3.1 Model Variables
This research has used Ordinary Least Square (OLS) and multiple linear regressions model to measure the time series data. It showed the relationship between the dependent variable (Female Employment) and the independent variables (Fertility rate, total and GDP per capita). The constant, C, is a proper apparatus. The female employment method stated that Female Employment (Y) consists of Fertility rate, total (F) and GDP per capita (G), represented by:
Y_t = f (F,G) (1)
Y = Female Employment (Persons, Thousands)
F = Fertility rate, total (Births per woman)
G = GDP per capita (Current US$)
We Will Write a Custom Essay Specifically
For You For Only $13.90/page!
The equation (1) substituted in equation (2) and as follows:
Y_t=?_0+??_(1 ) F?_(t-1)+??_2 G?_(t-1)+e_t (2)
t = Time trend, data range from 1990 to 2016 Yearly
e = error term
The equation (2) rephrased with lag different of 1 for Female Employment (Y) to avoid multicollinearity error of Fertility rate, total (F) and GDP per capita (G) into equation (3):
In Y_t=?_0+?_(1 ) ?(In F?_(t-1))+?_2 (?In G?_(t-1))+e_t (3)
3.2 Sources of Data
The yearly time series secondary data for female employment (Y) consists of fertility rate, total (F) are collected from the World Bank Open Data and GDP per capita (G) from OECD Statistics.
3.3 Sample Scope and Time Frame
There are 27 observations in total from 1990 to 2016 yearly for the data estimation period. The econometric analysis tool, Eviews used to analyse the time series data.
3.4 Introductory Analysis
Descriptive statistics summarized features of the collected data. The two types of measures are i) central tendency which includes mean and median ii) dispersion or variability which includes minimum, maximum value, standard deviation, kurtosis and skewness (Vetter, 2017). Refer TABLE 1.
Correlation showed whether the relationship between two variables is linear or non-linear. The coefficient correlation stated the pairs of variables that connection is strong or weak (Bewick, Cheek & Ball, 2003).
The four tests of diagnostic checking include i.) Heteroscedasticty test (White), ii.) Normality test (Jarque-Bera), iii.) Multicollinearity test (Variance Inflation Factor) and iv.) Serial Correlation test (LM) (Gujarati and Porter, 2009).