describing correlation
Standard Deviation and Correlation Coefficient
Standard Deviation: Measures the dispersion or spread of a set of values in a dataset.
Correlational Studies
Definition: Correlational studies examine two or more variables to determine whether a nonrandom relationship exists between them.
Quantitative Measurement: When both variables are numerical, their relationship's strength and direction can be assessed through the correlation coefficient.
Correlation Coefficient
Calculation: The correlation coefficient is calculated using a formula (as described in the Statistical Appendix).
Range: Values range from -1.00 to +1.00.
A positive correlation indicates an increase in one variable coincides with an increase in the other variable.
A negative correlation indicates an increase in one variable coincides with a decrease in the other variable.
Strength of Correlation: The absolute value of the correlation coefficient (ignoring the sign) reflects the correlation's strength:
Values close to +1.00 or -1.00 indicate a strong correlation, allowing predictions about one variable based on the other.
A correlation close to zero (0) implies no statistical relationship; knowledge of one variable does not help predict the other.
Example Study
Hypothetical Study Overview: Conducted with 10 college students to assess the correlation between their most recent test score and four other variables:
Hours spent studying for the test
Previous test score
Level of psychological depression (measured a day before the test)
Height in centimeters
Data Representation: Collected data was presented in a tabular format (Table 2.2), with each row showing a different student's data, rank-ordered by test scores.
Visualization of Relationships Through Scatter Plots
Purpose: To visualize the relationship between the test score and each of the other variables.
Scatter Plot Representation (Figure 2.2):
Plot A: Relationship between test score and hours of study.
Observation: Higher test scores tend to correspond with more hours spent studying—indicating a moderate positive correlation.
Calculated correlation coefficient: +.51.
Example Point: The point shown by a red arrow, representing a student who scored 85 and studied for 9 hours.
Plot B: Relationship between test score and previous test score.
Observation: Points closely align along an upward slant, indicating a strong positive correlation.
Correlation coefficient: +.93, meaning prior test scores are excellent predictors of new test scores.
Plot C: Relationship between test score and depression.
Observation: This scatter plot shows a moderate negative correlation, indicating that higher depression scores are associated with lower test scores.
Correlation coefficient: -.43.
Plot D: Relationship between test score and height.
Observation: The data appears uncorrelated, with a correlation coefficient near 0. Thus, height offers no predictive power regarding test scores.
Correlation coefficient: -.04.
Inferential Statistics
Nature of Correlation: A correlation is a form of inferential statistics, indicating that statistical variability in the data requires interpretation and understanding of possible relationships.
Importance of Understanding Correlations: Understanding the nature of correlations aids in drawing conclusions from relational data and assists in predicting outcomes.
Imagine you have two friends, and you want to see if they often do things together or if one does the opposite of the other! That's kind of what correlation helps us understand with numbers.
We don't actually 'calculate' it with simple counting, but grown-ups use a special math tool to figure out a "correlation number" between -1 and +1.
If the number is close to +1 (like going up the slide together): It means when one friend does something, the other friend usually does something similar. For example, the more you study (first friend), the better your test scores often get (second friend)! They go up together.
If the number is close to -1 (like one friend goes up, the other goes down): It means when one friend does something, the other friend often does the opposite. For example, maybe the more chocolate you eat (first friend going up!), the less hungry you are for dinner (second friend going down!).
If the number is close to 0 (like two friends doing completely different things): It means what one friend does doesn't really have anything to do with what the other friend does. Like how tall you are (one friend) compared to your test score (another friend) – they don't usually grow or shrink together!