Page 50 - Contributed Paper Session (CPS) - Volume 4

CPS2119 Noor Azlin M. S. et al.
there was a Family Well-Being Index Survey that was carried out by National
Population and Family Development Board of Malaysia (Noor et al. 2014). The
purpose of the study is to develop a set of indicators for measuring the well-
being of families among Malaysian and to produce a composite Index of
Family Well-Being. Confirmatory Factor Analysis (CFA) is used to identify the
significant domains of family well-being which are family relationships,
economic situation, health status and safety, community relationship and
religion/spirituality.
There are several different tools for measuring level of family well-being
depending on the different types of data collected and the scale of
measurement adopted for capturing perception of respondent. Regression
methods such as linear, logistic and ordinal regression are example of useful
tools to analyze the relationship between multiple explanatory variables and
level of family well-being (Fagerland & Hosmer 2016). In this study, the ordinal
regression method is used to model the relationship between the ordinal
outcome variable which is overall level of family well-being with several
demographic and social characteristics variables.
This article is organized as follows. Section 2 explains the ordinal
regression model which has been further discussed by Agresti (2007, 2011)
and Liu & Agresti (2005). Section 3 describes the application of ordinal
regression model on the family well-being data, while the results found are
reported in Section 4. Finally, the conclusion is discussed in Section 5.

2. Ordinal Regression Model
The ordinal regression model is a generalisation of the logistic regression
model, where the dependent variable is ordinal. There are several types of
ordinal regression models which can be described based on the specific
scenario of the data (Mccullagh 1980; Mccullagh et al. 2014; Mckelvey &
Zavoina 1975). However, the aim of this study is to model the dependence of
an ordinal response on discrete or continuous explanatory variables. The
proportional odds model which is considered to summarize the relationship
between the ordinal response and the explanatory variables, is as given in
Equation (1).
Pr ( ≤ |Χ)
′
() = ⌊ ⌋ = − Χ, = 1,2, … , − 1 (1)

Pr ( > |Χ )
Following the ordinal model above, let Y denotes an ordinal response
variable with c levels (1,..,c) which in this case is the level of family well-being
and x=(x_1,x_2,…,x_p)' be the vector of p explanatory variables. The higher
value of ordered response category for family well-being indicates the high
level of satisfaction of family well-being. For this study, x_i consists of
demographic, socioeconomic and social characteristics of the selected
households. The relationship between the response variable and the

39 | I S I W S C 2 0 1 9