A positive slope rises from left to right; a negative slope falls. Zero describes a horizontal line. Larger absolute values indicate more vertical change per horizontal unit, but comparisons make sense only with compatible axis scales.
Slope between two observations is a secant rate of change. It does not prove the relationship is linear between or beyond them. For measured data, uncertainty and rounding affect the result, especially when x values are close. A fitted regression slope uses all selected observations and is a different calculation.
The line through two points can be written in point-slope form as y − y₁ = m(x − x₁). Once m is known, substituting either point gives an equation for the same line. In slope-intercept form y = mx + b, calculate b as y₁ − mx₁. Substituting the second point is a useful independent check.
When slope represents a rate, direction matters. A result of −3 dollars per item means the modeled y value falls by three dollars for each additional item; it is not merely a steepness of three. A reciprocal slope answers a different question and is valid only when the original numerator is nonzero.
Close x-coordinates make the denominator small, which magnifies coordinate error. Record enough significant digits and consider an uncertainty analysis for measured points. In a graph, axis scaling can make the same numeric slope appear visually steep or flat, so read the labels rather than judging from appearance.
Parallel nonvertical lines share the same slope. Perpendicular nonvertical lines have slopes whose product is −1, provided neither is horizontal or vertical; the perpendicular slope is the negative reciprocal. These rules offer useful checks in geometry but still require attention to exceptional lines. Plotting both points provides another quick check of direction and sign.
Keep the original inputs beside the result for a clear, reproducible record.