To calculate an n-point moving average, sum the first values and divide by . For the next value, drop the first data point from the group, add the point, and calculate the new mean.
The general formula for the moving average in an -point series is: where represents the individual data observations.
Plotting Position: Each moving average must be plotted at the midpoint of the time interval it covers. For an even number of points like a 4-point average, the midpoint falls exactly between the second and third time periods in that specific group.
| Feature | Raw Time Series Data | Moving Averages |
|---|---|---|
| Visual Appearance | Jagged, showing all fluctuations | Smoother, showing general direction |
| Graphing Rule | Points are connected by lines | Points are NOT connected to each other |
| Plotting Location | Exactly on the time coordinate | At the midpoint of the time interval |
| Purpose | Shows exact values at specific times | Identifies long-term trends and patterns |
Check the Interval: Always ensure the number of points in your moving average matches the cycle described in the problem (e.g., use 4 points for quarterly data or 12 points for monthly data).
Midpoint Accuracy: A common mistake is plotting the moving average directly above a specific time period. Always calculate the halfway point; for a 4-point average of periods 1, 2, 3, and 4, the point must be plotted at .
The 'No-Join' Rule: Examiners frequently penalize students for connecting moving average points with a jagged line. Use these points only to guide the placement of a single straight trend line.
Calculation Verification: If the raw data is generally increasing, your moving averages should also show an increasing pattern. If a moving average value drops significantly while the data is rising, re-check your summation.
Over-smoothing: Using a window () that is too large can hide important trend changes or 'lag' the trend significantly behind the actual data. Conversely, a window that is too small fails to remove the seasonal noise.
End-Point Loss: Note that you will always have fewer moving average points than raw data points. You cannot calculate a 'full' moving average for the very beginning or very end of a data set because there aren't enough preceding or succeeding points to fill the window.
Misinterpreting the Trend: A moving average shows where the data has been on average. While it helps identify a trend, it does not automatically guarantee that the trend will continue into the future without further analysis.
