nominal or ordinal.

What are the conditions for using parametric tests?

Parametric tests are used when the data is on a ratio or interval scale and is normally distributed.

What are the conditions for using nonparametric tests?

Nonparametric tests are used when the data is on a nominal or ordinal scale or when the data does not follow a normal distribution.

What are the types of data scales mentioned in the script?

The types of data scales mentioned are ratio, interval, nominal, and ordinal.

What is an example of a ratio scale measurement?

An example of a ratio scale measurement is height or weight, where zero represents an absolute absence of the attribute.

What is an example of an interval scale measurement?

An example of an interval scale measurement is temperature, which can have negative values but does not have a true zero point.

What is the reference book mentioned in the script?

The reference book mentioned is 'Dasar Metodologi Penelitian Kuantitatif, Kualitatif dan Statistika' by Sarmanu, published in 2017.

What are some of the statistical tests mentioned in the script for different data scales and distributions?

Some of the statistical tests mentioned include t-tests, z-tests, Wilcoxon test, Mann-Whitney test, ANOVA, Kruskal-Wallis test, and logistic regression.

What is the purpose of using descriptive statistics?

The purpose of using descriptive statistics is to present data in a way that is easier to understand, such as through tables and graphs.

What is the purpose of using inferential statistics?

The purpose of using inferential statistics is to draw conclusions about a population based on sample data using statistical methods of analysis.

What is the script's advice for dealing with data that is not normally distributed?

The script advises using nonparametric tests when dealing with data that is not normally distributed.

What is the script's suggestion for further learning about nonparametric tests?

The script suggests discussing and sharing references or books on nonparametric tests to enhance understanding and become a collective learning source.

Outlines

00:00

📚 Introduction to Parametric and Nonparametric Statistics

The speaker begins by greeting the audience and introducing the topic of statistics, specifically parametric and nonparametric statistics. They mention that the material is based on a 2017 book titled 'Quantitative, Qualitative, and Statistical Research Methodology.' The speaker explains the difference between descriptive and inferential statistics, highlighting that the former presents data in an understandable format through tables and graphs, while the latter involves data analysis methods. The choice between parametric and nonparametric analysis is determined by the scale of the data and its distribution. Parametric tests are used for ratio and interval scales, assuming normal distribution, while nonparametric tests are applied for nominal and ordinal scales without normal distribution.

05:00

📐 Understanding Data Scales and Statistical Tests

This paragraph delves into the different scales of data: ratio, interval, nominal, and ordinal. The ratio scale is characterized by an absolute zero, such as weight or height, while the interval scale does not have an absolute zero, exemplified by temperature and exam scores. Nominal scale data are categorical without order or rank, like religion or gender, whereas ordinal scale data have ranks or order, such as survey responses on a Likert scale. The speaker then discusses the appropriate statistical tests for each scale and distribution type, including parametric tests like t-tests or z-tests for normally distributed data and nonparametric tests like Wilcoxon or Mann-Whitney tests for non-normally distributed data.

10:02

🔍 Advanced Statistical Analysis Techniques

The speaker discusses advanced statistical analysis techniques for different types of data and research questions. For ordinal data, they mention the use of nonparametric tests like Wilcoxon or Mann-Whitney tests. For comparing more than two treatments, they introduce ANOVA for normally distributed data and Kruskal-Wallis or Friedman tests for non-normally distributed data. The speaker also covers multivariate analysis for more than one dependent variable, mentioning MANOVA and Hotelling's T-squared test. They discuss regression analysis for the effect of variables on a dependent variable, logistic regression for nominal data, and various correlation tests for different data scales and distributions. The paragraph concludes with a mention of univariate and multivariate tests, and the speaker invites the audience to share references or books for further discussion on nonparametric tests not covered in the book.

Keywords

💡Statistics

Statistics is the science of collecting, organizing, analyzing, and interpreting data. In the video, the speaker discusses various statistical methods, emphasizing its importance in data analysis and research methodology. The term is central to the video's theme of understanding and applying statistical tests.

💡Parametric Statistics

Parametric statistics refers to a set of statistical methods that assume the data follows a particular probability distribution, often the normal distribution. The video mentions when to use parametric tests, such as when data is on a ratio or interval scale and is normally distributed.

