Scatterplots and Correlation: Direction, Form, Strength, and Outliers
Describe bivariate quantitative relationships precisely and use correlation only when its linear-association conditions make sense.
Direct answer
scatterplots and correlation: Describe a scatterplot by direction, form, strength, and unusual features before quoting correlation. Pearson r summarizes linear association only and does not establish causation.
Quick reference: Scatterplots and Correlation: Direction, Form, Strength, and Outliers
| Correlation | −1≤r≤1; direction and strength of linear association |
|---|---|
| No causation | Correlation alone does not establish a causal relationship. |
Scatterplots and Correlation: Direction, Form, Strength, and Outliers: complete lesson
A scatterplot is for two quantitative variables measured on the same cases
Scatterplots and correlation describe relationships between paired quantitative measurements. Each point represents one observational unit with coordinates (x,y). Before describing a pattern, identify what x and y mean and which variable is explanatory if the context suggests a directional modeling role. A scatterplot is not appropriate for two categorical variables; a two-way table is usually the relevant display there.
A strong description uses four features: direction, form, strength, and unusual features. “Positive” alone is incomplete. A better statement might be “There is a moderately strong, positive, roughly linear association between weekly study hours and exam score, with one high-leverage point near 18 study hours.” Context converts a geometric observation into statistical communication.
Direction describes whether larger x tends to accompany larger or smaller y
A positive association means larger x-values generally occur with larger y-values; a negative association means larger x-values generally occur with smaller y-values. No clear direction means neither monotone tendency dominates. Direction does not establish causation because lurking variables, reverse causation, selection, or common causes can create an association.
The axes matter. Swapping x and y leaves the sign and strength of correlation unchanged, but it changes the interpretation of a regression model because prediction treats one variable as explanatory and the other as response.
Form determines whether correlation is a sensible summary
Pearson correlation r measures the strength and direction of a linear relationship. A strong curved association can have r near 0 because positive and negative linear tendencies cancel. Therefore inspect a scatterplot before using r as the primary summary. “r is small” does not imply “there is no relationship.”
Clusters can also make one overall r misleading. If two subgroups each have weak association but their centers are far apart, the combined data can show a strong correlation driven by group membership. Coloring or stratifying by a relevant categorical variable can reveal that structure.
Correlation is unitless and bounded between −1 and 1
Because correlation standardizes both variables, changing from centimeters to meters or dollars to thousands of dollars leaves r unchanged, provided the transformation is positive linear. Adding a constant also leaves r unchanged. Multiplying one variable by a negative constant reverses the sign because it reverses that variable’s ordering.
Values near +1 or −1 indicate points concentrated around a straight line, while values near 0 indicate weak linear association. There is no universal threshold separating “strong” from “moderate”; context, sample size, measurement reliability, and the purpose of the analysis matter.
Outliers and influential points can change correlation substantially
A point far from the main cloud can affect r, especially in small samples. A point far in x-direction is often called high leverage. If it lies along the existing linear trend, it can make |r| larger; if it lies away from the trend, it can weaken or even reverse the correlation. Removing a point merely because it changes r is not defensible without a substantive reason such as measurement error or a clearly different population.
An outlier in y but near the center of x can create a large vertical residual without high leverage. Distinguishing vertical unusualness from x-leverage helps anticipate whether the point will strongly change a fitted line and correlation.
Correlation does not distinguish explanatory and response roles
r(x,y)=r(y,x), so correlation is symmetric. Regression is not: predicting y from x gives a different line from predicting x from y. This is why a correlation statement should describe association rather than prediction unless a regression model has also been specified.
Similarly, correlation cannot by itself adjust for confounding. A strong positive correlation between ice-cream sales and drowning incidents does not imply either causes the other; season or temperature can influence both. Design and subject-matter reasoning are required for causal claims.
Restricted range can weaken an observed correlation
If a study samples only cases from a narrow x-range, the scatterplot may show less variation and a weaker observed r than the broader population relationship. For example, the correlation between entrance-exam score and first-year GPA among students admitted only from a narrow high-score band can be smaller than in the full applicant population.
This is a selection issue, not proof that the variables are unrelated. Always interpret a correlation for the population and range actually represented by the data.
Worked scatterplot descriptions
Scatterplot case 1: Study hours vs exam score
A suitable description is: “The data show a moderately strong, positive, roughly linear association between study hours vs exam score; one high-study student follows the trend.” This sentence reports geometry first and then identifies the unusual structure that might affect a numerical summary or regression model.
