Microsoft Office Tutorials and References

In Depth Information

**Deviating from the Middle**

Skewed from the norm

There is deviation in a distribution, but who says the deviation has to be

uniform with deviation the same on both sides of the mean? Not all distributions

are normal — some are
skewed,
with more values clustered either below the

mean or above it:

✓
When more values fall below the mean, the distribution is
positively

skewed.

✓
When more values fall above the mean, the distribution is
negatively

skewed.

The following minitable has a few examples:

Values

Mean

Comment

1, 2, 3, 4, 5

3

No skew. An even number of values fall above

and below the mean.

1, 2, 3, 6, 8

4

The distribution is positively skewed. More

values fall below the mean.

1, 2, 8, 9, 10

6

The distribution is negatively skewed. More

values fall above the mean.

Figure 9-10 shows a distribution plot, where 1,000 values are in the

distribution, ranging between 1 and 100. The values are summarized in a table of

frequencies (discussed later in this chapter). The table of frequencies is the

source of the chart.

Figure 9-10:

Working

with skewed

data.