Correlational Research Strategies

Correlational Research Strategy

• Correlational research aims to demonstrate a relationship between variables, but it does not establish cause-and-effect.
• Experimental research, in contrast, demonstrates a cause-and-effect relationship.
• Correlational strategy is used to study the relationship between two or more variables.
• Correlation describes the nature of the relationship by specifying the direction and degree of the relationship between the variables.
• Data collection in correlational studies involves measuring, not manipulating variables.
• Methods include observations, surveys, and physiological measures.
• Correlational studies have high external validity but lack the ability to imply causality.

Outline of Correlational Research

• Correlational strategy
• Correlation coefficient
• Measuring correlations
• Strengths and weaknesses
• Timepoint-based correlations

Correlational vs. Experimental Research

• Correlational research focuses on observing relationships between variables, whereas experimental research manipulates one variable to observe its effect on another.

Correlational Strategy Details

• The goal is to examine the relationship between two (or more) variables.
• Correlation describes the nature of the relationship based on direction and magnitude.
• Examples of correlational studies: price of chocolate and quality; caffeine intake and alertness; movie topics and music preferences.

Correlational Strategy Data Collection

• The data collection procedure is correlational, which means no manipulation of variables.
• It involves measuring variables using observations, surveys, or physiological methods.

Correlational Strategy Validity and Causation

• High external validity (reflects natural events).
• Cannot imply causality (does not establish that one variable causes the other).
• Example provided: per capita cheese consumption correlates with deaths by bedsheet tangles (a humorous illustration demonstrating that correlation does not imply causation).

Examples of Correlational Studies

• Price of a box of chocolates and its quality.
• Caffeine intake and alertness.
• Movie topics and music preferences.

Visualizing Relationships: Scatterplots

• Scatterplots graph the relationship between variables.
• Each point represents a measurement.
• Shoe size and IQ score is an example (illustrative, not implying a true relationship).

Important Assumptions of Correlational Data

• Each item/person is represented by only one data point.
• Each point in the dataset is independent of other points; no two points can come from the same individual.

Line of Best Fit (Regression Line)

• A line of best fit summaries the relationship between two variables on a scatterplot.
• The closer the points to the line, the stronger the association between the variables.

Representing a Correlation

• Quantitative representation: Correlation coefficients range from -1 to +1.
• Ordinal variables use Spearman rho (rs).
• Ratio/interval variables use Pearson r.
• Important aspects include form (linear/non-linear), direction (positive/negative), and strength (absolute value between 0 and 1).

Non-Linear Correlations

• Change in one variable is not consistent with a change in another variable.
• The relationship between variables is not linear.

Spearman's Rank-Order Correlation

• Determines strength and direction of monotonic relationship between two variables.
• A monotonic relationship is one where both variables continue in the same direction or remain constant.

When to use Spearman's Rank-Order Correlation

• Used when data is ordinal (not interval or ratio scale).
• The data must be monotonic.

Calculating a Spearman Correlation (Example)

• Demonstrates steps to calculate the correlation coefficient.

When to Use Spearman's Rank-Order Correlation

• At least five pairs of data; preferably more than eight.
• Avoid using ranks with too few or too many tied ranks.

Calculating a Pearson Correlation

• A correlation coefficient describes the relationship between two variables: direction, form, and consistency/strength.
• Most behavioral research uses interval or ratio scale data. By default, correlation will mean Pearson's correlation (r) in behavioral research.

Direction of Correlation

• Positive correlation: larger values of one variable are associated with larger values of another (or smaller with smaller).
• Negative correlation: larger values of one variable are associated with smaller values of another (or smaller with larger).
• No correlation: no consistent relationship exists between the variables.

Form of Correlation

• Linear correlation: data points cluster around a straight line in a scatterplot.
• Nonlinear correlations: data points do not cluster around a straight line (can be monotonic and non-monotonic).

Strength of Correlation

• The degree of association between two variables.
• Expressed as a correlation coefficient ranging from -1.0 to +1.0; closer to +/- 1 indicates a stronger relationship.
• A strong correlation does not imply causality.

Interpreting the Strength of a Correlation

• Categorizes relationships based on the values of the correlation coefficient (r). (weak, moderate, strong).

Correlation: Outliers

• A data point that stands apart from the majority of points.
• Can significantly impact the strength and validity of a correlation.

Outliers in Correlations (Spearman vs. Pearson)

• Spearman correlation is less sensitive to outliers than Pearson correlation because it utilizes ranks, not raw scores.

Correlation: Significance

• Statistical significance: index of reliability of a correlation.
• P-value: probability the correlation was due to chance. Typically, p<0.05 suggests statistical significance.
• Sample size affects the criteria for statistical significance (larger n allows smaller r to reach significance).

Correlation: Timepoint-Based Correlations

• Cross-sectional, autocorrelations, and cross-lagged correlations are timepoint-based correlation type.

Correlation: Advantages

• Quick and efficient.
• Often the only method available in certain situations. For practical or ethical reasons.
• High external validity.

Correlation: Limitations

• Cannot establish causality.
• Highly sensitive to outliers.
• Directionality problem (order of effects).
• Third-variable problem (existence of a different variable affecting the relationship).

Coefficient of Determination

• Indicates the proportion of variance in one variable that can be explained by another.
• Expressed as r-squared (r²).
• r²is always positive.

Interpreting Correlations

• Evaluation of both strength and significance.
• The strength of a correlation (r squared) indicates the percentage of variability in one variable accounted for by the other variable.
• Significance of a correlation (p value) gives a measure of whether the correlation is due to chance, considering the sample size.

Practical Significance

• Addresses whether the result of a correlation is meaningful in a real-world context. A statistically significant correlation might not have practical significance.

Examples with Small N

• Obtain strong correlations even when no true relationship is present (sample size).

Next Steps

• The next chapter will cover experimental research strategies.

This quiz explores the fundamental concepts of correlational research, including its aims, methods, and how it contrasts with experimental research. You'll learn about correlation coefficients, measurement techniques, and the strengths and weaknesses of correlational studies. Test your understanding of this critical research strategy used in various fields.