Correlation is most informative here when the form is roughly linear. Because the form is approximately linear, Pearson r can supplement the scatterplot by summarizing the direction and linear strength, but it still does not establish causation. In “Scatterplot case 1: Study hours vs exam score,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Before fitting a line, check whether the unusual feature represents an error, a legitimately extreme case, or a subgroup. The point should not be discarded simply because it changes r. Any exclusion must be justified by the data-generating process and reported transparently. In “Scatterplot case 1: Study hours vs exam score,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Scatterplot case 2: Outside temperature vs heating use
A suitable description is: “The data show a strong, negative, roughly linear association between outside temperature vs heating use; no major outliers.” This sentence reports geometry first and then identifies the unusual structure that might affect a numerical summary or regression model.
Correlation is most informative here when the form is roughly linear. Because the form is approximately linear, Pearson r can supplement the scatterplot by summarizing the direction and linear strength, but it still does not establish causation. In “Scatterplot case 2: Outside temperature vs heating use,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Before fitting a line, check whether the unusual feature represents an error, a legitimately extreme case, or a subgroup. The point should not be discarded simply because it changes r. Any exclusion must be justified by the data-generating process and reported transparently. In “Scatterplot case 2: Outside temperature vs heating use,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Scatterplot case 3: Age vs reaction time in children
A suitable description is: “The data show a moderate, positive, curved association between age vs reaction time in children; older ages level off.” This sentence reports geometry first and then identifies the unusual structure that might affect a numerical summary or regression model.
Correlation is most informative here when the form is roughly linear. Because the form is curved, a single Pearson r would understate or distort the visible relationship; the curve itself is the main feature to describe. In “Scatterplot case 3: Age vs reaction time in children,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Before fitting a line, check whether the unusual feature represents an error, a legitimately extreme case, or a subgroup. The point should not be discarded simply because it changes r. Any exclusion must be justified by the data-generating process and reported transparently. In “Scatterplot case 3: Age vs reaction time in children,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Scatterplot case 4: Engine size vs fuel economy
A suitable description is: “The data show a moderately strong, negative, roughly linear association between engine size vs fuel economy; one sports car is high leverage.” This sentence reports geometry first and then identifies the unusual structure that might affect a numerical summary or regression model.
Correlation is most informative here when the form is roughly linear. Because the form is approximately linear, Pearson r can supplement the scatterplot by summarizing the direction and linear strength, but it still does not establish causation. In “Scatterplot case 4: Engine size vs fuel economy,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Before fitting a line, check whether the unusual feature represents an error, a legitimately extreme case, or a subgroup. The point should not be discarded simply because it changes r. Any exclusion must be justified by the data-generating process and reported transparently. In “Scatterplot case 4: Engine size vs fuel economy,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Scatterplot case 5: Rainfall vs crop yield
A suitable description is: “The data show a moderate, positive, curved association between rainfall vs crop yield; yield plateaus at high rainfall.” This sentence reports geometry first and then identifies the unusual structure that might affect a numerical summary or regression model.
Correlation is most informative here when the form is roughly linear. Because the form is curved, a single Pearson r would understate or distort the visible relationship; the curve itself is the main feature to describe. In “Scatterplot case 5: Rainfall vs crop yield,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Before fitting a line, check whether the unusual feature represents an error, a legitimately extreme case, or a subgroup. The point should not be discarded simply because it changes r. Any exclusion must be justified by the data-generating process and reported transparently. In “Scatterplot case 5: Rainfall vs crop yield,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Scatterplot case 6: Advertising spend vs sales
A suitable description is: “The data show a strong, positive, roughly linear association between advertising spend vs sales; one campaign has a large negative residual.” This sentence reports geometry first and then identifies the unusual structure that might affect a numerical summary or regression model.
Correlation is most informative here when the form is roughly linear. Because the form is approximately linear, Pearson r can supplement the scatterplot by summarizing the direction and linear strength, but it still does not establish causation. In “Scatterplot case 6: Advertising spend vs sales,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Before fitting a line, check whether the unusual feature represents an error, a legitimately extreme case, or a subgroup. The point should not be discarded simply because it changes r. Any exclusion must be justified by the data-generating process and reported transparently. In “Scatterplot case 6: Advertising spend vs sales,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Scatterplot case 7: Elevation vs air pressure
A suitable description is: “The data show a strong, negative, roughly linear association between elevation vs air pressure; no clear cluster.” This sentence reports geometry first and then identifies the unusual structure that might affect a numerical summary or regression model.
