Error bars show the range of uncertainty around your data points
Error bars are the short lines that extend above and below (or left and right of) each data point on a chart. They represent how much your measurement might be off — whether because of measurement error, variation in your sample, or natural fluctuation in what you're measuring. In Excel, you add them to an existing chart in a few clicks, and you can set them to show a fixed value, a percentage, or a standard deviation calculated from your data.
The most common reason to use error bars is to show that your data is not exact. If you're charting average sales by month, error bars might show the range of individual sales around that average. If you're plotting test results, they might show the margin of error in your measurement. They tell anyone reading your chart: "This point is roughly here, but it could reasonably be somewhere in this range."
Key Takeaways
- Error bars are added to a chart that already exists, not to the raw data in your spreadsheet.
- You can set error bars to a fixed value, a percentage of each point, or a standard deviation based on your data.
- Right-click the data series (the line or bars themselves) and select "Format Data Series" to access error bar options.
- Different data points can have different error bar values if you use a data column to define them.
How to add error bars to a column or bar chart
Start with a chart that is already in your spreadsheet. Click on the chart to select it, then click on the data series itself — the bars or columns you want to add error bars to. Right-click and choose "Format Data Series" from the menu that appears.
In the Format Data Series panel that opens on the right side of the screen, look for the "Error Bars" option. Click it to expand that section. You will see a dropdown that says "Error Amount" with options like "Fixed value", "Percentage", "Standard Deviation", and "Standard Error". Choose the one that matches what you want to show. If you pick "Fixed value", type the number you want the error bars to extend (for example, 5 means they go 5 units up and 5 units down). If you pick "Percentage", enter a percentage like 10 for ten percent of each data point's value.
Once you have set the error amount, the bars appear on your chart when ready. You can adjust the width, color, and style of the error bars in the same panel if you want them to look different from the default.
Using a data column to set different error values for each point
Sometimes different data points should have different error ranges. You might have more confidence in some measurements than others, or the natural variation might be larger for some values. In that case, create a column in your spreadsheet with the error value for each row — one error value per data point, in the same order as your chart data.
In the Format Data Series panel, choose "Custom" from the Error Amount dropdown. Excel will ask you to select the range of cells that holds your error values. Click the small icon next to the input box, then click and drag to select your error column. Click the icon again to confirm. Now each point will have its own error bar based on the value in that column.
This approach is useful when you have calculated the error or uncertainty separately for each measurement, or when you want to show confidence intervals that vary across your data.
Error bars on line charts and scatter plots
The process is the same for line charts and scatter plots. Click the chart, then click on the line or the data points themselves. Right-click and select "Format Data Series". The Error Bars section works identically — you choose your error amount type and enter the value or select your data column.
On a scatter plot, error bars can extend both horizontally and vertically, which is useful when both your X and Y values have uncertainty. In the Format Data Series panel, you can set the vertical error bars and then look for an option to add horizontal error bars as well. Not all chart types support both directions, but scatter plots do.
Choosing between fixed values, percentages, and standard deviation
A fixed value works when you know the exact margin of error — for example, if your scale is accurate to plus or minus 2 pounds, you would use 2 as the fixed value. This is straightforward but only works if the error is the same size across all your data points.
A percentage is useful when the error grows with the size of the data. If you measured something with 5 percent accuracy, you would use 5 percent. A measurement of 100 would have an error bar of 5, while a measurement of 200 would have an error bar of 10. This reflects the reality that larger values often have larger absolute errors even if the percentage error stays the same.
Standard deviation is the right choice if you have raw data and want to show how spread out your measurements are. If your data column contains multiple measurements for each point (or if you have calculated the standard deviation separately), this option will display that variation. Standard Error is similar but divides by the square root of the sample size, which is useful for showing the uncertainty in an average.
Removing or hiding error bars
If you add error bars and then decide you do not want them, click the chart, then click on the error bars themselves. Press Delete. The error bars disappear but your chart data stays intact.
If you want to hide them temporarily without deleting them, right-click the error bars, choose "Format Error Bars", and set the line color to "No Line". The error bar values are still there — you can turn them back on the same way — but they will not show on the chart.
Common mistakes and how to avoid them
The most common mistake is trying to add error bars to the data in your spreadsheet instead of to the chart. Error bars only exist on charts, not on the raw numbers. You must create your chart first, then add the error bars to it.
Another mistake is clicking the wrong part of the chart. If you right-click the chart background or the axis labels, you will not see the error bar option. You need to right-click the actual data series — the bars, line, or points themselves. If you are not sure, click once on the chart to select it, then click directly on one of the bars or data points.
A third mistake is forgetting that error bars show uncertainty, not the range of your raw data. If you have sales numbers from 10 to 50, do not set error bars to show that range. Instead, use error bars to show measurement error, sample variation, or confidence intervals — the amount of uncertainty in each individual point.
Frequently Asked Questions
Can I have error bars on only some of the data series in a chart?
Yes. Click the chart, then click on the specific data series you want to add error bars to. Right-click and format that series alone. Other series in the same chart will not be affected. This is useful when you have multiple lines or groups of bars and only some of them have uncertainty data.
What if my error values are in a different sheet?
When you select the Custom error amount option and are asked to pick a range, you can type the range directly instead of clicking and dragging. Use the format SheetName!A1:A10 to reference cells in another sheet. Make sure the range has the same number of cells as your data series.
How do I make error bars extend in only one direction?
In the Format Data Series panel, look for options labeled "Plus" and "Minus" under the Error Bars section. You can set one to a value and the other to zero, or set them to different values if your uncertainty is not symmetric.
Can error bars show different values above and below the same point?
Yes, if you use the Custom option with a data column. Create two columns — one for the plus direction and one for the minus direction — and select both when prompted. This is useful for showing asymmetric confidence intervals or when your error is larger in one direction than the other.