Error Bars on an Excel Scatter Plot: A Clear Guide

In Microsoft Excel for Microsoft 365 on Windows, select the scatter chart, then choose Chart Design > Add Chart Element > Error Bars > More Error Bars Options. In the Format Error Bars pane, choose an error amount, or select Custom > Specify Value to use worksheet ranges for positive and negative errors.

What error bars show

Error bars are lines drawn from data points to indicate variation or uncertainty. On a scatter plot, vertical bars describe variation in the Y values, while horizontal bars describe variation in the X values. A chart can show one direction or both. The bars do not explain themselves: the reader needs to know what they represent, such as standard deviation, standard error, a confidence interval, or measurement uncertainty.

That distinction matters. A standard deviation describes spread among observations; a standard error describes uncertainty in an estimated mean; and a confidence interval is an interval calculated using a chosen method and confidence level. Excel’s built-in error-bar choices are not interchangeable, and the chart does not automatically label which interpretation is appropriate. State the measure in the chart title, axis note, caption, or accompanying text.

Add error bars to a scatter chart

Start with a chart that has the correct X and Y series. To create one, select the X and Y data, then use Insert > Insert Scatter (X, Y) or Bubble Chart > Scatter. If the chart already exists, click its plot area or data series to show the chart tabs.

Select the scatter chart.
On the ribbon, choose Chart Design > Add Chart Element > Error Bars > More Error Bars Options.
In the Format Error Bars pane, choose the error-bar series and direction you want to edit. Click a horizontal or vertical bar on the chart to select that set; the pane updates for the selected bars.
Under Error Amount, choose Standard Error, Percentage, Standard Deviation, Fixed value, or Custom, depending on the measure you need.
For Custom errors, choose Specify Value and select the cells for the positive and negative error values.

Excel may add vertical error bars when you first use the command. If you need horizontal bars as well, select the X error bars and configure them separately. If a command or pane is difficult to find, confirm that the chart—not a cell or worksheet object—is selected. On current Microsoft 365 for Mac, chart-element controls are also available from the Chart Design tab, though pane layout and selection methods can differ. Excel for the web may not expose every custom error-bar control; open the workbook in the desktop app if you cannot set the ranges.

Choose the right error-bar type

Excel’s built-in options are convenient for quick charts, but their meaning depends on your data and question. Choose the measure before formatting the chart, rather than selecting an option simply because it makes the bars look reasonable.

Standard deviation: useful for showing the spread of a data series around its mean. It is not automatically a separate uncertainty estimate for each plotted point.
Standard error: useful when the plotted values represent estimates of means and you want to show the standard error calculated by Excel for the series. It is not the same as showing the spread of individual observations.
Percentage: draws bars as a percentage of the plotted value. It is a proportional display, not a statistical confidence interval.
Fixed value: applies the same error amount to each point. Use it when a constant uncertainty is justified, such as a stated instrument resolution.
Custom: uses values you supply and is usually the best choice for point-specific measurements, asymmetric intervals, or intervals calculated elsewhere.

For example, if a plotted value is 20 and you choose a 10 percent error amount, Excel uses a length of 2 for that point. This setting does not calculate a 10 percent confidence interval from the underlying observations. Likewise, Excel’s Standard Deviation setting should not be presented as a confidence interval unless you have separately performed the statistical calculation required for one.

Use custom values for different errors at each point

Custom error bars let you provide a different amount for every data point. Prepare one worksheet range for positive errors and another for negative errors, with the values in the same order as the points in the plotted series. In the Format Error Bars pane, choose Custom > Specify Value, then select the appropriate range in each field. The ranges should contain numeric error amounts and should match the number of points in that series.

These cells are lengths, not the upper and lower endpoints of an interval. Suppose a point is at Y = 12, its lower limit is 10, and its upper limit is 15. The negative error is 2 and the positive error is 3. Entering 10 and 15 would make bars much larger than intended because Excel adds or subtracts the entered amounts from the point. For a horizontal bar, calculate the distances in X units; for a vertical bar, calculate them in Y units.

Asymmetric errors are valid. A confidence interval might extend 2 units below a point and 3 units above it, for example. Use separate positive and negative ranges to preserve that shape. If the interval represents a particular statistical method, calculate it using that method first; Excel’s custom setting displays supplied values but does not calculate their statistical validity.