Correlation is most informative here when the form is roughly linear. Because the form is approximately linear, Pearson r can supplement the scatterplot by summarizing the direction and linear strength, but it still does not establish causation. In “Scatterplot case 7: Elevation vs air pressure,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Before fitting a line, check whether the unusual feature represents an error, a legitimately extreme case, or a subgroup. The point should not be discarded simply because it changes r. Any exclusion must be justified by the data-generating process and reported transparently. In “Scatterplot case 7: Elevation vs air pressure,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Scatterplot case 8: Practice trials vs error rate
A suitable description is: “The data show a strong, negative, curved association between practice trials vs error rate; rapid early decline then flattening.” This sentence reports geometry first and then identifies the unusual structure that might affect a numerical summary or regression model.
Correlation is most informative here when the form is roughly linear. Because the form is curved, a single Pearson r would understate or distort the visible relationship; the curve itself is the main feature to describe. In “Scatterplot case 8: Practice trials vs error rate,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Before fitting a line, check whether the unusual feature represents an error, a legitimately extreme case, or a subgroup. The point should not be discarded simply because it changes r. Any exclusion must be justified by the data-generating process and reported transparently. In “Scatterplot case 8: Practice trials vs error rate,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Scatterplot case 9: Height vs shoe length
A suitable description is: “The data show a moderate, positive, roughly linear association between height vs shoe length; two age-group clusters.” This sentence reports geometry first and then identifies the unusual structure that might affect a numerical summary or regression model.
Correlation is most informative here when the form is roughly linear. Because the form is approximately linear, Pearson r can supplement the scatterplot by summarizing the direction and linear strength, but it still does not establish causation. In “Scatterplot case 9: Height vs shoe length,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Before fitting a line, check whether the unusual feature represents an error, a legitimately extreme case, or a subgroup. The point should not be discarded simply because it changes r. Any exclusion must be justified by the data-generating process and reported transparently. In “Scatterplot case 9: Height vs shoe length,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Scatterplot case 10: Distance from city vs home price
A suitable description is: “The data show a weak to moderate, negative, roughly linear association between distance from city vs home price; several luxury outliers.” This sentence reports geometry first and then identifies the unusual structure that might affect a numerical summary or regression model.
Correlation is most informative here when the form is roughly linear. Because the form is approximately linear, Pearson r can supplement the scatterplot by summarizing the direction and linear strength, but it still does not establish causation. In “Scatterplot case 10: Distance from city vs home price,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Before fitting a line, check whether the unusual feature represents an error, a legitimately extreme case, or a subgroup. The point should not be discarded simply because it changes r. Any exclusion must be justified by the data-generating process and reported transparently. In “Scatterplot case 10: Distance from city vs home price,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Scatterplot case 11: Daily screen time vs sleep duration
A suitable description is: “The data show a weak, negative, roughly linear association between daily screen time vs sleep duration; wide vertical spread.” This sentence reports geometry first and then identifies the unusual structure that might affect a numerical summary or regression model.
Correlation is most informative here when the form is roughly linear. Because the form is approximately linear, Pearson r can supplement the scatterplot by summarizing the direction and linear strength, but it still does not establish causation. In “Scatterplot case 11: Daily screen time vs sleep duration,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Before fitting a line, check whether the unusual feature represents an error, a legitimately extreme case, or a subgroup. The point should not be discarded simply because it changes r. Any exclusion must be justified by the data-generating process and reported transparently. In “Scatterplot case 11: Daily screen time vs sleep duration,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Scatterplot case 12: Speed vs braking distance
A suitable description is: “The data show a strong, positive, curved association between speed vs braking distance; relationship steepens at high speed.” This sentence reports geometry first and then identifies the unusual structure that might affect a numerical summary or regression model.
