Part I. Data Description.
1.1. Original Data.
Due to the recent restructure of the company that resulted in personnel redundancy, the top management of JSC “TESS” * requested to investigate the average level of the job satisfaction (JSL) of the employees. To carry out the analysis we have collected 203 questionnaires (Appendix _) as a sample of total population. Total population: 1200 employees.
The main objectives of the analysis are:
* To compare the mean JSL with the top-management statement;
* To find out if there is a significant difference in mean JSL between the groups of employees with different work experience at the company.
* To investigate if the mean JSL ...view middle of the document...
After the data revision there are 194 cases left in the dataset. Although the distribution is still negativeley skewed we may observe the distribution is closer to normal in terms of kurtosis. (Appendix 2, Picture 1a, Table 1c). We checked the significance of non normal distribution by comparing the numeric value of kurtosis with twice the Std. Error of kurtosis. Looking at the range from minus twice the Std.Error of kurtosis to plus twice the Std.Error of kurtosis, we see that the kurtosis value falls within this range. Thus the non normal distribution is considered to be insignificant.
The JSL variable was also tested for the the distrubution normality depending on “Branch” and “Work Exp” variables. (Appendix, Histogram). Descriptive statistics of the subsamples shows that the skewness and kurtosis is acceptable , no outliers were detected, the mean and the meadian are close. As there are no adjustment needed the subsamples data distribution is accepted as close to normal.
1.3. Summary Statistics.
The analysed variable, i.e. the job satisfcation level (JSL), was meassured by its dispersion contrasted with central tendency and the symmetric bell-shaped distribution curve. Looking at the mean and median we may say that the distribution is not semetric. But we assum the distribution is close to normal as the deviation is not very significant. The distribution shape of the dataset is leptokurtic, kurtosis is -0,326 and skewness is -0,625. However we assume the distribution is close to normal.
We may observe that the data distribution depending on the “Branch” variable is skewed and leptokurtic as well. The mean JSL equals to 3,47 for variable 1, 3,52 for variable 2 and 3,76 for variable 3. (Appendix_, Table_)
We analysed total of 194 cases with the confidence interval of 95% that showed the mean job satisfaction level ranges between 3.41 and 3.74. With the confidence interval of 99% the mean lies between 3.36 and 3.79. Taking into consideration that we are analysing the preference, the precision is very important. Thus the confidence interval of 99% will be used for further analysis.
Descriptive Statistics |
| N | Minimum | Maximum | Mean | Std. Deviation | Skewness | Kurtosis |
| Statistic | Statistic | Statistic | Statistic | Statistic | Statistic | Std. Error | Statistic | Std. Error |
Job Satisfaction | 194 | 1 | 5 | 3,58 | 1,150 | -,625 | ,175 | -,326 | ,347 |
Valid N (listwise) | 194 | | | | | | | | |
1. DATA ANALYSIS
2.4. Hypothesis Testing of a Single Mean
Reffering to the 99% confident interval the mean job satisfaction level lies between 3.36 and 3.79. The top management of the company states that the mean level of the job satisfaction equals “4” provided the scale range is from 1 to 5. The hypothesis test is to be carried out to disprove if the claimed satisfaction level.
1. The null hypothesis: “The mean job satisfaction level equals 4”.
H0: µ = 4