Polygons
To construct a polygon, you begin by listing the
A frequency distribution graphs is basically a picture of numerical scores along the X-axis, then:
information available in a frequency distribution table. We
start with two perpendicular lines called axes. The vertical a. A dot is centered above each score so the
line is called the Y- axis, or ordinate and the horizontal line is vertical position of the dot corresponds to the
called X-axis. The measurement scale (set of values) is listed frequency for each category
along the X- axis in increasing value from left to right. b. A continuous line is drawn from dot to dot to
Frequencies are listed on the Y-axis in increasing value from connect the series of dots.
bottom to top. As a general rule, the point where the two c. The graph is completed by drawing a line down
axes intersect should have a value of zero for both the scores to the X – axis (zero frequency) at each end of
and the frequencies. the range of scores. The final lines are usually
drawn so that they reach the X-axis at a point
Graphs for interval or ratio data
that is one category below the lowest score on
1. Histogram the left side and one category above the highest
score on the right side.
To construct the histogram, you first list the numerical
scores along the X-axis. Then you draw a bar above X e.g. using the same data:
value so that:
x f
a. The height of the bar corresponds to the frequency 6 1
for that category 5 2
b. The width of the bar extends to the real limits of the 4 2
category. 3 4
e.g. 2 2
1 1
x f
6 1
5 2
4 2
3 4
2 2
1 1
x f
5 4
4 5
3 3
2 1
1 1
The cumulative frequencies show the number of individuals P 50 = L.B. + N (.50) - < cf b i
located at or below each score. To find percentiles, we must
convert these frequencies into percentages. The resulting f
values are called cumulative percentages because they show
P50 = 59.5 + 42 (.50) – 12 5
the percentage of individuals who are accumulated as you
move up the scale. 9
P# = L.B. + N (.x) - <cf b i What is the score that corresponds to 87th percentile?
4
P87 = 74.5+ 36.54 - 35 5 Prank = 12 + 1.8 (5) X 100
4 42
P87 = 76.42 42
class 42
interval midpoint f < cf cum %
90 - 94 92 1 42 100.00 Prank = 50%
85 - 89 87 1 41 97.62 Prank of score 64.5 = 50 %
80 - 84 82 1 40 95.24
75 - 79 77 4 39 92.86
70 - 74 72 3 35 83.33
What is the percentile rank of score 58?
65 - 69 67 11 32 76.19
60 - 64 62 9 21 50.00 Prank = cumfb + (f/i) (X – XL) X 100
55 - 59 57 5 12 28.57
N
50 - 54 52 1 7 16.67
45 - 49 47 3 6 14.29 Prank = 7+ (5/5) (58– 54.5) X 100
40 - 44 42 1 3 7.14
42
35 - 39 37 1 2 4.76
30 - 34 32 0 1 2.38 Prank = 7+ 3.5 X 100
25 - 29 27 1 1 2.38
42 42
Prank = 25%
Prank = cumfb + (f/i) (X – XL) X 100
What is the percentile rank of score 76.42?
N
Prank = cumfb + (f/i) (X – XL) X 100
Cum fb – cumulative frequency before
N
F – frequency
Prank = 35 + (4/5) (76.42 – 74.5) X 100
I = class interval size
42
X = score
Prank = 35 + (.8) (1.92) X 100
XL = lower boundary of the class that contains the score
42
e.g. 2 3 3 4 7 9 10 86 X = Xi + (Σfd/N)i
presence of extreme score greatly affects the value of the Where: Xi = midpoint
mean
Σ = summation
mean = 15
Fd = frequency times deviation
Median = 7
I = class interval size
Formula for the mean:
f = frequency X = Xi + (Σfd/N)i
= 37 + (-74/50)5
Xi = midpoint
=37 + -7.4
N = number of cases
= 29.6
Try placing the 0 to another class interval step. Would you get the Where:
same value if the hypothetical mean is 42?
LB = lower boundary
scores f d fd
N/2 = ½ of N
55 - 59 1 3
50 - 54 1 2 <cfb = less than cumulative frequency before
= 24.5 + 3.75
= 28.25
Median – unlike the mean it does not incorporate all the values; it is
the value that divides the distribution in half, 50 % above it and 50%
below it.
D2 = difference between the f of the modal class and the next higher
class
e.g. 5, 7, 20, 24, 25 Median = 20 (mean = 16.2)
Mode = LB + (d1/d1+d2)i
= 24.5 + (6/6+5)5
= 24.5 + 2.727
= 27.23