Correlation is most informative here when the form is roughly linear. Because the form is curved, a single Pearson r would understate or distort the visible relationship; the curve itself is the main feature to describe. In “Scatterplot case 12: Speed vs braking distance,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Before fitting a line, check whether the unusual feature represents an error, a legitimately extreme case, or a subgroup. The point should not be discarded simply because it changes r. Any exclusion must be justified by the data-generating process and reported transparently. In “Scatterplot case 12: Speed vs braking distance,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Association-diagnostics laboratory
Association-diagnostics laboratory 1: school survey
For each scatterplot description, separate visual form from numerical correlation. Identify clusters, curvature, leverage, and outliers before deciding whether r is a useful summary, and keep association language separate from causal claims. In this school survey, write the target in words before using notation so the calculation remains tied to the variable, population, or model actually being studied.
A high-quality solution should also include a self-check tailored to this topic: compare the sign, units, boundary, probability range, or design logic with what the scenario makes plausible. If the numerical output contradicts that check, revisit the setup before changing the conclusion. In “Association-diagnostics laboratory 1: school survey,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Finally, distinguish what the statistical method establishes from what the study design does not establish. The strongest response gives the numerical or graphical evidence and then limits the claim to the population, process, association, or legacy-enrichment scope justified by the data-generating mechanism. In “Association-diagnostics laboratory 1: school survey,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Association-diagnostics laboratory 2: public-health study
For each scatterplot description, separate visual form from numerical correlation. Identify clusters, curvature, leverage, and outliers before deciding whether r is a useful summary, and keep association language separate from causal claims. In this public-health study, write the target in words before using notation so the calculation remains tied to the variable, population, or model actually being studied.
A high-quality solution should also include a self-check tailored to this topic: compare the sign, units, boundary, probability range, or design logic with what the scenario makes plausible. If the numerical output contradicts that check, revisit the setup before changing the conclusion. In “Association-diagnostics laboratory 2: public-health study,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Finally, distinguish what the statistical method establishes from what the study design does not establish. The strongest response gives the numerical or graphical evidence and then limits the claim to the population, process, association, or legacy-enrichment scope justified by the data-generating mechanism. In “Association-diagnostics laboratory 2: public-health study,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Association-diagnostics laboratory 3: manufacturing process
For each scatterplot description, separate visual form from numerical correlation. Identify clusters, curvature, leverage, and outliers before deciding whether r is a useful summary, and keep association language separate from causal claims. In this manufacturing process, write the target in words before using notation so the calculation remains tied to the variable, population, or model actually being studied.
A high-quality solution should also include a self-check tailored to this topic: compare the sign, units, boundary, probability range, or design logic with what the scenario makes plausible. If the numerical output contradicts that check, revisit the setup before changing the conclusion. In “Association-diagnostics laboratory 3: manufacturing process,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Finally, distinguish what the statistical method establishes from what the study design does not establish. The strongest response gives the numerical or graphical evidence and then limits the claim to the population, process, association, or legacy-enrichment scope justified by the data-generating mechanism. In “Association-diagnostics laboratory 3: manufacturing process,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Association-diagnostics laboratory 4: transportation system
For each scatterplot description, separate visual form from numerical correlation. Identify clusters, curvature, leverage, and outliers before deciding whether r is a useful summary, and keep association language separate from causal claims. In this transportation system, write the target in words before using notation so the calculation remains tied to the variable, population, or model actually being studied.
A high-quality solution should also include a self-check tailored to this topic: compare the sign, units, boundary, probability range, or design logic with what the scenario makes plausible. If the numerical output contradicts that check, revisit the setup before changing the conclusion. In “Association-diagnostics laboratory 4: transportation system,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Finally, distinguish what the statistical method establishes from what the study design does not establish. The strongest response gives the numerical or graphical evidence and then limits the claim to the population, process, association, or legacy-enrichment scope justified by the data-generating mechanism. In “Association-diagnostics laboratory 4: transportation system,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Association-diagnostics laboratory 5: consumer study
For each scatterplot description, separate visual form from numerical correlation. Identify clusters, curvature, leverage, and outliers before deciding whether r is a useful summary, and keep association language separate from causal claims. In this consumer study, write the target in words before using notation so the calculation remains tied to the variable, population, or model actually being studied.
