Lines of Best Fit & Regression Lines

• The Double Mean Point, denoted as $(\bar{x}, \bar{y})$, is the central anchor of any accurate line of best fit. It is calculated by finding the arithmetic mean of all $x$-coordinates and all $y$-coordinates in the data set; the regression line must pass through this point.

• The Least Squares Principle is the mathematical foundation for regression lines. It works by minimizing the sum of the squares of the vertical distances (residuals) between each data point and the line, ensuring the line represents the 'average' path of the data as accurately as possible.

• The equation of a regression line is typically given in the form $y = a + bx$. In this context, $a$ represents the y-intercept (the value of $y$ when $x=0$) and $b$ represents the gradient or slope (the change in $y$ for every one-unit increase in $x$).

Drawing a Line of Best Fit by Eye

• Use a ruler to draw a single straight line that passes through the center of the data cluster. Ensure there are roughly an equal number of points above and below the line, and that the line follows the general direction of the correlation.

• If the double mean point $(\bar{x}, \bar{y})$ is known, the line must be drawn specifically through this coordinate. This provides a more objective starting point than simply guessing the center of the cluster.

Using the Regression Equation

• To plot a regression line from its equation $y = a + bx$, select two distinct $x$-values within the range of the data. Substitute these values into the equation to find their corresponding $y$-values, plot these two points, and connect them with a straight line.

• To make a prediction using the line, substitute the desired $x$-value into the equation. This is generally more accurate than reading a value off a hand-drawn graph, as it eliminates human error in visual estimation.

• Interpret in Context: When asked to explain the gradient $b$, always use the units of the axes. For example, if $x$ is 'hours studied' and $y$ is 'test score', the gradient represents the 'increase in test score per hour of study'.

• Check the Intercept: The y-intercept $a$ represents the value of $y$ when $x=0$. Always evaluate if this value makes sense in a real-world context; sometimes the intercept is a theoretical starting value (like a fixed fee) even if $x=0$ is not possible in practice.

• Identify Outliers: Before drawing a line of best fit, look for points that fall far away from the general pattern. These outliers should usually be ignored when positioning the line to prevent them from skewing the trend.

• Sanity Check Predictions: If a prediction seems physically impossible (e.g., a negative height or a price of zero for a luxury item), check if you have used extrapolation or if the linear model is simply not suitable for that range.