Using SPSS computerized statistical tool - Research Paper Example

Using SPSS computerized statistical tool Part A Statistics can be best understood if researchers get to understand the basics of statistics. This question will talk about the four measurement scales and types of central tendency mean, median, or mode that can be used to best describe each level of measurement. Nominal scale is called labeling variables, and do not have quantitative value. This level of measurement have no overlap with no numeric significance. Examples include gender, race, and place of residence among others. This level of measurement can be best described by mode, which will give researchers the chance to identify the most appearing name or label (Trochim 12). Ordinal scale is also known as rank sale where variables are placed in a certain orderwhile the difference between the variables is not known. Examples of its application are in Likert scale, such as measuring satisfaction, happiness, and discomfort among others. It can be best described by median. Interval level of measurement is numerical where the order is distinct and there is a difference between the values.Examples include time and temperature in Celsius. However, this level of measurement does have true zero values. Both mode, median and mean can be used to describe the data (Trochim 26). Lastly, ratio scale or level of measurement is characterized by order, exact difference of value between units and having an absolute zero value. Central tendency that can be applied in describing ratio data include median, mode and mean. (2) A measure of dispersion isaimed at measuring how spread out a set of data is. For ratio scale, measures of dispersion that can be used to describe it include standard deviation, interquartile range, range and coefficient of variation. When dealing with the interval level of measurement, the best measure of dispersion that can best describe it include standard deviation, interquartile range and range. There is no known measure of dispersion that can be used to describe nominal and interval level of measurement (Trochim 45). Part B Topic: An investigation into the relationship between absenteeism and job related attitudes. Hypotheses 1. There is a link between absenteeism and the lack of pro work-life policy. 2. Motivated employees take less sick-leave. 3. Companies regulated by collective agreement record less absenteeism. 4. HRM can meet employee motivation, expectations, and, as a result reducing absenteeism. 5. There is a relationship between high staff turnover rate and absenteeism. Multiple choice questions 1. Gender Male Female 2. Nature of employment Full-time Part-time 3. Have you been absent from work in the past 6 months? Yes No 4. How much leave in hours have you taken in the past 6 months for any of the reasons stated below If none, then indicate ‘0’? Family issues and responsibly Voluntary work Medical appointments Personal business Just a day off Travel 5. In 5 level state your level of agreement with the following statement: pro work-life policy in work environment is related to absenteeism. Strongly agree Agree Neutral Do not agree Strongly disagree 6. Based on the current situation, do you feel motivated? Yes No 7. How satisfied are you with the aspect of your job? Very satisfied Satisfied Neutral Dissatisfied Very Dissatisfied 8. What is your level of agreement with the following statement, ‘Companies regulated by collective agreement record less absenteeism’ Strongly Agree Agree Neutral Disagree Strongly Disagree 9. Since you were employed in this organization, how many staff(s) have left voluntarily. 10. Do you see yourself working for this organization in the next 5 years? Yes No Part C Question 1 As discussed previously, descriptive statistics is an important concept that help researchers to summarize large data set and draw meaning from them. Tables, graphs and pie charts have been deemed the best in describing data. From the analysis, it is evident that with regards to highest qualification of the respondents, a greater proportion of respondents had acquired high school education 1003 representing 49.6%, followed by those who were bachelor’s holders 355 representing 17.5% of the sample, then individual with lower high school being 297 representing 14.7% of the sample, those with post graduate came forth with 194 individual contributing to 9.6% of the sample and lastly those who have junior college qualification were 173 representing 8.6% of the sample (Fig. 1). Fig. 1 Bar Graph of