Let's say your data is all in column A. Furthermore, we'll say we have 21 items in this column. Make sure this data column is sorted from smallest to largest. If this isn't done, then weird things will happen when it comes to connecting the dots later on.
Column B will be used later, but for now, let's move to cell C1. In this cell, type =AVERAGE(A1:A21)
The equal sign is important to tell Excel that we have a function here, and not just plain text. As the function implies, we're finding the average (aka mean) of values A1,A2,...,A20,A21. This computes the arithmetic mean.
In cell D1, type in =STDEV(A1:A21) to compute the sample standard deviation
There's nothing special about C1 or D1. We just need something off to the side. Think of it like scratch work.
-------------------------
Now move back to column B. In cell B1, type =NORMDIST(A1,$C$1,$D$1,FALSE)
The $C$1 and $D$1 refer to cells C1 and D1 respectively. The dollar signs lock in those rows and columns so they don't change. We don't want the mean and standard deviation to alter. The "False" at the end tells Excel that we don't want a cumulative normal distribution.
The first parameter A1 is the x value. Much like the curve y = x^2, we plug in various x values to find corresponding y values. We'll plot various (x,y) points to connect them with a curve. The more points, the more accurate the curve. Instead of the x^2 function, we're using the NORMDIST function.
Once cell B1 is filled out, click on the bottom right corner of the cell. You should see a very small square at this corner. Dragging this square down until you reach cell B21 will fill in the remaining 20 cells with the formula mentioned in bold; however, the A1 will change depending on what row you're on. Everything else stays the same hence the use of the dollar signs.
-------------------------
In figure 1 below, I'm showing an example of what I've discussed so far. Figure 1 shows the formulas in each cell. This is before they are evaluated.
Figure 2 shows what the formulas evaluate to. The data in column A can be anything you want. I randomly generated the data values between 2 and 17.
Once column B is figured out, you would insert a chart of (x,y) scatter points. Make sure to connect the dots so that a curve forms. This will complete the normal distribution graph. Again, the data in column A must be sorted or else the dots won't be connected properly.
To be perfectly honest, Excel doesn't really do a good job at drawing curves like this. In my opinion, programs like GeoGebra are better suited for the task. Though spreadsheet programs like Excel are still used in a lot of settings, which is why this is good practice.