💡Nonparametric Statistics

Nonparametric statistics are statistical methods that do not rely on the assumption of a specific distribution of the data. The video script discusses the use of nonparametric tests when data does not follow a normal distribution, emphasizing their applicability in various scenarios.

💡Descriptive Statistics

Descriptive statistics are used to summarize and organize data in a meaningful way, such as through tables and graphs. The video explains that this type of statistics helps in making data more comprehensible to others, showcasing the data presentation aspect of statistics.

💡Inferential Statistics

Inferential statistics involve making inferences about a population based on a sample. The video contrasts this with descriptive statistics, highlighting that inferential statistics use analytical methods to go beyond the data at hand to make broader conclusions.

💡Data Scales

Data scales refer to the types of scales on which data can be measured, such as ratio, interval, nominal, and ordinal. The script explains the characteristics of each scale and their relevance to choosing the appropriate statistical tests.

💡Normal Distribution

A normal distribution is a continuous probability distribution in which data points are symmetrically distributed around the mean. The video script discusses the importance of data being normally distributed for the use of parametric tests.

💡Hypothesis Testing

Hypothesis testing is a process of making decisions about a population parameter based on a sample. The video mentions various tests like t-tests and z-tests for parametric data and Wilcoxon or Mann-Whitney tests for nonparametric data, illustrating the application of hypothesis testing in statistics.

💡ANOVA (Analysis of Variance)

ANOVA is a statistical method used to compare the means of more than two groups to see if they are significantly different. The video script refers to ANOVA as a parametric test used when data is normally distributed and the researcher wants to compare multiple treatments.

💡Kruskal-Wallis Test

The Kruskal-Wallis test is a nonparametric method for comparing three or more independent samples to determine if they originate from the same distribution. The video mentions this test as an alternative to ANOVA when data does not follow a normal distribution.

💡Regression Analysis

Regression analysis is a set of statistical processes that estimates the relationships among variables. The video script discusses using regression analysis when the data is on a ratio or interval scale to examine the impact of variables on one another.

💡Logistic Regression

Logistic regression is a statistical method for analyzing a dataset in which the response variable is categorical. The video mentions using logistic regression when the data is nominal, indicating its use for binary outcome predictions.

💡Correlation

Correlation measures the extent to which two variables are linearly related. The video script differentiates between Pearson correlation for normally distributed data and Spearman's rank correlation for non-normally distributed ordinal data, showing the importance of selecting the right correlation method based on data characteristics.

💡Multivariate Analysis

Multivariate analysis deals with datasets that involve multiple interdependent variables. The video briefly touches on MANOVA (Multivariate Analysis of Variance) and Hotelling's T-squared test as examples of multivariate statistical methods.

Highlights

Introduction to the topic of statistics, specifically parametric and nonparametric statistics.

Explanation of when to use parametric tests if the data follows a normal distribution and nonparametric tests otherwise.

Differentiation between parametric and nonparametric statistics based on the scale of data: ratio and interval for parametric, nominal and ordinal for nonparametric.

Description of data scales: ratio, interval, nominal, and ordinal, with examples for each.

Mention of the reference book 'Dasar Metodologi Penelitian Kuantitatif, Kualitatif dan Statistika' by Sarmanu (2017) as the source of the material.

Clarification on the use of t-tests or z-tests for parametric data analysis and Wilcoxon or Mann-Whitney tests for nonparametric data.

Discussion on the use of ANOVA for comparing more than two treatments with normally distributed data.

Introduction of Kruskal-Wallis or Friedman tests for non-normally distributed data when comparing more than two treatments.

Explanation of using chi-square tests for the difference between two means when data is nominal.

Mention of multivariate analysis for data with more than one variable, such as MANOVA or Hotelling's T-squared test.

Discussion on regression analysis for the impact of variables on a ratio or interval scale and its nonparametric alternatives.

Introduction of logistic regression for the impact of variables on nominal data.

Explanation of Pearson correlation for the relationship between variables on a ratio or interval scale with normal distribution.

Introduction of Spearman's rank correlation for relationships between ordinal variables.

