commit to user 34
3. Dependent variable The dependent variable in this research is job performance. Job
Performance is the competence of employees in conducting the managerial activity
includes planning,
investigating, coordinating,
evaluating, supervising, staffing, negotiating and representing.
Job performance was measured with a self-evaluation questionnaire developed by Mahoney et al. 1963. Respondents were asked to respond on a
nine-point Likert-type scale. They were required to rate their performance on eight dimensions – planning, investigating, coordinating, evaluating,
supervising, staffing, negotiating and representing. They were also asked to rate their overall performance. Although such self-evaluation might be
criticized as being a subjective measure, similar techniques have been widely adopted in many studies Brownell and Hirst 1986; Frucot at all 2006;Yuen
2007; Sardjito and Muthaher 2007.
E. Analysis Technique
The data analysis in this research conducted: descriptive statistic, validity and reliability test and hypothesis testing. The computation of analysis technique is
done with SPSS version 16.0.
commit to user 35
1. Descriptive statistic Descriptive statistic consists of the measurement of mean, median,
standard deviation, maximum, and minimum value from each data sample. This analysis means to give the picture of concerning distribution and the data sample
behavior. 2. Validity and reliability test
Research variable measurement using questionnaire instrument must use quality examination of data to obtain the test of validity and reliability. This
examination aims to know whether the instrument used is valid and reliable or not. This is done since the truth of data analysis assesses the quality of research
result. The aim of validity test is to measure the quality of instrument used and
to show the level of validity of the instrument, and also how well a concept can be defined by a size of measurement Ghozali 2006. Instrument can be categorized
as valid if instrument can measure what the researcher want and show the accurate data. Examination conducted by using factor analysis. This research used
Confirmatory Factor Analysis CFA test to measure validity test. Confirmatory Factor Analysis CFA test used to test whether the construct have
unidimentionality or the indicators can confirm the construct or variable. Data is able to be conducted by factor analysis if the Kaiser’s MSA is above 0.5 Ghozali
2006.
commit to user 36
Reliability test conducted by calculating Cronbach Alpha ά to test the eligibility to consistency of all used scale. Instrument told reliable if owning
Cronbach Alpha ά more than 0.60 this is based on the Nunnally criteria Ghozali 2006.
3. Hypotheses testing Hypothesis will be tested by path model for the data analysis. Path
coefficients representing the relationships between variables were estimated by standardizing the
β regression coefficients Yuen 2007. The relationship between the variables in the path model can be stated as two equations, such as the follows:
BP = P
31
NA + P
32
WA+ P
3a
R
i
................................................................ 1 Where, BP : is the participative budgeting,
NA : the need for achievement, WA : the work attitude,
P : the standardized partial regression coefficients path coefficients,
Ri : the standardized residual JP =
P
41
NA + P
42
WA+ P
43
BP +P
4b
R
i....................................................................
2 Where, JP : is the job performance,
NA : the need for achievement, WA : the work attitude,
BP : the participation in budgeting. P
: the standardized partial regression coefficients path coefficients, Ri : the standardized residual
commit to user 37
Before the multiple regression test, the examination must use classical assumption test to ensure the validity of research data, not bias, consistent, and
further it can estimate the efficient regression coefficient Sekaran 2003. a. Multicollinierity test
The goal of multicollinierity test is to test whether the regression model found the correlation between the independent variables. The good
regression model must not have correlation between the independent variables. If the independent variables have correlation, so the variables
are not orthogonal. To detect the multicollinierity, this research used the tolerance value and the VIF Variance Inflation Factors value. To confirm
that multicollinierity not exist in the regression model, if the Tolerance value is above 0.10 and the value of VIF is 10 Ghozali 2006.
b. Heteroskedasticity test The goal of heteroskedasticity test is to test whether in the
regression model there is inequality variance from residual of the certain research to another. If there is fix variance of residual of the certain
research to another, it called homoskedasticity. The good regression model is if there is homoskedasticity. If the graph scartterplots result show that
dot disseminate at random and also spread over on above and also under number 0 zero at axis of the ordinate, this matter inferential that the
heteroskedasticity is not happened at the regression model Ghozali 2006.
commit to user 38
c. Normality test The goal of Normality test is to test whether the regression model,
disturbing variable or residual normally distributed. Researcher uses the analysis of statistic One-Sample Kolmogorov-Sminorv test with the
significant level 0.05, thus the data of residual is normally distributed. If the Histogram and Normal P-P Plot of Regression Standardized Residual
shows that histogram graphic gives the normal pattern distribution and the normal plot graphic shows that the dot spread around the diagonal line and
follow the diagonal line. It means that the histogram graphic show the normal plot distribution thus the regression model is fulfill the normality
assumption Ghozali 2006.
commit to user 39
CHAPTER IV DATA ANALYSIS AND DISCUSSION