Respondent’s Qualification With regards to general happiness, 54.6% (1100) indicated that they were pretty happy, 29.7% (599) were very happy and only 15.7% (316) indicated that they were not too happy (Fig. 2). Fig. 2 Bar Graph of General Happiness From Fig. 3, 77.1% (1559) were white, 13.9% (281) were black and 9.0% (183) fell under the category labelled other which included Hispanic, among others. Fig. 3 Bar Graph of Respondent’s Qualification Concerning hours worked in the last week, the number of children and total family income, the descriptive table below summarizes it all. It is evident from the analysisthat on average, families had 2children with the biggest family having 8 children while the smallest having no child. On average, respondents worked for 42 hours in the last week, having a standard deviation of 14.480.The highest number of hours worked standing at 89 while the shortest being 1 hour. Based on coding adopted, the maximum and minimum total family income are over $150,000 and under $1,000. Table 1 Descriptive Statistics N Range Minimum Maximum Mean Std. Deviation NUMBER OF CHILDREN 2020 8 0 8 1.94 1.698 TOTAL FAMILY INCOME 1774 24 1 25 16.86 5.745 NUMBER OF HOURS WORKED LAST WEEK 1203 88 1 89 42.00 14.480 Valid N (listwise) 1098 Question 2 From the histogram above, the mean age is 77.7 years with a standard deviation of 17.35 years, meaning that majority fell between 30.36 and 65.06 years. With the normal curve fitted on the histogram, it is evident that the age is normally distributed. Fig. 4 Histogram showing the age of the respondents Question 3 a. The median income of respondents aged 30 years and over who work full-time In order to respond to this question, the researcher selected respondents who fell in the stated categories, that is those who were over 30 years and worked full-time. Based on the coding adopted, the median value of the income of the respondents was between $40,000 and $49,999. It is worth noting that individuals who fell in this category were 704 but 122 did not provide information with regards to their income. Table 2 Statistics RESPONDENT'S INCOME N Valid 704 Missing 122 Mean 16.95 Median 18.00 Std. Deviation 4.600 Minimum 1 Maximum 25 b. The mean socioeconomic index of men versus women To respond to this question an independent Samples T-test is used when the task wishes to determine whether the difference between means for two sets of scores is significant. Table 3 Group Statistics RESPONDENT'S SEX N Mean Std. Deviation Std. Error Mean RESPONDENT SOCIOECONOMIC INDEX MALE 887 49.109 19.4399 .6527 FEMALE 1024 48.458 19.5677 .6115 From table 3, it is evident that the mean differences in terms of social-economic index between males and females is not different. It is apparent that the mean respondents’ social-economic index for male is 49.109 while for female it stood at 48.458. There is no statistical significance in the means between males and females. c. The mean number of hours worked last week of women who work part-time The statistical technique used above will be used in this question. From table 4, it is evident that the mean of women who worked full-time was significantly higher than those women working part-time. The means stood at 41.56 and 28.08 hours for women who worked full-time and part-time respectively. Table 4 Group Statistics WAS RESPONDENT'S WORK PART-TIME OR FULL-TIME? N Mean Std. Deviation Std. Error Mean NUMBER OF HOURS WORKED LAST WEEK FULL-TIME 419 41.56 10.377 .507 PART-TIME 132 28.08 13.170 1.146 d. The mean number of children of black versus white respondents When asked about the number of children based on race of the respondents, it emerged that on average, black had 2 children while the whites had on average one child per family. Table 5 Group Statistics RACE OF RESPONDENT N Mean Std. Deviation Std. Error Mean NUMBER OF CHILDREN WHITE 1557 1.91 1.665 .042 BLACK 280 2.13 1.874 .112 Part D Hypotheses Testing 1. Satisfaction with financial situation is higher among male than female respondents To correctly answer the above hypothesis, a cross-tabulation will be conducted. A cross tabulation is the merging of the frequency distributions of two or more variables in a single table. Ideally, the technique helps in understanding how one variable relates to another variable. Based on the percentages of both males and female responses to each category and by looking at the clustered bar chart, it is evident that a higher percentage of female respondents were not at all satisfied with their financial situation (female=32.0%, male=29.8). On the other hand, a higher percentage of males 29.2% compared to female, 27.7% were satisfied with their financial situation. Can be explained by the fact that male tend to start jobs with higher pay and they have a higher probability of getting a promotion and pay increment at a rate slightly higher than their female counterparts. Table 6 RESPONDENT'S SEX * SATISFACTION WITH FINANCIAL SITUATION Crosstabulation SATISFACTION WITH FINANCIAL SITUATION Total SATISFIED MORE OR LESS NOT AT ALL SAT RESPONDENT'S SEX MALE Count 271 380 276 927 % within RESPONDENT'S SEX 29.2% 41.0% 29.8% 100.0% FEMALE Count 302 439 348 1089 % within RESPONDENT'S SEX 27.7% 40.3% 32.0% 100.0% Total Count 573 819 624 2016 % within RESPONDENT'S SEX 28.4% 40.6% 31.0% 100.0% Fig. 5 Cross-tabulation of financial satisfaction and gender 2. The respondents’ socioeconomic indexis related to age among respondents whose highest qualification is High School Education. Correlation as a statistical technique will be used to test this hypothesis. It was chosen since it is best used when exploring the relationship between two numerical variables. A Pearson correlation analysis was undertaken to determine if there was a relationship between respondents’ socioeconomic indexis related to age among respondents of High School Education qualification. Results indicated that there was NO statistically significant relationship between the two variables (r = 0.056, p = 0.087). The findings suggest that according to individuals with high school education qualification, the age does not determine one’s social-economic index. Table 7 Correlations AGE OF RESPONDENT RESPONDENT SOCIOECONOMIC INDEX AGE OF RESPONDENT Pearson Correlation 1 .056 Sig. (2-tailed) .087 N 1002 953 RESPONDENT SOCIOECONOMIC INDEX Pearson Correlation .056 1 Sig. (2-tailed) .087 N 953 954 3. The number of hours worked last weekis higher among male than female respondents. To answer the hypothesis, independent samples T-test is used since it is best suited to determine the difference between means for two sets of scores is significant. Table 7 Group Statistics RESPONDENT'S SEX N Mean Std. Deviation Std. Error Mean NUMBER OF HOURS WORKED LAST WEEK MALE 636 45.50 15.164 .601 FEMALE 567 38.08 12.574 .528 Table 8 Independent Samples Test Levene's Test for Equality of Variances t-test for Equality of Means F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper NUMBER OF HOURS WORKED LAST WEEK Equal variances assumed 16.594 .000 9.181 1201 .000 7.426 .809 5.839 9.012 Equal variances not assumed 9.279 1.195E3 .000 7.426 .800 5.856 8.996 An independent samples t-test was undertaken to determine whether there was a significant mean difference in the number of hours respondents worked in the last week respondents between males and females. The results indicate there was a statistically significant difference in the number of hours worked last week between the groups (t = 9.279, p = 0.000). This suggests that Males (mean = 45.50) worked for longer hours in the last week compared to females (mean = 38.08). Thus, the hypothesis is supported. 4. Male respondents tend to be employed more on a full-time basis than female respondents. To get statistical backing on such notion, a cross-tabulation was used. According to the percentages of both males and female responses to each category and by looking at the clustered bar chart, it is evident that a higher percentage of males (54.4%) against females (33.1%) were employed full time. On the other hand, a higher percentage of females 66.9% against males 45.6% are employed part-time. This finding warrant acceptance of the hypothesis formulated. Table 9 Respondent's sex * was respondent's work part-time or full-time? crosstabulation WAS RESPONDENT'S WORK PART-TIME OR FULL-TIME? Total FULL-TIME PART-TIME RESPONDENT'S SEX MALE Count 621 104 725 % within WAS RESPONDENT'S WORK PART-TIME OR FULL-TIME? 54.4% 33.1% 49.8% FEMALE Count 521 210 731 % within WAS RESPONDENT'S WORK PART-TIME OR FULL-TIME? 45.6% 66.9% 50.2% Total Count 1142 314 1456 % within WAS RESPONDENT'S WORK PART-TIME OR FULL-TIME? 100.0% 100.0% 100.0% Fig. 6 Cross-tabulation of Status of work and gender 5. When compared to respondents who earn less than $90,000, those who earn $90,000 or more are significantly more likely to believe that taxes on high income people are too high. In order to answer this question, a cross-tabulationis used. However, the respondent's income was recoded into 2 groups, value 1 represented earnings less than $90,000 