Discussion on multiple correlation tests for relationships involving more than one independent variable.

Introduction of canonical correlation for relationships between variables with more than one dependent and independent variable.

Explanation of univariate and multivariate tests, including examples like t-tests for univariate and Hotelling's T-squared for multivariate.

Acknowledgment of the video's background noise due to a technical issue with the camera overheating.

Conclusion emphasizing the practicality of nonparametric tests when data does not follow a normal distribution.

Call to action for viewers to help identify and discuss unspecified nonparametric tests if they have references or books.

Closing remarks with a New Year's greeting and an invitation for viewers to join in further statistical discussions.

Transcripts

Hai baik A1 mualaikum warahmatullahi

wabarakatuh Selamat malam teman-teman

semua dimanapun anda berada baik di awal

tahun 2021 ini ya saya akan kembali

share kepada temen-temen semua mengenai

statistika ya Jadi pada kesempatan ini

saya akan sharing mengenai statistika

parametrik dan statistika nonparametrik

Nah jadi di video ini nanti kita akan

tahu jika ya Eh data yang akan kita

analisis itu tidak bisa kita menggunakan

uji parametrik maka uji no

tak matic apa yang dapat kita gunakan

Jadi mohon maaf saya agak sariawan jadi

suara saya agak berbeda ya C ini

materinya adalah statistika parametrik

dan statistika non parametrik ya Nah ini

adalah referensinya jika temen-temen

ingin mengutip Ya materi yang ada yang

saya sampaikan di video ini ini berasal

dari buku sarmanu tahun 2017 judulnya

dasar metodologi penelitian kuantitatif

kualitatif dan statistika ya Nah sebelum

kita masuk ke Apa itu statistika

parametrik dan nonparametrik kita review

lagi apa itu statistika ya jadi

statistik sendiri adalah ya Ilmu yang

membahas tentang bagaimana

acara penyajian data pengolahan data dan

juga bagaimana untuk menarik

kesimpulannya nah dalam penarikan

kesimpulannya ya that I've itu dibagi

menjadi dua yaitu statistika deskriptif

dan statistika inferensial ya Nah

statistika deskriptif Yusni adalah

statistika dimana kita menyajikan data

itu agar lebih mudah dimengerti oleh

orang lain Jadi data hasil penelitian

kita itu bisa kita buat dalam bentuk

tabel bisa kita buat dalam bentuk grafik

sehingga orang dengan melihat tabel

membaca tabel dan melihat grafik kita

sudah paham Apa maksud dari data

penelitian kita Nah kemudian ada

statistika inferensial bedanya kalau

statistika deskriptif kan kita menarik

kesimpulan itu hanya berdasarkan grafik

atau berdasarkan tabel saja tapi

baru-baru statistika inferensial kita

menggunakan metode analisis data nah

metode analisis datanya dibagi menjadi

dua yaitu metode analisis data

menggunakan statistika parametrik dan

mengedan statistika nonparametrik Kapan

sih kita menggunakan statistika

parametrik dan kapan kita menggunakan

statistika non parametrik untuk

statistika parametrik itu data

penelitian kita itu dia skalanya adalah

skala rasio dan skala interval Nah kalau

nonparametrik itu adalah skalanya adalah

skala nominal dan skala ordinal ya kalau

parametrik itu syarat wajibnya data

harus berbisnis ribu si normal kalau

nonparametrik data tidak berdistribusi

normal jadi saat data tidak berisi B

berdistribusi normal kita menggunakan

uji nonparametrik nah data

shio atau interval yang tidak

berdistribusi normal dia diuji dengan

menggunakan uji non parametrik Nah tadi

kan ada skala skala data nah ini saya

akan mengulang jenis skala data dan

contohnya pertama itu adalah skala rasio

ya skala rasio itu adalah setelah data

