Scatterplot: A graph showing the relationship between two quantitative variables by plotting points on a coordinate plane
Bivariate Data: Data consisting of two variables measured on the same individuals
Explanatory Variable (x): The variable that may explain or predict the other variable
Response Variable (y): The variable that responds to changes in the explanatory variable
- Identify the explanatory variable (x-axis) and response variable (y-axis)
- Create a coordinate system with appropriate scales
- Plot each data point as coordinates (x, y)
- Analyze the overall pattern of the points
Explanatory variable (x): Study hours
Response variable (y): Test scores
Points: (2,65), (3,70), (4,75), (5,80), (6,85), (7,90), (8,95)
Points form an upward-sloping linear pattern
As study hours increase, test scores increase
Correlation appears strong and positive
The scatterplot shows a strong positive linear relationship between study hours and test scores. As study hours increase from 2 to 8 hours, test scores consistently increase from 65 to 95. The points form a clear upward-sloping straight line, indicating a strong positive association between the two variables.
• Variable assignment: Explanatory on x-axis, response on y-axis
• Pattern analysis: Look for direction, form, and strength
• Linear relationship: Points form a straight-line pattern
Direction: Positive (upward slope) or negative (downward slope)
Strength: How closely the points follow a pattern (weak, moderate, strong)
Form: The shape of the relationship (linear, curved, clusters)
- Identify the direction (positive or negative)
- Assess the strength (how tightly points cluster)
- Determine the form (linear or non-linear)
- Look for outliers or unusual patterns
Downward slope from left to right = negative relationship
Depends on how closely points follow the downward pattern
As temperature increases, ice cream sales decrease
Form: Linear (assumed)
Relationship: As temp increases, sales decrease
The downward-sloping pattern indicates a negative relationship between temperature and ice cream sales. This means that as temperature increases, ice cream sales tend to decrease. This is contrary to the expected positive relationship, suggesting there might be an error in the data or an unusual situation (perhaps the data represents winter months where hot weather leads to more air conditioning costs, reducing spending on treats).
• Direction: Upward = positive, Downward = negative
• Strength: Closer clustering = stronger relationship
• Context: Always consider if the relationship makes sense
Form: The overall shape of the relationship (linear, quadratic, exponential)
Outlier: A data point that falls far from the overall pattern
Curved Relationship: Non-linear pattern suggesting exponential or polynomial relationship
- Identify the overall form of the relationship
- Look for points that deviate significantly from the pattern
- Consider if the outlier is valid or an error
- Assess how the outlier affects the relationship
Most points follow a curved pattern, indicating non-linear relationship
One point is far from the curved pattern
Determine if it's a legitimate data point or an error
Outlier: Present
Pattern: Age vs. reaction time typically curved
The scatterplot shows a curved (non-linear) relationship between age and reaction time, which is typical as reaction time tends to be fast in young ages, slowest in older ages, with a gradual change. The outlier point is a data point that doesn't follow this curved pattern. It could represent an unusual individual or potentially an error in data collection. The outlier should be investigated to determine if it should be included in analysis.
• Form recognition: Look for linear, curved, or clustered patterns
• Outlier identification: Points that deviate significantly
• Outlier investigation: Determine validity before analysis
Correlation Coefficient (r): A number between -1 and 1 measuring the strength and direction of a linear relationship
Strong Positive Correlation: r close to +1
Strong Negative Correlation: r close to -1
- Identify the direction of the relationship
- Assess the strength based on point clustering
- Estimate the correlation coefficient value
- Interpret the meaning of the correlation
Upward slope indicates positive correlation
Very tight pattern indicates strong correlation
Strong positive linear relationship: r ≈ 0.8 to 0.99
Strength: Strong
Correlation coefficient: r ≈ 0.8 to 0.99
The correlation coefficient would be close to +1, likely between 0.8 and 0.99. A very tight linear pattern sloping upward indicates a strong positive linear relationship. The closer the points are to forming a perfect straight line, the closer the correlation coefficient is to +1. This suggests a very strong positive association between the two variables.
