Analyzing and interpreting correlations, correlation coefficients

Question 1

What are correlations and what do they show?

Answer 1

Correlations are a technique for analyzing the strength of the relationship between variables, called co-variables. The data for a correlation is usally obtained from a non-experimental source such as a survey.

The analysis will show one of three things…

• a positive correlation, as one variable increases so does the other one.
• a negative correlation, as one variable increases the other decreases.
• no correlation, there is no relationship between the two variables.

You can measure the strength of a correlation between -1 and 1. 0 means there is no correlation, 1 a strong positive correlation and -1 a strong negative correlation.
The strength of a correlation is known as a correlation coefficient.

Question 2

Correlation coefficient

Answer 2

Often times a statistical test, such as Spearman’s Rho and Pearson’s r, is used to analyze correlational fata to decide whether to accept of reject a hypothesis.

When analyzing the correlational data with these tests , you will get a number known as the observed or calculated value. You compare this against a critical value in the relevant table for the test you are conducting.

The observed value is the same of correlation coefficient and is measured from -1 to 1.

Question 3

Advantages of correlations

Answer 3

1) The technique allows psychologists to establish the strength of the relationship between two variables and measure it precisely.
2) It allows experimenters to investigate things that could not be manipulated experimentally for ethical or practical reasons, ie someone’s age or wealth.
3) Once a correlation has been conducted predictions can be made about one of the variables based on information on the other variable.

Question 4

Disadvantages of correlations

Answer 4

1) Correlational analysis cannot show cause and effect; we cannot tell which variable influences the other.
2) Even if there is a correlation between two variables it may be the case that the variables are not actually related but a third unknown variable influences them both (confounding variable).
3) Correlations can only measure linear relations and does not detect curvilinear relationships. This is when a relationship may be positive up to a point and then it becomes negative or vice versa.