Statistical Analysis and Standard Error in Scientific Error Calculation

The mean is the basic average of a data set. To calculate it, all measured numbers are added together, and the sum is then divided by the total number of measurements taken.
On standard equation sheets, the mean is represented using the summation symbol (sigma) to indicate adding up all measured values.
The formula for the mean can be expressed using $n$ (the number of data points) or as a fraction $1/n$ multiplied by the sum of the measurements.
Calculating the average is the first step in statistical analysis and provides the value used in subsequent calculations for standard deviation and standard error.
If calculating a mean for grades, for example, the process remains the same: sum the grades and divide by the number of individual grades recorded.

Standard deviation is a measure of the amount of variation or dispersion in a set of values. While the calculation process is multi-step, the math itself relies on fundamental arithmetic.
The process for calculating standard deviation involves several specific steps:
- Determine the difference between each individual measurement ( $x$ ) and the calculated mean ( $\bar{x}$ ).
- Square each of those individual differences.
- Add all of the squared differences together ( $\sum(x - \bar{x})^2$ ).
- Divide the resulting sum by the total number of measurements minus one ( $n - 1$ ).
- Take the square root of that entire quotient.
The formula for standard deviation ( $s$ ) is expressed as: $s = \sqrt{\frac{\sum(x - \bar{x})^2}{n - 1}}$
Using a specific dataset of eight values where the mean was calculated as $232.5$ , the step-by-step application is as follows:
- For a value of $241$ : $(241 - 232.5)^2 = 72.25$
- For a value of $236$ : $(236 - 232.5)^2 = 12.25$
- The sum of all these squared differences for the eight trials resulted in a numerator of $576$ .
- Since there were eight values ( $n = 8$ ), the divisor is $n - 1 = 7$ .
- The calculation proceeds as $576 / 7$ , followed by taking the square root of the result.
- The final standard deviation for this example is approximately $9.1$ .

Standard error is a statistical term that measures the accuracy with which a sample distribution represents a population. It quantifies how certain one can be that the average obtained would remain the same if the experiment were repeated.
To calculate the standard error, the standard deviation ( $s$ ) is divided by the square root of the number of samples ( $n$ ): $\text{Standard Error} = \frac{s}{\sqrt{n}}$
Using the previous example values where $s = 9.1$ and $n = 8$ :
- Calculate the square root of $8$ .
- Divide $9.1$ by $\sqrt{8}$ .
- In the example provided, the standard error results in approximately $3.2$ .
The standard error provides a range of certainty:
- One Standard Error (±1 SE): There is approximately a $68\%$ confidence level that the true mean of the population falls within this range.
- Two Standard Errors (±2 SE): By doubling the standard error, the confidence level increases to approximately $95\%$ . This suggests that the actual average is very likely to fall within this broader range.

When graphing data, the mean is plotted as the primary point or bar height.
Error bars are added to the graph using the standard error values to represent the range of uncertainty.
To correctly represent standard error bars for a $95\%$ confidence interval, the standard error must be doubled ( $2 \times \text{SEM}$ ).
Error bars must be drawn in both directions: adding the value to the mean and subtracting it from the mean ( $\pm 2 \text{SE}$ ). Failure to include both the top and bottom of the bar is a common error in data visualization.

The most critical application of standard error and graphing is identifying statistical differences between datasets.
Overlapping Error Bars: If the error bars of two different data sets overlap, there is no statistical difference in the data. This means that if the experiment were repeated with a larger sample size, the averages could potentially end up being identical.
No Overlap: If the error bars do not overlap, the data sets are considered statistically different. In such cases, one sample can be definitively described as being greater than or less than another.
Care must be taken with terminology: when comparing separate groups that show no overlap, it is more accurate to say one is "greater than" the other rather than saying it "increased," as the groups represent separate entities rather than a progression of the same group.

How do you handle the calculator order for standard deviation? It is recommended to follow the order of operations strictly when using a calculator. This involves subtracting the mean from the value first, squaring that specific result, repeating for all values, and finally summing them. For the final steps, one should divide by $n - 1$ (which was $7$ in the example of $8$ trials) before applying the square root.
Where did the number seven come from in the calculation? The number seven is the result of $n - 1$ . Since there were eight values measured, the formula requires dividing by $8 - 1$ . If an experiment has three trials, this number would be two.
What was the sum used in the standard deviation example? The sum of the squared differences (the top part of the equation) was $576$ .
What was the final standard error for the fiddler crab example? After dividing the standard deviation of $9.1$ by the square root of $8$ , the simplified standard error was recorded as $3.2$ .