A small example: standard error

Suppose five repeated measurements are 8, 10, 9, 11, and 12, and you want to calculate the standard error of their mean. With the five measurements in cells B2:B6, enter =STDEV.S(B2:B6)/SQRT(COUNT(B2:B6)) in another cell. The result is approximately 0.7071: the sample standard deviation is about 1.5811, divided by the square root of 5.

This example calculates one standard error for those five measurements. It does not mean that every point in a different scatter series should receive that same value. If each plotted point comes from a separate set of replicates, calculate an appropriate error for each set, place the results in a range aligned with the plotted points, and use Custom error bars. Also distinguish a chart of individual measurements from a chart of group means: they call for different interpretations.

Format and interpret the bars

After choosing the error amount, adjust the appearance so the bars are legible without overpowering the data. In the Format Error Bars pane, options include direction, end style, and line formatting. Direction controls whether the bar extends in both directions or only toward the positive or negative side; end style controls whether caps appear at the ends. A cap is a visual choice, not a different statistical calculation.

Keep the scale in mind. A bar of a given length represents units on its own axis, so an X error of 0.5 and a Y error of 0.5 need not look equally long if the axes use different scales. Avoid changing axis limits merely to make the bars look more prominent. If several series appear on one chart, use clearly distinguishable formatting and make the legend or caption explain which bars belong to which series.

Where readers could mistake one measure for another, write a concise note such as “bars show ±1 standard error” or “bars show the observed measurement range.” A plotted interval alone does not tell a reader whether it is a standard deviation, standard error, confidence interval, or instrument tolerance.

Common problems and fixes

Only vertical bars appear: select the horizontal X error bars and configure them separately. Scatter charts can show errors in both directions, but adding one set does not guarantee that the other is present.
Custom values are rejected: check that each selected range contains numeric cells, has the right number of entries, and is ordered to match the series points. Select the error amounts, not the original measurement values or interval endpoints.
Bars look enormous or point the wrong way: verify that the cells contain distances from each point, rather than upper or lower endpoint coordinates. Also check that X errors use X-axis units and Y errors use Y-axis units.
The wrong bars change when formatting: click the intended horizontal or vertical bars before editing, or use the chart’s current-selection controls when available. If the chart contains multiple series, make sure the selected error bars belong to the intended series.
Bars are hard to see: check whether they are hidden behind markers, use a contrasting line color, and review the axis scale. A very small error can be correct even when it is barely visible.

When the data changes, custom bars linked to worksheet cells update with those cells. If you add points to the series, check that the custom ranges cover the new points too; a range that no longer matches the series can produce an error or omit the intended values.

When error bars are not the best choice

Error bars are compact, but they can become difficult to read when many points overlap or when the distribution is strongly asymmetric. Consider whether the reader needs to see the underlying observations or the shape of the distribution, rather than only a summary around each point.

Show individual observations alongside group means when the sample size is small and the raw values are informative.
Use a box-and-whisker chart when the goal is to compare distributions, medians, and spread across groups.
Use a confidence-interval plot when the main comparison is between estimated values and their calculated intervals.
Use a table or a separate chart when several overlapping error-bar series make the scatter plot unclear.

These alternatives answer different questions; none is automatically more correct. Choose the display that matches how the data were collected and what the plotted estimate represents.

Frequently Asked Questions

Can an Excel scatter plot have both horizontal and vertical error bars?
Yes. Horizontal bars represent uncertainty or variation in X, and vertical bars represent it in Y. Add and format each direction separately, using values measured in the corresponding axis units.

How do I add different error bars to each point?
Choose Custom > Specify Value in the Format Error Bars pane. Select worksheet ranges containing positive and negative error amounts in the same order as the points in the series.

Why are my error bars too large?
Custom ranges should contain distances from the plotted values, not the upper and lower limits themselves. Also check that the values use the correct axis units and that the chart scale has not changed unexpectedly.

Are Excel standard-error bars confidence intervals?
No. Standard error and a confidence interval are related but are not the same thing. A confidence interval requires a defined calculation and assumptions; Excel’s Standard Error option does not automatically label or calculate a chosen confidence level.

Can I add custom error bars in Excel for the web?
Available chart-formatting controls can vary in Excel for the web. If the custom range fields are unavailable, open the workbook in Microsoft Excel for Windows or Mac desktop to configure the bars.

Try it: the Formula Explainer: paste any formula and see each function and reference in plain English.

More in Errors & fixes

Sources: Function behaviour and menu paths: Microsoft Support documentation for Excel. Method & sources

Spotted an error in this guide? Let us know.