while 2 represented earning greater than $90,000. Analysis indicates that individuals earning less than 90,000 were more likely to believe that taxes on high income people are too high (Fig. 7). Fig. 7 Cross-tabulation of income and Taxes 6. The respondents’ Fathers’ Socioeconomic Indexis related to the respondents’ total family incomeamong white persons but not among black persons. Descriptive statisticswere used to answer this hypothesis. From the box plot, it is apparent that whites have a significantly higher total family income, followed by those in other race then the black respondents come third. Fig. 8 Box plot of Total family income, race o respondent 7. White respondentsare more in self-employmentwhen compared to black respondents or respondents from other races. A cross-tabulation best answers the question. Based on the percentages of white and black respondent in terms of self-employment and by looking at the clustered bar chart, it is evident that a higher proportion of white are self-employed (86.3%) compared to black (6.2%). Similarly, a greater portion of whites is employed by other compared to black respondents. This can be explained by a number of factors such as discrimination against the blacks or the blacks lack necessary educational skills and experience to be employed. Table 10 Race of respondent * respondent self-employed or works for somebody crosstabulation RESPONDENT SELF-EMPLOYED OR WORKS FOR SOMEBODY Total SELF-EMPLOYED SOMEONE ELSE RACE OF RESPONDENT WHITE Count 195 1319 1514 % within RESPONDENT SELF-EMPLOYED OR WORKS FOR SOMEBODY 86.3% 76.2% 77.3% BLACK Count 14 256 270 % within RESPONDENT SELF-EMPLOYED OR WORKS FOR SOMEBODY 6.2% 14.8% 13.8% OTHER Count 17 157 174 % within RESPONDENT SELF-EMPLOYED OR WORKS FOR SOMEBODY 7.5% 9.1% 8.9% Total Count 226 1732 1958 % within RESPONDENT SELF-EMPLOYED OR WORKS FOR SOMEBODY 100.0% 100.0% 100.0% Fig. 9. Cross-tabulation of Race and Nature of employment 8. White respondents have significantly fewer childrenthan both black respondents and respondents from other races. There has been a notion that white people usually have fewer children when compared to other races, especially the blacks, this question tests whether this is true or false. In doing so, both exploratory and correlation techniques are employed. From the exploratory descriptive table, the average number of children for whites stood at 1.91, blacks were 2.13 and for other races it stood at 1.91. Correlation results show that there is no statistical significance between the race and number of children the respondents have, r=0.020 p=0.380 (Table 11). Table 11 Correlations NUMBER OF CHILDREN RACE OF RESPONDENT NUMBER OF CHILDREN Pearson Correlation 1 .020 Sig. (2-tailed) .380 N 2020 2020 RACE OF RESPONDENT Pearson Correlation .020 1 Sig. (2-tailed) .380 N 2020 2023 9. The number of hours worked last weekis higher among respondents who have no children than among those who have children. It has been assumed for a long time that individuals with no children spend most of their time working, hence work longer times than their counterparts who do not have children. Since the data comes from the same group, a paired sample test was used. From table 12, it is evident that the difference is significant (t = -96.359, p= 0.000). This suggests that respondents who had no child worked for a longer time compared to those who did had a child or children. Table 12 Paired Samples Test Paired Differences t df Sig. (2-tailed) Mean Std. Deviation Std. Error Mean 95% Confidence Interval of the Difference Lower Upper Pair 1 NUMBER OF CHILDREN - NUMBER OF HOURS WORKED LAST WEEK -40.340 14.514 .419 -41.162 -39.519 -96.359 1201 .000 10. The respondents’ condition of health worsens with age. It has been believed that the older one becomes, the poorer their health conditions. To test this hypothesis, a correlation analysis was used. Results in table 13indicated that there was a weak statistically significant relationship between the two variables (r = 0.207, p = 0.000). This means that the older an individual becomes, the higher the chances that their health condition worsen. Table 13 Correlations AGE OF RESPONDENT CONDITION OF HEALTH AGE OF RESPONDENT Pearson Correlation 1 .207** Sig. (2-tailed) .000 N 2013 1347 CONDITION OF HEALTH Pearson Correlation .207** 1 Sig. (2-tailed) .000 N 1347 1351 **. Correlation is significant at the 0.01 level (2-tailed). 