• Correlation range: -1 ≤ r ≤ 1
• Direction: Positive slope = positive r
• Strength: Tighter clustering = |r| closer to 1
Non-linear Relationship: Relationship that doesn't follow a straight line
Diminishing Returns: Situation where additional input produces smaller increases in output
Threshold Effect: Point beyond which additional input has little effect
- Identify the overall pattern and any changes
- Consider the real-world context
- Interpret the meaning of the pattern
- Consider practical implications
Positive relationship initially, then levels off after 30 inches
Plants need water but too much can be harmful
Additional rain beyond 30 inches provides little benefit
Threshold: ~30 inches
Meaning: Diminishing returns after threshold
The scatterplot shows a positive relationship between rainfall and crop yield up to about 30 inches, after which the relationship levels off. This suggests that additional rainfall beyond 30 inches provides diminishing returns for crop yield. In agricultural terms, this could be because plants can only absorb a certain amount of water, and excessive rainfall may lead to waterlogging, nutrient leaching, or other problems that prevent further yield improvements.
• Context matters: Interpret patterns in real-world terms
• Non-linear relationships: Common in real-world phenomena
• Threshold effects: Many relationships have optimal ranges
Direction: Positive (upward) or Negative (downward)
Form: Linear, Curved, or Clustered
Strength: Weak, Moderate, or Strong
Outliers: Points that deviate significantly
Bivariate Data: Data that consists of two quantitative variables measured on the same individuals or objects. Scatterplots are the primary tool for visualizing bivariate data.
Association vs. Causation: Scatterplots show association between variables, but association does not imply causation. Other variables may influence the relationship.
Correlation Coefficient: A numerical measure ranging from -1 to +1 that quantifies the strength and direction of a linear relationship between two variables.
- Data Preparation: Identify explanatory and response variables
- Scatterplot Creation: Plot points on coordinate system
- Pattern Analysis: Assess direction, form, and strength
- Outlier Identification: Look for unusual points
- Interpretation: Explain relationship in context
• Correlation range: -1 ≤ r ≤ 1
• Direction: Positive slope = positive r, Negative slope = negative r
• Strength: |r| closer to 1 = stronger relationship
• Form: Correlation only measures linear relationships
Residual: The difference between observed and predicted values in a scatterplot, used to assess how well a line fits the data.
Lurking Variables: Variables that are not included in the analysis but may influence the relationship between the two plotted variables.
Scatterplot Smoothing: Techniques like LOESS to reveal underlying patterns in noisy scatterplots.
- Variable Identification: Correctly assign explanatory and response variables based on research question
- Scale Selection: Choose appropriate axis scales to clearly show patterns
- Pattern Recognition: Distinguish between linear and non-linear relationships
- Outlier Analysis: Determine if outliers are errors or meaningful data points
- Contextual Interpretation: Connect statistical patterns to real-world phenomena
Questions & Answers
Question: Can a scatterplot show causation between two variables, or only correlation?
Answer: Scatterplots can only show correlation (association), not causation. A strong correlation between two variables doesn't prove that one causes the other. For example, ice cream sales and drowning incidents might be correlated, but eating ice cream doesn't cause drowning - both are related to hot weather. To establish causation, you need controlled experiments or additional evidence beyond observational data. Always remember: correlation does not imply causation.
Question: What's the difference between correlation and regression? Are they the same thing?
Answer: Correlation and regression are related but different concepts:
- Correlation: Measures the strength and direction of the linear relationship between two variables (just a number, r)
- Regression: Creates an equation that predicts one variable from another (a line of best fit)
You can think of correlation as describing how closely two variables move together, while regression gives you a formula to predict one variable from the other. Correlation comes first - you need to see a relationship before creating a predictive model.
Question: How do I know if a correlation coefficient is statistically significant?
Answer: The correlation coefficient alone doesn't tell you if the relationship is statistically significant. Significance depends on both the correlation coefficient and the sample size:
- Small samples: Even strong correlations might not be significant
- Large samples: Weak correlations might be significant
To determine significance, you need to perform a hypothesis test for the correlation coefficient, which considers both the r value and the sample size (n). For now, focus on interpreting the strength and direction of the relationship shown in your scatterplot.