Calculate and interpret Pearson correlation

Try

A synthetic experiment records five paired inputs and responses: (0,1),(1,2),(2,4),(3,4),(4,6). Each row belongs to one observation. Plotting the pairs suggests a positive relationship, but the points do not fall exactly on a line. Pearson correlation measures the direction and strength of linear association after centering and scaling both variables.

A coefficient alone cannot reveal curvature, a broken pairing or a causal mechanism. We will enter actual paired data, inspect the scatterplot, compute the coefficient and independently check its intermediate arithmetic.

Predict

Predict the sign of correlation before using the calculator. Would adding 100 to every response change it? Would converting the input to a positive multiple change it? Now imagine sorting the response column independently: that would create different pairs, not simply reorganize the same observations.

Experiment

Fit, inspect residuals and challenge the model

Each row is one paired observation. Predict a model family and its residual pattern before revealing calculations. The default data are synthetic. State the units and sampling process supplied by the surrounding investigation; the tool itself does not assign physical units.

Paired observations, one x,y pair per line0,2 1,4 2,7 3,10 4,15 5,21

Use 3–40 pairs of decimal numbers, each between −100 and 100. Rows may repeat; they are separate observations, not automatically independent samples. Pairing and observation order are retained.

Model to inspectLinearQuadraticExponential (log-linear fit)Reset paired dataTry noisy linear dataTry constant yShow paired-data calculations

Record the raw pairs, model criterion, equation, residual display and context-dependent conclusion in your practice response. Computations use floating-point arithmetic; values are rounded for display. Editing data or selecting a model hides old results. This temporary tool neither saves responses nor awards mastery.

Replace the initial data with the five worked pairs, one row per pair. Predict, reveal and record the scatterplot and Pearson r. The workbench also shows model fits; for this activity focus on correlation and the paired raw rows. Check that every entered row represents the intended observation before interpreting any number.

For an arithmetic check, the means are x̄=2 and ȳ=3.4. Write centered differences and their products:

x,y	x−2	y−3.4	Product
0,1	−2	−2.4	4.8
1,2	−1	−1.4	1.4
2,4	0	0.6	0
3,4	1	0.6	0.6
4,6	2	2.6	5.2

The product sum is 12, the centered x squared sum is 10, and the centered y squared sum is 15.2. Therefore r=12/√152≈0.9733. Retain enough digits to compare the calculation, but do not describe the population association as known to four decimal places from five synthetic observations.

Understand

Write Sxx for the sum of squared centered x values, Syy for the corresponding y sum, and Sxy for the sum of centered cross-products. Then r=Sxy/√(SxxSyy), provided both squared sums are positive. Positive products occur when a pair is above both means or below both; opposite deviations contribute negatively. The normalization removes measurement units and places r between −1 and 1.

Adding a constant to a variable disappears after centering. Multiplying it by a positive constant multiplies numerator and denominator by the same factor, leaving r unchanged. Multiplying exactly one variable by a negative constant reverses its sign. A constant variable has no variation: its squared sum is zero, so r is undefined, not evidence of zero association. The constant preset exposes that boundary.

A zero coefficient can coexist with a strong nonlinear relationship. Try the five pairs with x=−2,−1,0,1,2 and y=x². The positive and negative cross-products cancel by symmetry, while the U-shaped scatterplot is exact. Correlation addresses linear association, not every possible dependence. Conversely, one unusual far-away pair can strongly change r; inspect the scatterplot and source data rather than deleting a point solely to improve a result.

The source of the pairs matters as much as the formula. Hours studied and scores may be associated with prior preparation, access to resources or selection into the sample. A large positive r does not prove that changing one variable causes the other to increase. Randomized interventions and careful design answer different questions from a descriptive coefficient.

Apply

For the fresh paired dataset in practice, record the technology input, plotted points, coefficient and the three centered sums. State the variables and units, describe the shape, and name a limitation on causal or population interpretation. If a row seems wrong, check its original record and document any correction.

A reproducible workflow keeps whole rows together when sorting or filtering. Stable observation IDs help detect accidental mismatching. The workbench is temporary, so copy the requested settings and results into your written response; revealing a graph does not itself save evidence or award understanding.

Continue

Next fit competing model families and inspect residual plots. Correlation describes a linear association, while residuals help identify what a particular predictive equation leaves unexplained.