11. Total family income is significantly higher among respondents whose spouse is in full-time employment or part-time employment, than among those whose spouse is retired. From the cross-tabulation bar graph, it is evident that individuals who are married, havea higher total income, then followed by those who are never married, then those who have divorced, widowed and lastly those who have separated. However, to test the significance, analysis of variance was conducted. From ANOVA results, since p=0. 000 it means that there is a significant difference between at least two of the groups (Table 14). A Tukey post hoc test indicated that people who have their spouses working on either full-time or part-time basis have a significantly higher total family income compared to those who are retired. Table 14 TOTAL FAMILY INCOME Tukey HSD SPOUSE LABOUR FORCE STATUS N Subset for alpha = 0.05 1 2 3 SCHOOL 8 14.75 OTHER 25 16.64 16.64 RETIRED 110 17.60 17.60 17.60 UNEMPL, LAID OFF 13 18.00 18.00 18.00 KEEPING HOUSE 101 18.58 18.58 TEMP NOT WORKING 24 19.17 19.17 WORKING PARTTIME 87 19.34 19.34 WORKING FULLTIME 467 20.25 Sig. .100 .284 .311 Means for groups in homogeneous subsets are displayed. Fig. 10 Cross-tabulation of Marital status and Total family income 12. Male respondents’ marital status is related to their general happiness. To answer the above hypothesis, paired sample T-test was selected. From the table below the sample mean for each question are indicated in yell as 1.0 and 1.70 for marital arts ad general happiness. It's worth noting that all male respondents we married. In this sample of 469 men, the mean general happiness stood at 1.70 (Table 15). Since t=-24.215, p=0.000, it is evident as there isa relationship between the two variable, hencemen who are married tend to be happier (Table 16). This conclusion could have been made convincing in the sample selected could have had unmarried, divorced or separated male respondents. Table 15Paired Samples Statistics Mean N Std. Deviation Std. Error Mean Pair 1 MARITAL STATUS 1.00 469 .000 .000 GENERAL HAPPINESS 1.70 469 .627 .029 Table 16Paired Samples Test Paired Differences t df Sig. (2-tailed) Mean Std. Deviation Std. Error Mean 95% Confidence Interval of the Difference Lower Upper Pair 1 MARITAL STATUS - GENERAL HAPPINESS -.701 .627 .029 -.758 -.645 -24.215 468 .000 13. Married respondentsare more likely to be in employmentthan non-married respondents. From cross-tabulation analysis, it is evident that respondents who were married were likely to be in employment when compared to their counterparts. Approximately 50.0% and 47.7% of married respondents are employed full-time and part-time respectively (Table 17). Table 17 MARITAL STATUS * LABOUR FORCE STATUS Crosstabulation LABOUR FORCE STATUS Total WORKING FULLTIME WORKING PARTTIME TEMP NOT WORKING UNEMPL, LAID OFF RETIRED SCHOOL KEEPING HOUSE OTHER MARITAL STATUS MARRIED Count 501 96 27 21 169 12 128 18 972 % within LABOUR FORCE STATUS 50.0% 45.7% 50.9% 28.4% 50.6% 21.1% 56.4% 30.0% 48.2% WIDOWED Count 27 12 5 0 94 0 21 5 164 % within LABOUR FORCE STATUS 2.7% 5.7% 9.4% .0% 28.1% .0% 9.3% 8.3% 8.1% DIVORCED Count 157 22 4 8 45 4 24 17 281 % within LABOUR FORCE STATUS 15.7% 10.5% 7.5% 10.8% 13.5% 7.0% 10.6% 28.3% 13.9% SEPARATED Count 43 7 0 4 6 0 4 6 70 % within LABOUR FORCE STATUS 4.3% 3.3% .0% 5.4% 1.8% .0% 1.8% 10.0% 3.5% NEVER MARRIED Count 275 73 17 41 20 41 50 14 531 % within LABOUR FORCE STATUS 27.4% 34.8% 32.1% 55.4% 6.0% 71.9% 22.0% 23.3% 26.3% Total Count 1003 210 53 74 334 57 227 60 2018 % within LABOUR FORCE STATUS 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 14. Older respondentsare more likely to think that it is better for men to work and for women to tend homes. This hypothesis was tested using correlation analysis, generation of a scatter-plot. From, the table below, it is apparent that there is a weak negative significant relationship between age and the idea or thought that it is better for men to work and for women to tend homes, r=-0.155, p=0.000 (Table 18). This suggests that the older one is, the stronger is the thought or the idea that it is better for men to work and for women to tend homes. Table 18 Correlations AGE OF RESPONDENT BETTER FOR MAN TO WORK, WOMAN TEND HOME AGE OF RESPONDENT Pearson Correlation 1 -.155** Sig. (2-tailed) .000 N 2013 1300 BETTER FOR MAN TO WORK, WOMAN TEND HOME Pearson Correlation -.155** 1 Sig. (2-tailed) .000 N 1300 1308 **. Correlation is significant at the 0.01 level (2-tailed). Work Cited Trochim, William. Descriptive statistics: Research Methods Knowledge Base.Oxford University: Oxford University Press, 2003. Print. Read More