RESEARCH METHOD
The research method which will be used in this paper is explanatory study. In this paper, the writer wants to test the theory of job stress. There are eight independent variables and one dependent variable in this research. The independent variables are work-life balance, resource and communication, work relationship, overload, job security, job characteristics, job control, and pay and benefits. The dependent variable is job stress. the type of data which will be used is the nominal scales, ratio scales and interval scales; and the data measurement used is the rating scales, to be exact Likert scale. There are three screening questions in the questionnaire which are gender, age group and marital status. Gender and marital status will use nominal scale while age group will use ordinal scale. In this research, the 5 level Likert scale will be used since the 5 level Likert scale is proven to be enough to gather necessary data. The use of 5 level Likert scale also common in the social studies. This research will use the simple random sampling. The population will be housekeeping department in Hotel ABC. The respondents will be the employees who already have contract with Hotel ABC. The questionnaire will be distributed by giving it to the admin in the housekeeping department. The admin will then distribute it to the employees randomly. Using the formula developed by Harris (1985), the sample size that is targeted in this research is 58 employees. Ghozali (2011) explains to carry out validity test, the r-value of each variable’s indicator will be compared with the r-table. The r-value generated from each variable must be greater than the r-value from the r-table with the degree of freedom (df) = n – 2 (n is number of sample). To measure the
reliability, the writer will use the Cronbach’s Alpha test. If the value is larger then 0.6, data is considered reliable and can be used for further test. The statistical method which will be used by the writer is the multiple linear regression. Multiple linear regression is used to analyze the impact of the independent variables towards the dependent variable. Multiple linear regression has four assumption which must be fulfilled by the researcher (Osborne & Water, 2002). When these assumptions are not met, the results of the statistical method may not be trustworthy. Ghozali (2011) suggests that the simplest way to examine residuals’ normality is by examining histogram. Normal distribution will create a straight diagonal line, while the observed data will be represented by the plots. The data are said to pass normality test when the observed data are following the straight diagonal line. The data will also be tested for normality using KolmogorovSmirnov Test to further justify the distribution of data. The data is said to pass the normality test if the significance value shows a result above 0.05 (Priyanto, 2012). This research will omit autocorrelation test since autocorrelation is usually used for a time series study or longitudinal study. This research is an explanatory study, where the data has no natural order, autocorrelation test will be omitted (Armstrong, 2001; Doane & Seward, 2011). Multicollinearity can be detected by using the correlation table between independent variables. A way to check is by using tolerance value and VIF (Variance Inflation Factor). The cutoff value usually used is when the tolerance value is below 0.10 or the VIF is above 10. The purpose of heteroscedasticity test is to check whether the variance inequality of residuals between one observation to another (Ghozali, 2011). A way to check heteroscedasticity is by using Breusch-Pagan Test and Koenker Test. The Koenker Test will be used since it is more appropriate for small sample size. The test will be conducted using syntax code developed by Granero (2002) through SPSS. The purpose of F-Test is to determine whether independent variables included in the multiple linear regression model give significant influence to the model simultaneously (Ghozali, 2011). The criteria used in this research to see whether the independent variables have significant influence toward dependent variable simultaneously is the significance F (P-value). The hypothesis tested is as follows:
H0: β1 = β2 = β3 = ... = βk = 0 H1: β1 ≠ β2 ≠ β3 ≠ ... ≠ βk ≠ 0 If the significance F (P-value) is lower than 5%, then it means the independent variables have