A high-quality solution should also include a self-check tailored to this topic: compare the sign, units, boundary, probability range, or design logic with what the scenario makes plausible. If the numerical output contradicts that check, revisit the setup before changing the conclusion. In “Association-diagnostics laboratory 5: consumer study,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Finally, distinguish what the statistical method establishes from what the study design does not establish. The strongest response gives the numerical or graphical evidence and then limits the claim to the population, process, association, or legacy-enrichment scope justified by the data-generating mechanism. In “Association-diagnostics laboratory 5: consumer study,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Association-diagnostics laboratory 6: environmental monitoring
For each scatterplot description, separate visual form from numerical correlation. Identify clusters, curvature, leverage, and outliers before deciding whether r is a useful summary, and keep association language separate from causal claims. In this environmental monitoring, write the target in words before using notation so the calculation remains tied to the variable, population, or model actually being studied.
A high-quality solution should also include a self-check tailored to this topic: compare the sign, units, boundary, probability range, or design logic with what the scenario makes plausible. If the numerical output contradicts that check, revisit the setup before changing the conclusion. In “Association-diagnostics laboratory 6: environmental monitoring,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Finally, distinguish what the statistical method establishes from what the study design does not establish. The strongest response gives the numerical or graphical evidence and then limits the claim to the population, process, association, or legacy-enrichment scope justified by the data-generating mechanism. In “Association-diagnostics laboratory 6: environmental monitoring,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Association-diagnostics laboratory 7: sports analysis
For each scatterplot description, separate visual form from numerical correlation. Identify clusters, curvature, leverage, and outliers before deciding whether r is a useful summary, and keep association language separate from causal claims. In this sports analysis, write the target in words before using notation so the calculation remains tied to the variable, population, or model actually being studied.
A high-quality solution should also include a self-check tailored to this topic: compare the sign, units, boundary, probability range, or design logic with what the scenario makes plausible. If the numerical output contradicts that check, revisit the setup before changing the conclusion. In “Association-diagnostics laboratory 7: sports analysis,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Finally, distinguish what the statistical method establishes from what the study design does not establish. The strongest response gives the numerical or graphical evidence and then limits the claim to the population, process, association, or legacy-enrichment scope justified by the data-generating mechanism. In “Association-diagnostics laboratory 7: sports analysis,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Association-diagnostics laboratory 8: education program
For each scatterplot description, separate visual form from numerical correlation. Identify clusters, curvature, leverage, and outliers before deciding whether r is a useful summary, and keep association language separate from causal claims. In this education program, write the target in words before using notation so the calculation remains tied to the variable, population, or model actually being studied.
A high-quality solution should also include a self-check tailored to this topic: compare the sign, units, boundary, probability range, or design logic with what the scenario makes plausible. If the numerical output contradicts that check, revisit the setup before changing the conclusion. In “Association-diagnostics laboratory 8: education program,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Finally, distinguish what the statistical method establishes from what the study design does not establish. The strongest response gives the numerical or graphical evidence and then limits the claim to the population, process, association, or legacy-enrichment scope justified by the data-generating mechanism. In “Association-diagnostics laboratory 8: education program,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Association-diagnostics laboratory 9: service operation
For each scatterplot description, separate visual form from numerical correlation. Identify clusters, curvature, leverage, and outliers before deciding whether r is a useful summary, and keep association language separate from causal claims. In this service operation, write the target in words before using notation so the calculation remains tied to the variable, population, or model actually being studied.
A high-quality solution should also include a self-check tailored to this topic: compare the sign, units, boundary, probability range, or design logic with what the scenario makes plausible. If the numerical output contradicts that check, revisit the setup before changing the conclusion. In “Association-diagnostics laboratory 9: service operation,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Finally, distinguish what the statistical method establishes from what the study design does not establish. The strongest response gives the numerical or graphical evidence and then limits the claim to the population, process, association, or legacy-enrichment scope justified by the data-generating mechanism. In “Association-diagnostics laboratory 9: service operation,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Association-diagnostics laboratory 10: technology experiment
For each scatterplot description, separate visual form from numerical correlation. Identify clusters, curvature, leverage, and outliers before deciding whether r is a useful summary, and keep association language separate from causal claims. In this technology experiment, write the target in words before using notation so the calculation remains tied to the variable, population, or model actually being studied.
A high-quality solution should also include a self-check tailored to this topic: compare the sign, units, boundary, probability range, or design logic with what the scenario makes plausible. If the numerical output contradicts that check, revisit the setup before changing the conclusion. In “Association-diagnostics laboratory 10: technology experiment,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Finally, distinguish what the statistical method establishes from what the study design does not establish. The strongest response gives the numerical or graphical evidence and then limits the claim to the population, process, association, or legacy-enrichment scope justified by the data-generating mechanism. In “Association-diagnostics laboratory 10: technology experiment,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Association-diagnostics laboratory 11: community poll
For each scatterplot description, separate visual form from numerical correlation. Identify clusters, curvature, leverage, and outliers before deciding whether r is a useful summary, and keep association language separate from causal claims. In this community poll, write the target in words before using notation so the calculation remains tied to the variable, population, or model actually being studied.
