Mode vs Mean

The mode is found by tallying rather than by arithmetic: count how many times each value or category occurs, then report the one carrying the largest count. It has no standard symbol, and unlike the mean it works on categorical data, because counting never requires the values to be numbers.

In the data set 4, 5, 5, 5, 7, 8, 9 the counts are one 4, three 5s, one 7, one 8 and one 9. The largest count is 3, so the mode is 5. Answering that the mode is 3 is the classic mistake: 3 is the frequency, and the mode is the value that has it.

A data set can have no mode, one mode, or several. In 2, 4, 6, 8 every value occurs exactly once, the counts are all tied, and the convention is to say the set has no mode rather than to call all four values modes. That case is routine with measured quantities such as heights or reaction times, where exact repeats are rare. In 1, 1, 4, 7, 7, 9 both 1 and 7 occur twice, so the set has two modes.

Once data are grouped the mode becomes the modal class, the tallest bar of an equal-width histogram, because individual values have been absorbed into bins. That is also why bimodal describes a shape, two clear peaks separated by a dip, rather than a tie in the counts. Two peaks in a histogram of real data almost never have exactly equal heights, and a bimodal shape usually means two groups have been mixed together.

The mode is the only measure of center that survives nominal data such as favorite color or blood type, which is where it earns its keep. For quantitative data AP Statistics leans on the mean and the median instead: the mean uses every observation, and the median uses the position of every observation in the sorted list, while the mode can be created or destroyed by a single repeated value and says nothing about the rest of the distribution. In 4, 5, 5, 5, 7, 8, 9 the mode of 5 happens to equal the median, but nothing forces that.

The sample mean is $\bar{x} = \frac{\sum x_i}{n}$ ("x-bar"), the sum of the values divided by how many there are. Compute the same thing for a whole population and it is called $\mu$ ("mu"). The arithmetic is identical; what differs is whether the numbers came from a sample or from everyone. Balance point is the exact version of the intuition: the deviations from the mean always add to zero, for every data set without exception.

For 2, 5, 5, 8, 10 the sum is 30, so $\bar{x} = 30/5 = 6$. The deviations are $-4$, $-1$, $-1$, $2$, and $4$, and they cancel to 0: the pull from below the mean exactly matches the pull from above. Now change the 10 to 40. The sum becomes 60 and the mean becomes 12, while the median has not budged from 5. Every value contributes its full size, so a single value moving 30 units drags the mean $30/5 = 6$ units after it.

Here is the sentence to drop: "the mean is the value half the data fall below." That is the median. In 2, 5, 5, 8, 40 the mean is 12 and four of the five values sit below it, 80 percent of the data. Nothing makes the mean split a data set evenly, and the more skewed the data, the further from an even split it lands.

The everyday version of the same error is averaging averages. A class of 10 students averages 75 and a class of 30 averages 85, so the combined average is 80. It is not. The 30 students count three times as heavily, and the real figure is $\frac{10(75) + 30(85)}{40} = \frac{3300}{40} = 82.5$. Means combine correctly only through their totals, which is exactly what a weighted mean does.

The mean also needs values it makes sense to add. Averaging jersey numbers or zip codes is arithmetic with nothing behind it. And the mean need not be a value the variable can take: 1.75 siblings is an ordinary answer. Summary statistics for one quantitative variable are Unit 1 topic 1.7.