The Mean Rating for Job Satisfaction: Business Statistics Using GSS.SAV

Subject: Sciences
Pages: 5
Words: 1115
Reading time:
4 min
Study level: PhD

For the three variables in question, the computed means are 42.5 hours worked per week, 13.2 years of educational attainment, and 3.85 siblings. Nominally, the mean for the first independent variable means that Americans typically work either 40 hours (5 days) or an additional half-day on Saturdays. The fact that the mode and median are both at 40 hours alerts us to the reality (see also Figure 3 overleaf) that a five-day workweek is a norm. That the mean is slightly higher than the mode or median shows that a small number of workers do put in extra time.

Table 1: Comparing Measures of Central Tendency

Statistics
Number of hours worked last week Highest year of school completed NUMBER OF BROTHERS AND SISTERS
N Valid 909 1415 1414
Missing 510 4 5
Mean 42.51 13.23 3.85
Std. Error of Mean .455 .077 .084
Median 40.00 13.00 3.00
Mode 40 12 2
Std. Deviation 13.730 2.895 3.144
Variance 188.514 8.378 9.883

On the matter of educational attainment, a mean and median suggesting that one year of college is representative of the working population are preferable to the mode of 12 years. Barring exceptional circumstances, children do have to stay in school at least until Grade 12. But the truth is that educational attainment is a bimodal distribution because a good number of the youth do go on to take some college and even obtain a degree. And this finding is more important than sadly conceding that families can typically afford to send children only up to Junior College.

The matter of family size is tricky. More often than not, adults of today have two siblings which equates to a family size of three. However, the median reveals that half the sample have more than 3 siblings; the mean of 3.85 siblings reveals the upward pull of a distribution skewed by a substantial number of families with unusually large numbers of children (figure 1 below). A last point in this respect is that sibling count needs to be analyzed by age cohort since the trend toward smaller family sizes makes it likely that the youth have fewer siblings than the middle-aged among us. As well, there are also likely to be significant differences by race.

Comparing Measures of Central Tendency
Figure 1. Comparing Measures of Central Tendency

Job satisfaction is an ordinal variable; the hypothesized independent variables “education” and “age” are scale variables while “rincome” is ordinal. This permits the use of the Pearson correlation coefficient.

Table 2. Correlations Between Job Satisfaction and Independent Variables: Total Sample

Correlations Between Job Satisfaction and Independent Variables: Total Sample
JOB OR HOUSEWORK Age of respondent Respondent’s highest degree RESPONDENTS INCOME FOR LAST YEAR
JOB OR HOUSEWORK Pearson Correlation 1.000 -.089** -.048 -.049
Sig. (2-tailed) .003 .117 .131
N 1081.000 1079 1075 938
Age of respondent Pearson Correlation -.089** 1.000 -.090** .181**
Sig. (2-tailed) .003 .001 .000
N 1079 1417.000 1409 992
Respondent’s highest degree Pearson Correlation -.048 -.090** 1.000 .297**
Sig. (2-tailed) .117 .001 .000
N 1075 1409 1411.000 989
RESPONDENTS INCOME FOR LAST YEAR Pearson Correlation -.049 .181** .297** 1.000
Sig. (2-tailed) .131 .000 .000
N 938 992 989 993.000
**. Correlation is significant at the 0.01 level (2-tailed).

On the whole, the mean rating for job satisfaction is 1.68 or close to “moderately satisfied” (on a four-point scale where 1= “very satisfied” and 2 = “moderately satisfied”). Given a standard deviation of 0.77, two-thirds of the sample would theoretically lie between a job satisfaction rating of 0.91 (non-existent) and 2.45, still in the “moderately satisfied” portion of the scale.

SPSS flagged solely the inverse correlation between age and job satisfaction as being significant at the 0.99 confidence level. This means only one chance in a hundred that the relationship between the two variables could be due to chance. In real-world terms, a correlation of -0.089 is hardly meaningful. Accepting it at face value, the finding suggests that older workers are more likely to be satisfied with their work. This is usually accounted for by greater skill or going up the career ladder. Since this happens to only a minority of workers, the relationship is not especially remarkable.

The findings remain even when switching to Spearman’s rho to account for the ordinal nature of the dependent variable:

Table 3. Spearman’s Rank-Order Correlations

Spearman’s Rank-Order Correlations
JOB OR HOUSEWORK Age of respondent Respondent’s highest degree RESPONDENTS INCOME FOR LAST YEAR
Spearman’s rho JOB OR HOUSEWORK Correlation Coefficient 1.000 -.090** -.057 -.054
Sig. (2-tailed) . .003 .063 .100
N 1081 1079 1075 938
Age of respondent Correlation Coefficient -.090** 1.000 -.090** .211**
Sig. (2-tailed) .003 . .001 .000
N 1079 1417 1409 992
Respondent’s highest degree Correlation Coefficient -.057 -.090** 1.000 .326**
Sig. (2-tailed) .063 .001 . .000
N 1075 1409 1411 989
RESPONDENTS INCOME FOR LAST YEAR Correlation Coefficient -.054 .211** .326** 1.000
Sig. (2-tailed) .100 .000 .000 .
N 938 992 989 993
**. Correlation is significant at the 0.01 level (2-tailed).

For the most part, the relationship among satisfaction, education, and income does not differ materially for men and women. Table 4 (appended) shows that partial correlations controlled for sex do not meet the required significance level of even 0.05 except in the case of the relationship between job satisfaction and income. This stands to reason since earning more money permits one many more comforts and amenities and thus increased job satisfaction.

Table 4. Partial Correlations, Controlling for Gender

Partial Correlations, Controlling for Gender
Control Variables JOB OR HOUSEWORK Age of respondent Highest year of school completed Respondent’s income; ranges recoded to midpoints Respondent’s sex
-none-a JOB OR HOUSEWORK Correlation 1.000 -.064 -.003 -.054 .024
Significance (2-tailed) . .058 .938 .111 .472
df 0 865 865 865 865
Age of respondent Correlation -.064 1.000 .042 .187 .052
Significance (2-tailed) .058 . .211 .000 .129
df 865 0 865 865 865
Highest year of school completed Correlation -.003 .042 1.000 .393 -.023
Significance (2-tailed) .938 .211 . .000 .500
df 865 865 0 865 865
Respondent’s income; ranges recoded to midpoints Correlation -.054 .187 .393 1.000 -.271
Significance (2-tailed) .111 .000 .000 . .000
df 865 865 865 0 865
Respondent’s sex Correlation .024 .052 -.023 -.271 1.000
Significance (2-tailed) .472 .129 .500 .000 .
df 865 865 865 865 0
Respondent’s sex JOB OR HOUSEWORK Correlation 1.000 -.066 -.002 -.049
Significance (2-tailed) . .053 .951 .147
df 0 864 864 864
Age of respondent Correlation -.066 1.000 .044 .209
Significance (2-tailed) .053 . .199 .000
df 864 0 864 864
Highest year of school completed Correlation -.002 .044 1.000 .402
Significance (2-tailed) .951 .199 . .000
df 864 864 0 864
Respondent’s income; ranges recoded to midpoints Correlation -.049 .209 .402 1.000
Significance (2-tailed) .147 .000 .000 .
df 864 864 864 0
a. Cells contain zero-order (Pearson) correlations.