A high-quality solution should also include a self-check tailored to this topic: compare the sign, units, boundary, probability range, or design logic with what the scenario makes plausible. If the numerical output contradicts that check, revisit the setup before changing the conclusion. In “Association-diagnostics laboratory 11: community poll,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Finally, distinguish what the statistical method establishes from what the study design does not establish. The strongest response gives the numerical or graphical evidence and then limits the claim to the population, process, association, or legacy-enrichment scope justified by the data-generating mechanism. In “Association-diagnostics laboratory 11: community poll,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Association-diagnostics laboratory 12: quality-control review
For each scatterplot description, separate visual form from numerical correlation. Identify clusters, curvature, leverage, and outliers before deciding whether r is a useful summary, and keep association language separate from causal claims. In this quality-control review, write the target in words before using notation so the calculation remains tied to the variable, population, or model actually being studied.
A high-quality solution should also include a self-check tailored to this topic: compare the sign, units, boundary, probability range, or design logic with what the scenario makes plausible. If the numerical output contradicts that check, revisit the setup before changing the conclusion. In “Association-diagnostics laboratory 12: quality-control review,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Finally, distinguish what the statistical method establishes from what the study design does not establish. The strongest response gives the numerical or graphical evidence and then limits the claim to the population, process, association, or legacy-enrichment scope justified by the data-generating mechanism. In “Association-diagnostics laboratory 12: quality-control review,” apply this check to the named variables and numerical direction rather than treating it as a reusable sentence from another exercise. Inspect form and unusual points before treating correlation as a useful summary, and keep association language separate from causal language.
Scatterplots and Correlation: Direction, Form, Strength, and Outliers: 22 multiple-choice questions
These items practice complete scatterplot descriptions, limits of correlation, transformations, outliers, clusters, and causal overreach.
Question 1. Scatterplots and Correlation
A scatterplot for 29 parts at a regional manufacturer in Riverbend during a service-improvement study is roughly linear with correlation r=-0.62. Interpret direction and strength, and state one limitation of r.
Answer: C
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 1. Scatterplots and Correlation: The association is moderate and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 2. Scatterplots and Correlation
A scatterplot for 23 appointments at a university advising center in Midwest consortium during a multiweek validation study is roughly linear with correlation r=0.28. Interpret direction and strength, and state one limitation of r.
Answer: A
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 2. Scatterplots and Correlation: The association is weak and positive: larger x values tend to occur with larger y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 3. Scatterplots and Correlation
A scatterplot for 42 samples at a food safety laboratory in Metro East during a fall 2026 audit is roughly linear with correlation r=-0.86. Interpret direction and strength, and state one limitation of r.
Answer: B
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 3. Scatterplots and Correlation: The association is strong and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 4. Scatterplots and Correlation
A scatterplot for 67 visitors at a state park in Westview during a winter readiness review is roughly linear with correlation r=-0.35. Interpret direction and strength, and state one limitation of r.
Answer: A
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 4. Scatterplots and Correlation: The association is weak and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 5. Scatterplots and Correlation
A scatterplot for 30 appointments at a university advising center in Pacific Northwest during a weekday operations study is roughly linear with correlation r=-0.35. Interpret direction and strength, and state one limitation of r.
Answer: A
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 5. Scatterplots and Correlation: The association is weak and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 6. Scatterplots and Correlation
A scatterplot for 61 visitors at a county library in South Harbor during a spring 2027 pilot is roughly linear with correlation r=-0.86. Interpret direction and strength, and state one limitation of r.
Answer: B
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 6. Scatterplots and Correlation: The association is strong and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 7. Scatterplots and Correlation
A scatterplot for 46 ballots at a county election office in Westview during a fall 2026 audit is roughly linear with correlation r=-0.35. Interpret direction and strength, and state one limitation of r.
Answer: D
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 7. Scatterplots and Correlation: The association is weak and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 8. Scatterplots and Correlation
A scatterplot for 71 students at a school district in Sunbelt district during a weekday operations study is roughly linear with correlation r=-0.62. Interpret direction and strength, and state one limitation of r.
