Materi Analisis Data Kategori
CATEGORICAL
DATA ANALYSIS
Eni Sumarminingsih, S.Si, MM
CATEGORICAL RESPONSE
DATA
A categorical variable is one for which the
measurement scale consists of set of categories
Ex :
political philosophy : liberal, moderate,
conservative
Smoking status : never smoked, former smoker,
current smoker
Recovery from an operation : completely
recovered, recovered, nearly recovered, only
somewhat recovered, not at all recovered
Response/Explanatory
Distinction
Most statistical analyses distinguish between
response (or dependent) variables and
explanatory (or independent) variables
This course is concerned with the analysis of
categorical response variables.
For some methods, the explanatory
variables must also be categorical ; for
others, they can be continous or
categorical
Measurement Scala
Distinction
Nominal :
categorical variable for which levels do not have
a natural ordering. Ex: religious affiliation, mode
of transportation, choice of residence, race,
gender, and marital status.
Ordinal :
categorical variable for which do have ordered
levels. Ex : social class, size of automobile
Interval :
variables that have numerical distance between
any two levels for the scale. Ex : blood presure
level, income
Continous/Discrete
Distinction
Variables are classified as continous or
discrete, according to the number of value
they can attain
Quantitative/Qualitative
Disticntion
Qualitative : distinct levels differ in quality,
not in quantity (nominal variable)
Quantitative : disticnt levels have differing
amounts of the characteristic of interest
(interval variable)
What we learn in this
course
Descriptive and inferential methods for
bivariate categorical response
Basic measures of association
Develops the basic of model building
Logit model : used with binomial or multinomial
response
Loglinear model : used with Poisson response
Refference
Categorical Data Analysis
Alan Agresti
University of Florida
Gainesville, Florida
John Wiley & Son
Source from Internet
DATA ANALYSIS
Eni Sumarminingsih, S.Si, MM
CATEGORICAL RESPONSE
DATA
A categorical variable is one for which the
measurement scale consists of set of categories
Ex :
political philosophy : liberal, moderate,
conservative
Smoking status : never smoked, former smoker,
current smoker
Recovery from an operation : completely
recovered, recovered, nearly recovered, only
somewhat recovered, not at all recovered
Response/Explanatory
Distinction
Most statistical analyses distinguish between
response (or dependent) variables and
explanatory (or independent) variables
This course is concerned with the analysis of
categorical response variables.
For some methods, the explanatory
variables must also be categorical ; for
others, they can be continous or
categorical
Measurement Scala
Distinction
Nominal :
categorical variable for which levels do not have
a natural ordering. Ex: religious affiliation, mode
of transportation, choice of residence, race,
gender, and marital status.
Ordinal :
categorical variable for which do have ordered
levels. Ex : social class, size of automobile
Interval :
variables that have numerical distance between
any two levels for the scale. Ex : blood presure
level, income
Continous/Discrete
Distinction
Variables are classified as continous or
discrete, according to the number of value
they can attain
Quantitative/Qualitative
Disticntion
Qualitative : distinct levels differ in quality,
not in quantity (nominal variable)
Quantitative : disticnt levels have differing
amounts of the characteristic of interest
(interval variable)
What we learn in this
course
Descriptive and inferential methods for
bivariate categorical response
Basic measures of association
Develops the basic of model building
Logit model : used with binomial or multinomial
response
Loglinear model : used with Poisson response
Refference
Categorical Data Analysis
Alan Agresti
University of Florida
Gainesville, Florida
John Wiley & Son
Source from Internet