Partial Correlations With IBM SPSS

Introduction

As the earlier implies, the correlation is carried out on the variables partially. So, from the above statement, we can determine the relationship between any given variable using the partial correlations in SPSS. Partial correlation is different from normal correlation as partial correlation takes account of the effect of other variables. One of the advantages of partial correlation is that when normal correlation tells you that two variables are not correlated if you carry out partial correlation you may be shocked of the result.

A scenario

A popular radio talk show host has just received the latest government study on public health care funding and has uncovered a startling fact: As health care funding increases, disease rates also increase! Cities that spend more seem to be worse off than cities that spend less.

Unbelievably, according to government research on public health care expenditures, it is said that health care funding rises, sickness rates likewise increasing funding seems to leave cities in worse shape than less funding.

Data

Does government funding increase the citizen’s illness rate concerning health care? Is this the case

There is not likely to be a correlation between illness rates and health care spending. If the figures are accurate, then further elements that might provide the impression of a link when none truly exists?

Well, let us find out from the partial correlation analysis. Most importantly, scale variables should be used for each variable.

This illustration makes use of the data file health_funding.

Running the Analysis

From the dataset link above, download the data as CSV, then import data into Spss using the comma (“ , ”) as a delimiter; then, to get partial correlations Option From the Spss menus, select: Analyze => Correlate => Partial as shown in figure 1.

Figure 1: Partial Correlation in SPSS

The next step is to choose the variables Reported disease rate, Health care funding and select Visits to health care providers as the control variable, then click (check) Zero-order correlations and click Continue to return to Partial Correlations dialogue afterwards, and click OK to run the procedure, as shown in figure 2.

Figure 2: Variable Partial Correlation

Interpretation

The Partial Correlations table in this illustration displays the zero-order correlations and correlations between all three variables without considering any control factors and the partial correlation between the first two variables, which accounts for the effects of the third variable.

However, from figure 3, the green highlighted cell, the partial correlation that accounts for the frequency of doctor visits is negligible (0.013) and not statistically significant (p = 0.928).

According to one explanation of this data, the reported positive “relationship” between health care funding and illness rates is caused by underlying connections between each factor and the frequency of visits to medical facilities.

In figure 3, the red highlighted cell shows that the zero-order relationship between health care spending and illness rates is both statistically significant (p < 0.001) and relatively strong (0.737). This suggests that sickness rates seem to rise as money grows for healthcare.

Figure 3: Correlations tables

From figure 3, the yellow highlighted row, returning to the zero-order correlations, it is clear that the control variable, the frequency of visits to healthcare providers, is substantially positively correlated with the rates of reported disease and the funding for healthcare.

Conclusion

However, it is essential to know that if we need to check the relationship between two or more variables, we can do such using the partial correlation analysis in Spss once the measures are in the correct data type.

Finally, from the above analysis, it is evident that more funding from the government correlates positively with the number of patients that visit the hospital to receive treatment. This could be that people are motivated to seek the doctor’s intervention when they realized that the government have funded hospitals.

Related blog: Pearson Correlation Analysis