For this assignment, you will use the six-step hypothesis testing process (noted below) to run and interpret a correlation analysis using SPSS. The following vignette will inform you of the context for this assignment. A data file is provided in the week’s resources for use in this assignment. Also review the section on Presentation of Statistical Results and Explaining Quantitative Findings in a Narrative Report in the NCU School of Business Best Practice Guide for Quantitative Research Design and Methods in Dissertations.

A manager is interested in better understanding job satisfaction by studying the associations between a number of variables. These variables are age, years of experience, level of education, employee engagement, job satisfaction, and job performance levels.

**Part 1**

She thinks there is a relationship between job satisfaction and

years of experience

educational level

employee engagement

job performance

State the null and alternative hypotheses.

Identify critical values for the test statistics and state the decision rule concerning when to reject or fail to reject the null hypothesis of no relationship.

Run the Pearson correlation analysis and include the correlation matrix in your assignment response.

Report and interpret the correlation coefficient and p-value for each variable paired with job satisfaction.

Explain what decisions the manager might make using these findings.

Part 2

She thinks that younger employees will perform at a higher level, on average.

State the null and alternative hypotheses sets.

Select the significance level.

Select the test statistics and calculate its value.

Identify critical values for the test statistics and state the decision rule concerning when to reject or fail to reject the null hypothesis.

Compare the calculated and critical values to reach a conclusion for the null hypothesis.

Explain what decisions the manager might make using these findings.

Length: 4 to 6 pages not including title and reference page

References: Include a minimum of 3 scholarly resources.

………………………………………………………….

Hypothesis Testing: Parametric Correlation Analysis

A common form of statistical analysis is the exploration of the relationships, associations, or correlations between variables. These analyses provide important knowledge that helps managers understand workplace phenomenon. However, it must be made clear that simply because two variables share a relationship, we cannot use correlation analysis to say that one causes another.

For example, we have more than 70 years of research available to us indicating that job satisfaction and job performance share what we consider to be a strong correlation. However, it has not been established whether one causes the other.

On the face of it, we might argue that if I am satisfied with my work (i.e., strong emotional attachment and ownership), I may work harder. Conversely, I may have a high level of satisfaction because I perform so well in the first place. “Correlation does not imply causation” is a common caution noted in statistics classrooms (Weiers, 2011).

Correlations can be identified by looking at scatterplots. In the following figure, we can see how the points on the scatterplot trend in one direction or another, or not at all. Correlations can also be measured using a variety of formulas. The most common for normally distributed data is the Pearson Product Moment Correlation, noted in the literature as “r.” The Pearson correlation ranges from -1 (perfectly inverse/negative correlation) to a +1 (perfectly direct/positive correlation). A correlation that hovers around zero shows that two variables are not related. It must be noted that the sample size plays a big role in interpreting correlation values. A sample of 1,000 will reveal statistically significant correlations even though the correlation values themselves might be fairly small.

Perfect correlations, in either direction, are not common. In fact, if two variables are too highly correlated, they are said to be multicollinear. In other words, they measure pretty much the same phenomenon. A good range to identify a moderate to strong correlation that is meaningful regardless of any reasonable sample size would be from -.70 to -.30 or +.30 to +.70.

Reference:

Weiers, R.M. (2011). Introduction to Business Statistics (7th ed.). Boston, MA: South-Western Cengage Learning.