yang bersifat Absolut atau mempunyai

angka nol mutlak contohnya itu adalah

berat badan tinggi badan ya Nah Mengapa

disebut disebut memiliki nilai nol

mutlak karena tinggi badan itu paling

tidak atau paling-paling kecil nol Ya

tidak ada tinggi badan negatif berat

badan juga sama berat badan itu paling

kecil nol tidak ada berat badan negatif

ya Nah sedangkan skala skala interval

itu dia tidak bersifat Absolut artinya

Dia memiliki nilai negatif contohya

skala interval itu

soalnya adalah suhu ya suhu itu ada

nilai negatifnya negatif 5 negatif satu

gitu ya tuh tetap ada artinya berarti

dia adalah skala interval ya kemudian

nilai ujian juga termasuk pada skala

interval karena dia memiliki rentang

0-100 begitu ya Oke kemudian skala

nominal Apa itu skala nominal ya skala

nominal dibalas Skala yang berbentuk

kategori namun dia tidak memiliki

tingkatan atau tidak memiliki strata

contohnya itu jenis agama jenis kelamin

hobi ya Misalnya kalian mengumpulkan

data jenis agama keren bikin di situ

satu agama Islam dua agama Kristen

sehingga agama Hindu 4 agama budaya gitu

ya jadi sini 1234 ini Dia tidak memiliki

kasta ya Tidak ada yang lebih tinggi

walaupun

ada disitu 1234 ada angkanya tidak

ber-ac A4 lebih tinggi dari angka 1 itu

maksud dari skala nominal kalau skala

ordinal ya Ia berbentuk kategori tapi

dia memiliki strata nah ini biasanya

data-data jika kalian mengumpulkan data

dengan kuesioner atau angket kalian

menggunakan skala likert contohnya

kurang cukup baik sangat baik kurang

kalian military angka satu cukup kalian

beri angka2 baik kalian beli angka 3 dan

sangat baik kalian memberi angka 4 angka

4 disini artinya lebih tinggi dibanding

dengan angka 1 2 dan 3 ya Nah sekarang

kita masuk ke uji parametrik dan ujan

uji nonparametrik ya Nah pertama jika

teman teman mau melakukan analisis data

korban Daan dua rata-rata skalanya skala

rasio atau interval

seks kalau dia berdistribusi normal dia

kita menggunakan uji parametrik

menggunakan uji-t atau uji-z ya kalau

dia tidak berdistribusi normal

menggunakan uji wilcoxon atau uji

mann-whitney jadi jangan khawatir Jika

data teman-teman yang jangkauannya itu

di metodenya menggunakan uji-t agar

uji-z saat diuji normalitasnya tidak

normal maka teman-teman bisa menggunakan

uji wilcoxon atau uji mann-whitney u

Hai kemudian kalau skalanya ordinal ini

perbedaan dua rata-rata setelah ordinal

Ya jelas kalau segalanya sudah ordinal

langsung menggunakan uji wilcoxon atau

uji mann-whitney kemudian perbedaan

lebih dari 200 maksudnya adalah di sini

ada tiga perlakuan yang kalian

bandingkan kalau tadi kan hanya dua

Perlakuan di sini ada tiga perlakuan Nah

itu menggunakan uji Anova kalau dia

berdistribusi normal ya Nah kalau dia

tidak berdistribusi normal menggunakan

uji kruskal-wallis atau uji trik Man ya

Nah kemudian jika ya karena ingin

membandingkan lebih dari tiga perlakuan

skalanya ordinal itu menggunakan uji

kruskal-wallis atau uji treatment ya Nah

perbedaan dua rata-rata skalanya nominal

ya Nah kalau skala nominal menggunakan

uji t

kwe atau kwadrat uji magmar atau uji

cover kemudian perbedaan multivariat

multivariat itu artinya ya variabel

tergantung nya atau variabel yang diukur

ya itu lebih dari satu ya Nah kalau dia

bergizi busi normal menggunakan uji

Manova manopause sendiri kepanjangan

dari multivariat anak multivariate

Analysis of variance ya atau menggunakan

uji hotelling kalau datanya tidak berisi

busi normal disini tidak dijelaskan

ganti dari uji Manova atau hotel yang

itu uji apa di teman-teman silakan kalau

ada yang sudah membaca referensi

silahkan komen ya nanti ya Nah kalau

perbedaan multivariat data nominal atau

