|
PAGE
FOUR
A Basic
Guide to Statistics
by Jonathan Dolhenty, Ph.D.
The
Famous Bell Curve
Most of you have probably have some acquantance
with what is called a "bell" curve. The concept of
a normal curve is important in understanding the
use of statistics. Most direct measures used in the
study, for instance, of varying traits in human
beings and most psychological measures, such as IQ
scores, have been found to approximate closely a
mathematical model called the normal
distribution.
The graph of this normal distribution is a
continuous, symmetrical, bell-shaped curve.
Frequencies tend to concentrate around the median
and become fewer and fewer at either end, resulting
in a frequency curve which is high in the middle
and low at the ends.
The normal curve is a limited curve which is
approached by many distributions when a large
number of measurements is made or, as is often
said, when there is a large number of cases. It is
necessary to assume that these measurements or
cases are taken at random, or that there is no bias
or systematic error.
Of course, frequency distributions are not
always of the normal curve type and the term
skewness is used in describing abnormal
distributions. In a normal curve, the right and
left halves of the curve are mirror images of each
other. If this is not the case, the curve is said
to be skewed, either positively (to
the right) or negatively (to the left. Thus,
if the scores tend to be concentrated toward the
high end of the score scale, the curve is
negatively skewed. If they are concentrated
toward the low end of the score scale, they are
positively skewed.
A
Practical Illustration of the Normal Probability
Curve
One way to illustrate the normal curve is by
tossing coins or dice. If we represent "heads" by
H and "tails" by T, the expression
H+T represents the probabilities for any
toss of one coin, namely equal probabilities of a
head or a tail.
If we toss the coin 100 times, the results will
approximate 50 heads and 50 tails. If we toss two
coins, the possibilities are two heads, head and
tail, tail and head, two tails, or H2+2HT+T2.
If we toss two coins 100 times we would
theoretically get 25H2+50HT+25T2, or both coins
heads 25 times, one head and one tail 50 times, and
both coins tails 25 times.
Similarly, we can predict the theoretical
frequency with which each possible combination of
any number of coins tossed simultaneously any given
number of times will occur.
In tossing coins, if there is an equal chance
for each coin to fall head or tail each time, every
possible combination can occur, but the
probabilities of getting ten heads or ten tails
when we toss ten coins are less than those for
getting other combinations.
The most probable combination is five heads and
five tails, since each coin has an equal chance of
falling heads or tails.
By expanding (H+T)10, we get the probabilities
of each possible combination occurring if ten coins
were tossed an infinite number of times. The
expression becomes H10 + 10H9T + 45H8T2 + 120H6T4 +
252H5T5 + 210H4T6 + 120H3T7 + 45H2T8 + 10HT9 +
T10.
This means that the chances of getting five
heads and five tails in tossing ten coins are 252
in 1,024. The probabilities of getting ten heads or
ten tails are, respectively, one in 1,024. Where
ten coins have thus been tossed, it has been found
that the frequencies with which possible
combinations do occur approach the theoretical
values as limits.
The frequency distribution below shos the
results of tossing ten pennies 1,000 times, based
on an actual experiment.
Although the number of tosses is 1,000 instead
of 1,024, it is evident that the frequency with
which each possible combination actually occurred
closely approximated the theoretical values. The
resulting curve is very close to the theoretical
normal curve of distribution, the perfect "bell"
curve.
Another situation in which the typical
bell-shaped curve occurs is in measurement of
natural phenomena. Thousands of measurements have
been made of barometric and temperature readings
over a long period of time. The results approximate
the "bell" curve.
The frequency distribution below, a study of
height using over 8,000 adult males, also comes
very close to a normal "bell" curve.
To Page 5
Enrich
Your Life With a Philosophy Book...
Enrich
Your Life With a Philosophy
Magazine...
|
Academy
Showcase Specials
|
|
|
|
|
|
|