Answer: A
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 8. Scatterplots and Correlation: The association is moderate and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 9. Scatterplots and Correlation
A scatterplot for 50 students at a public high school in North Valley during a follow-up evaluation period is roughly linear with correlation r=0.57. Interpret direction and strength, and state one limitation of r.
Answer: C
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 9. Scatterplots and Correlation: The association is moderate and positive: larger x values tend to occur with larger y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 10. Scatterplots and Correlation
A scatterplot for 45 students at a school district in Pine Ridge during a yearly program evaluation is roughly linear with correlation r=0.28. Interpret direction and strength, and state one limitation of r.
Answer: C
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 10. Scatterplots and Correlation: The association is weak and positive: larger x values tend to occur with larger y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 11. Scatterplots and Correlation
A scatterplot for 59 accounts at a municipal water office in Midwest consortium during a community outreach cycle is roughly linear with correlation r=0.81. Interpret direction and strength, and state one limitation of r.
Answer: C
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 11. Scatterplots and Correlation: The association is strong and positive: larger x values tend to occur with larger y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 12. Scatterplots and Correlation
A scatterplot for 28 travelers at a regional airport authority in Prairie District during a fall 2026 audit is roughly linear with correlation r=-0.62. Interpret direction and strength, and state one limitation of r.
Answer: A
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 12. Scatterplots and Correlation: The association is moderate and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 13. Scatterplots and Correlation
A scatterplot for 75 visitors at a county library in Lakeside district during a winter readiness review is roughly linear with correlation r=-0.86. Interpret direction and strength, and state one limitation of r.
Answer: B
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 13. Scatterplots and Correlation: The association is strong and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 14. Scatterplots and Correlation
A scatterplot for 56 visitors at a state park in Pine Ridge during a yearly program evaluation is roughly linear with correlation r=0.28. Interpret direction and strength, and state one limitation of r.
Answer: D
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 14. Scatterplots and Correlation: The association is weak and positive: larger x values tend to occur with larger y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 15. Scatterplots and Correlation
A scatterplot for 50 students at a school district in Desert County during a quarterly performance study is roughly linear with correlation r=-0.62. Interpret direction and strength, and state one limitation of r.
Answer: D
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 15. Scatterplots and Correlation: The association is moderate and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 16. Scatterplots and Correlation
A scatterplot for 61 enrolled learners at a community college in Pacific Northwest during a monthly quality review is roughly linear with correlation r=0.57. Interpret direction and strength, and state one limitation of r.
Answer: B
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 16. Scatterplots and Correlation: The association is moderate and positive: larger x values tend to occur with larger y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 17. Scatterplots and Correlation
A scatterplot for 30 visitors at a county library in Coastal Plains during a community outreach cycle is roughly linear with correlation r=-0.35. Interpret direction and strength, and state one limitation of r.
Answer: C
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 17. Scatterplots and Correlation: The association is weak and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 18. Scatterplots and Correlation
A scatterplot for 24 samples at a food safety laboratory in Capital Region during a follow-up evaluation period is roughly linear with correlation r=-0.86. Interpret direction and strength, and state one limitation of r.
Answer: D
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 18. Scatterplots and Correlation: The association is strong and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 19. Scatterplots and Correlation
A scatterplot for 29 students at a school district in New England network during a regional benchmarking study is roughly linear with correlation r=0.28. Interpret direction and strength, and state one limitation of r.
Answer: B
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 19. Scatterplots and Correlation: The association is weak and positive: larger x values tend to occur with larger y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 20. Scatterplots and Correlation
A scatterplot for 59 accounts at a municipal water office in Atlantic Corridor during a spring 2027 pilot is roughly linear with correlation r=-0.35. Interpret direction and strength, and state one limitation of r.
Answer: C
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 20. Scatterplots and Correlation: The association is weak and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 21. Scatterplots and Correlation
A scatterplot for 57 parts at a regional manufacturer in Lakeside district during a winter readiness review is roughly linear with correlation r=-0.62. Interpret direction and strength, and state one limitation of r.
Answer: B
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 21. Scatterplots and Correlation: The association is moderate and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Question 22. Scatterplots and Correlation
A scatterplot for 79 students at a public high school in Desert County during a semester-long cohort study is roughly linear with correlation r=-0.86. Interpret direction and strength, and state one limitation of r.
