Quiz Results
Your report appears here after final submission.
Question Review
Quiz Settings
Configure the assessment before loading your question set.
Scatter Plot Analysis Guide
Scatter plots reveal how two variables move together. Each point represents a paired observation from the dataset. Patterns become meaningful after scales and context are checked.
Reading Direction and Strength
Direction describes whether values rise or fall. Strength describes how tightly points follow a pattern. Strong association does not prove a causal relationship.
Understanding Form and Structure
Linear patterns follow a straight band. Curved patterns require models or transformed variables. Separate clusters may represent hidden groups needing independent analysis.
A horizontal fan can reveal changing variance across predictions. This pattern is commonly called heteroscedasticity in regression diagnostics. Weighted methods or transformations may improve the resulting model.
Evaluating Outliers and Influence
Outliers sit unusually far from the main point cloud. Some outliers change correlation and slope substantially. Others look extreme but have little statistical influence.
High leverage points occupy unusual positions along the horizontal axis. Their influence depends on vertical distance from the fitted line. Cook's distance helps combine leverage and residual information.
Regression Insights
Residuals measure observed values minus fitted values. Random residual scatter supports a simple linear specification. Curved residual patterns suggest omitted nonlinear structure.
Correlation Limitations
Correlation can remain small despite a strong curved relationship. Restricted ranges can also hide meaningful associations. Mixed populations may create misleading pooled correlations.
Simpson's paradox occurs when grouped and pooled trends disagree. Group labels should therefore be examined before final interpretation. Subject knowledge helps identify plausible grouping variables.
Reliable Visual Decisions
Axis limits can exaggerate or conceal apparent changes. Logarithmic scales help when values span several magnitudes. Labels and units prevent incorrect practical conclusions.
Formula Used
Pearson's coefficient equals covariance divided by both standard deviations. Regression slope equals covariance divided by predictor variance. The intercept equals mean response minus slope times mean predictor.
The coefficient of determination equals explained variation divided by total variation. A standardized score equals value minus mean, divided by deviation. These formulas support several calculation questions within the quiz.
How to Use This Quiz
Choose the question count and preferred analysis focus. Select timing, feedback, and scoring options before starting. Inspect every plot before reviewing the available answers.
Use calculations when values or summaries are provided. Record confidence to compare certainty against actual accuracy. Submit the quiz to generate detailed performance charts.
Review missed questions by topic and reasoning error. Repeat weak categories with fewer questions and immediate feedback. Consistent practice builds faster, more reliable visual interpretation skills.
Regression slope: b₁ = Σ[(xᵢ − x̄)(yᵢ − ȳ)] ÷ Σ(xᵢ − x̄)²
Regression intercept: b₀ = ȳ − b₁x̄
Coefficient of determination: R² = 1 − SSE ÷ SST
Frequently Asked Questions
1. What does a scatter plot show?
A scatter plot displays paired numeric observations. Point direction, spacing, clusters, curves, and unusual values reveal possible relationships. Interpretation should consider scales, sample design, and domain context.
2. Does strong correlation prove causation?
No. Strong correlation can arise from confounding, selection, shared trends, or coincidence. Causal conclusions require suitable design, credible assumptions, and supporting evidence beyond visual association.
3. How is Pearson correlation interpreted?
Pearson correlation ranges from negative one to positive one. Its sign gives linear direction. Its magnitude gives linear strength, but it may miss curved relationships.
4. Why can one point change the fitted line?
A point with an unusual predictor value has high leverage. When it also has a large residual, it can strongly alter slope, intercept, and correlation.
5. What does a fan-shaped pattern indicate?
A fan shape often indicates nonconstant variance. Regression errors become more dispersed across predictor values. Transformations, robust errors, or weighted methods may then be appropriate.
6. When should logarithmic axes be used?
Logarithmic axes help when values span several orders of magnitude. They can linearize multiplicative relationships and improve visibility. Zero and negative values require special handling.
7. What is Simpson's paradox?
Simpson's paradox occurs when pooled and grouped trends differ. A lurking grouping variable changes the apparent relationship. Analysts should inspect meaningful subgroups before accepting pooled conclusions.
8. How does sample size affect interpretation?
Small samples produce unstable patterns and uncertain correlations. Larger samples improve precision, but bias can remain. Sampling quality matters more than point count alone.
9. Are quiz results saved automatically?
Active progress is saved within browser storage after each answer. Results remain local to the current browser. Clearing site data removes stored attempts and progress.