In simple words, the standard error bars attempt to show the range of "likely values", but not the more extreme values. . . Example 1) In measurements, there are often slight errors.. . For instance, a coach who uses a stopwatch to time a runner at 45.3 seconds may find that the actual time is probably between 45.2 and 45.4 seconds.. . Example 2) In sampling data, there are slight variations due to the random nature of sampling (a particular sample of people may have slightly higher blood pressures than another sample).. . Suppose a doctor examines 50 patients and records their blood pressure. The average from that sample can be used to predict the average blood pressure of a much larger population without examining everyone in the larger group.. However, when she graphs the average from that sample, she knows that the predicted average of the larger group might be a little lower or a little higher than the value from the sample of the particular 50 patients. The standard error bars attempt to show this likely variation.. . The amount of "likely" deviation is called a standard deviation, and there are formulas for calculating that value.