Answer: C
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — Question 22. Scatterplots and Correlation: The association is strong and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Scatterplots and Correlation: Direction, Form, Strength, and Outliers: 6 free-response questions
For each free-response prompt, describe direction, form, strength, and unusual features, then explain whether correlation is an appropriate summary and what the design permits you to claim.
FRQ set 1: Scatterplots and Correlation
Scenario. A scatterplot for 79 visitors at a state park in North Valley during a weekday operations study is roughly linear with correlation r=0.57. Interpret direction and strength, and state one limitation of r.
- Define the variable, units, groups, and requested distribution or model feature.
- Show the required calculation or graphical/model reasoning with labeled quantities.
- Interpret the numerical result in the context of the data rather than as an isolated number.
- Identify an unusual feature, limitation, or condition that affects the conclusion.
Model response
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — FRQ set 1: Scatterplots and Correlation: The association is moderate and positive: larger x values tend to occur with larger y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
FRQ set 2: Scatterplots and Correlation
Scenario. A scatterplot for 26 ballots at a county election office in Lakeside district during a fall 2026 audit is roughly linear with correlation r=-0.62. Interpret direction and strength, and state one limitation of r.
- Define the variable, units, groups, and requested distribution or model feature.
- Show the required calculation or graphical/model reasoning with labeled quantities.
- Interpret the numerical result in the context of the data rather than as an isolated number.
- Identify an unusual feature, limitation, or condition that affects the conclusion.
Model response
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — FRQ set 2: Scatterplots and Correlation: The association is moderate and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
FRQ set 3: Scatterplots and Correlation
Scenario. A scatterplot for 54 samples at a food safety laboratory in Central County during a follow-up evaluation period is roughly linear with correlation r=-0.62. Interpret direction and strength, and state one limitation of r.
- Define the variable, units, groups, and requested distribution or model feature.
- Show the required calculation or graphical/model reasoning with labeled quantities.
- Interpret the numerical result in the context of the data rather than as an isolated number.
- Identify an unusual feature, limitation, or condition that affects the conclusion.
Model response
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — FRQ set 3: Scatterplots and Correlation: The association is moderate and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
FRQ set 4: Scatterplots and Correlation
Scenario. A scatterplot for 28 accounts at a municipal water office in Pine Ridge during a quarterly performance study is roughly linear with correlation r=-0.35. Interpret direction and strength, and state one limitation of r.
- Define the variable, units, groups, and requested distribution or model feature.
- Show the required calculation or graphical/model reasoning with labeled quantities.
- Interpret the numerical result in the context of the data rather than as an isolated number.
- Identify an unusual feature, limitation, or condition that affects the conclusion.
Model response
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — FRQ set 4: Scatterplots and Correlation: The association is weak and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
FRQ set 5: Scatterplots and Correlation
Scenario. A scatterplot for 74 ballots at a county election office in Central County during a six-week field trial is roughly linear with correlation r=0.57. Interpret direction and strength, and state one limitation of r.
- Define the variable, units, groups, and requested distribution or model feature.
- Show the required calculation or graphical/model reasoning with labeled quantities.
- Interpret the numerical result in the context of the data rather than as an isolated number.
- Identify an unusual feature, limitation, or condition that affects the conclusion.
Model response
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — FRQ set 5: Scatterplots and Correlation: The association is moderate and positive: larger x values tend to occur with larger y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
FRQ set 6: Scatterplots and Correlation
Scenario. A scatterplot for 57 students at a school district in Atlantic Corridor during a two-month observation window is roughly linear with correlation r=-0.86. Interpret direction and strength, and state one limitation of r.
- Define the variable, units, groups, and requested distribution or model feature.
- Show the required calculation or graphical/model reasoning with labeled quantities.
- Interpret the numerical result in the context of the data rather than as an isolated number.
- Identify an unusual feature, limitation, or condition that affects the conclusion.
Model response
Scatterplots and Correlation: Direction, Form, Strength, and Outliers — FRQ set 6: Scatterplots and Correlation: The association is strong and negative: larger x values tend to occur with smaller y values. Correlation describes linear association only, is sensitive to outliers, has no units, and does not establish causation.
Continue with the next connected AP Statistics skill
The most useful next step is to connect this topic to a neighboring method rather than repeating the same question type indefinitely.
