How to use the correlation and simple linear regression to test

the strength of the relationship (or association) between 2

continuous variables.

The graph below is a simple linear regression to test the relationship between Sodium (Na) and BP
(Blood Pressure).

You can see that there really are no obvious outliers for this. The dotted red line is an estimate of the
line of best fit through all points. Notice the majority of the points form a reasonably straight line which
moves up and to the right or in a positive direction. Just looking at this scatter dot you could guess that
the correlation will be strong and positive.

©2014 Walden University Academic Skills Center

Next you fit the linear regression model. Now you have to designate one of your variables as dependent
and one as independent. Based on the original researcher’s hypothesis and/or research question the
variable for blood pressure is the dependent one. In this case that makes sense as it is not likely that
your blood pressure will have any effect on your sodium levels or intake. If you only had the labels of y
and x, y is the dependent variable and x is the independent variable. To do this in SPSS use
Analyze > Regression > Linear
Move bp into the box labeled y and Sodium into the box labeled x. Notice that you can have several x
variables but only 1 y variable. You will get several tables of output but the following two are the ones
you need to focus on here:
Model Summary

Adjusted R Std. Error of

Model R R Square Square the Estimate

1 .920a .847 .827 6.163

a. Predictors: (Constant), Na (g)

Coefficientsa

Model Standardize
Unstandardized d 95.0% Confidence
Coefficients Coefficients Interval for B

Lower Upper
B Std. Error Beta t Sig. Bound Bound

1 (Constan -280.910 68.541 -4.098 .003 -438.965 -122.854

t)

Na (g) 64.010 9.636 .920 6.643 .000 41.789 86.231

a. Dependent Variable: Blood pressure (mm Hg)

©2014 Walden University Academic Skills Center

Interpretation:
• This demonstrates a strong positive correlation with R = 0.920, approaching 1.
• The coefficient of determination is 0.847 or R2. This means that 84.7% of the variability in BP,
can be explained by regression on sodium intake. In common terms, about 85% of the
differences you observe in a person’s blood pressure are due to the level of sodium in their
blood (or diet). These types of results are why many physicians restrict sodium intake in people
with high blood pressure.
• The notice that the y intercept is -280.910 and the slope representing the relationship with
sodium is 64.010. This yields an equation of Y = -280.910 + 64.01x. Thus for every 64.01g
increase in Na expect the Blood Pressure to increase 1 mm/Hg.
• Increased sodium appears to be a significant predictor for increased blood pressure. The strong
correlation is shown below with the r of .920, significance < 0.001. The ANOVA of the regression
(last table)also demonstrates a statistical significance of <.001

Correlations

Blood pressure
Na (g) (mm Hg)

Na (g) Pearson Correlation 1.000 .920**

Sig. (2-tailed) .000

N 10.000 10

Blood pressure (mm Hg) Pearson Correlation .920** 1.000

Sig. (2-tailed) .000

N 10 10.000

ANOVA

Model Sum of Squares df Mean Square F Sig.

1 Regression 1675.776 1 1675.776 44.125 .000a

Residual 303.824 8 37.978

©2014 Walden University Academic Skills Center

 Total 1979.600 9

a. Predictors: (Constant), Na (g)

b. Dependent Variable: Blood pressure (mm Hg)

©2014 Walden University Academic Skills Center