Scatter plots address the very important question of correlation. How do these two variables interact? When to use a scatter plot can be confusing because you essentially have two dependent variables. Although you will measure each separately, you are curious as to what their relationship is. Defining that relationship and showing a pattern in the data (or not) will establish a correlation between your two variables.
Correlation
- Positive correlation: As one variable increases, the other also increases (points go up and to the right).
- Negative correlation: As one variable increases, the other decreases (points go down and to the right).
- No correlation: There’s no discernible pattern between the variables (points are scattered randomly).
Detecting Outliers
Trendlines
A trendline is different than a line graph. In a line graph, you connect each data point. A trendline (also called a line of best fit) on a scatter plot is a straight or curved line that best represents the relationship between the two variables being plotted. The trendline is drawn to show the general direction or pattern of the data points, helping to visualize correlations or trends more clearly.
- Linear Trendline: A straight line used when the relationship between the two variables is approximately linear (i.e., the data points generally form a straight line). For a linear trendline, the line is often fitted using least squares regression, which minimizes the distance (or errors) between the data points and the line, ensuring the line represents the best fit.
- Exponential, Polynomial, or Logarithmic Trendlines: These are used when the relationship is nonlinear, meaning it follows a curved pattern rather than a straight line. For nonlinear trendlines, different methods (like polynomial fitting or exponential regression) are used depending on the nature of the data.
- Moving Average Trendline: This smooths out short-term fluctuations and highlights longer-term trends by averaging the data over a set number of periods.
Example:
If you’re plotting data on the number of hours studied (x-axis) and exam scores (y-axis), a trendline could help indicate whether there is a general upward trend, where more study hours lead to higher exam scores. If you apply a linear trendline, it might show a straight line that fits the data points, making it easier to interpret the overall relationship.