ordinal ini menggunakan uji

nonparametrik tapi diri buku ini tidak

dijelaskan uji apa uji nonparametrik nya

ya

Hai Nah kalau penelitiannya mengenai

pengaruh variabel terhadap variabel lain

pengaruh ya ini datanya skala rasio atau

skala interval itu kita menggunakan uji

regresi dan uji save jika datanya

berdistribusi normal jika datanya tidak

berdistribusi normal tidak dijelaskan ya

ganti uji dari regresi itu apa aldya

Hai Nah kalau penelitian pengaruh ya

Fariq terhadap variabel lain datanya

nominal Nah kalau datanya nominal ini

langsung menggunakan regresi logistik ya

kemudian ini penelitian hubungan

antarvariabel ya kalau variabel bebas

dan variabel terikatnya hanya satu ya

skalanya itu rasio atau interval

menggunakan korelasi pearson ya kalau

dia berbisnis berdistribusi normal kalau

tidak berdistribusi normal tidak tidak

tahu ya tidak ada dijelaskan ganti dari

uji korelasi Vixion untuk uji non

parametrik yaitu tidak dijelaskan

menggunakan uji apa ya Nah kalau

penelitian hubungan antarvariabel ya

Variabel terikat dan bebasnya hanya satu

datanya skala ordinal kita menggunakan

korelasi rank spearman rank spearman

nge-review kalau penelitiannya ya

peningkatan sini adalah hubungan

antarvariabel lagi tapi di sini variabel

bebas lebih dari satu dan Variabel

terikat hanya satu jadi variabel keknya

ada lebih dari satu variabel ininya ada

satu maka menggunakan uji korelasi

berganda jika dananya normal jika tidak

normal juga tidak dijelaskan di buku ini

ya kemudian penelitian hubungan antar

variabel variabel bebas dan Variabel

terikat lebih dari satu ya Nah kalau

datanya berdistribusi normal menggunakan

korelasi kanonikal Ya saya juga belum

tahu nanti saya akan pelajari dan saya

akan juga bahas apa itu korelasi

kanonikal kalau dia tidak berdistribusi

normal tidak dijelaskannya menggunakan

uji apa ya Nah ini ada bonus ada uji

Yuni varietas dan uji multivariat Apa

itu kalau uji

Hai univariat itu adalah ujian digunakan

untuk mengolah data yang jumlah variabel

terikatnya hanya satu ya jika contohnya

uji-t gitu ya variabel terikatnya ada

satu kalau uji multivariat itu variabel

terikatnya lebih dari satu contoh uji

hotelling uji multivariat analisis atau

menopause kemudian uji same structure

question modelling sama korelasi

kanonikal ya oke teman-teman eh Demikian

Ya video saya kali ini Jadi teman-teman

jika teman-teman melihat ada suara kipas

angin yaitu memang sengaja karena

ternyata saya sudah menemukan penyakit

di kamera saya itu adalah dia itu sukat

atau sering terjadi overheating ya jadi

dia sering mati sendiri jika durasi

hanya lebih dari 10 menit kadang saya

untuk menjelaskan video statistika ini

saya membutuhkan waktu pedang lebih ya

dari 10 menit ya jadi saya terpaksa

menggunakan kipas angin ya tapi tidak

apa-apa ini tidak mengurangi makna dari

video yang telah yang saya buat Semoga

dengan informasi yang saya berikan di

video ini teman-teman jadi paham ya dan

tidak khawatir lagi jika datanya tidak

berdistribusi normal Ternyata kita bisa

menggunakan uji non parametrik Nah tadi

ada beberapa uji yang tidak dijelaskan

Jika dia tidak berdistribusi normal

menggunakan uji nonparametrik nya uji

apa gitu ya teman-teman Tolong bantu

saya jika teman-teman punya referensi

atau punya bukunya Coba kita bahas

bersama kita serkan di channel lebih

statisika sehingga dapat menjadi

informasi atau sumber belajar bagi kita

semua

I make key demikian video saya kali ini

ya Saya ucapkan selamat tahun baru ya

Semoga di tahun ini kita dapat jadi

pribadi yang lebih baik lagi ya oke

demikian video ini saya akhiri

wassalamu'alaikum warahmatullahi

wabarakatuh dan selamat malam

hai hai