Data presentation

1

• Data presentation

A guide to good tables
Marianne W. Zawitz zawitzm_at_ojp.usdoj.gov
Bureau of Justice Statistics October 11, 2000
2
Elements of a table
Table 1. Number of widgets and gadgets by color



Number
Widgets
Color
Gadgets
Total
Blue
20
14
6
23
Red
28
5
3
Green
18
15
True color is aqua not green.

• Bureau of Justice Statistics

3
Elements of a table
Table 1. Number of widgets and gadgets by color



Number

Widgets
Color
Gadgets
Total
Blue
20
14
6
Stub
23
Red
28
5
3
Green
18
15
True color is aqua not green.

• Bureau of Justice Statistics

4
Elements of a table
Table 1. Number of widgets and gadgets by color
Header



Number
Widgets
Color
Gadgets
Total
Blue
20
14
6
23
Red
28
5
3
Green
18
15
True color is aqua not green.

• Bureau of Justice Statistics

5
Elements of a table
Table 1. Number of widgets and gadgets by color
Spreader Head



Number
Widgets
Color
Gadgets
Total
Blue
20
14
6
23
Red
28
5
3
Green
18
15
True color is aqua not green.

• Bureau of Justice Statistics

6
Elements of a table
Table 1. Number of widgets and gadgets by color



Number
Data field
Widgets
Color
Gadgets
Total
Blue
20
14
6
23
Red
28
5
3
Green
18
15
True color is aqua not green.

• Bureau of Justice Statistics

7
Elements of a table
Table 1. Number of widgets and gadgets by color



Number
Cell
Widgets
Color
Gadgets
Total
Blue
20
14
6
23
Red
28
5
3
Green
18
15
True color is aqua not green.

• Bureau of Justice Statistics

8
Elements of a table
Table 1. Number of widgets and gadgets by color



Number
Row
Widgets
Color
Gadgets
Total
Blue
20
14
6
23
Red
28
5
3
Green
18
15
True color is aqua not green.

• Bureau of Justice Statistics

9
Elements of a table
Table 1. Number of widgets and gadgets by color



Number
Column
Widgets
Color
Gadgets
Total
Blue
20
14
6
23
Red
28
5
3
Green
18
15
True color is aqua not green.

• Bureau of Justice Statistics

10
Elements of a table
Table 1. Number of widgets and gadgets by color
Title



Number
Widgets
Color
Gadgets
Total
Blue
20
14
6
23
Red
28
5
3
Green
18
15
True color is aqua not green.

• Bureau of Justice Statistics

11
Elements of a table
Table 1. Number of widgets and gadgets by color



Number
Widgets
Color
Gadgets
Total
Blue
20
14
6
23
Red
28
5
3
Green
18
15
Notes
True color is aqua not green.

• Bureau of Justice Statistics

12

• Use tables, when you need -

• to present 10 or fewer data points
• to find exact numeric values
• to make localized comparisons

Bureau of Justice Statistics
13

• Tables have an architecture

• Use the structure of the table to reinforce its
  purpose
• Take advantage of how we learned to do math

Bureau of Justice Statistics
14

327
483
123
42
7
18








327





483





123





42





7





18


Bureau of Justice Statistics
15

XXXX
XXX
XX
X
Bureau of Justice Statistics
16

• Types of tables -

• Exploratory tables
• Resource tables
• Presentation tables

Bureau of Justice Statistics
17
Example of an exploratory table

12 and
13-14
15
16
17
18
19
20
21

under








1979
6.7
122
236
326.4
404
472.1
425
400.9
403.9
1980
6.7
126.3
251
319.4
400.8
457.5
429.9
399.6
407.4
1981
7.2
138.4
262.5
339.7
406
460.9
450.5
407
410.9
1982
6.8
128.5
266.6
336.9
420.9
471.6
465.8
443.5
432.1
1983
7.5
130.8
245.5
316.3
393.4
433.2
413.4
393.6
399.8
1984
7.4
141.2
267.1
336.5
414.8
468.7
433.6
392.4
410
1985
8.2
165.6
302.9
374.6
444.3
501.2
469.5
418.9
415
1986
7.2
166.9
297.3
388.7
476.4
560.5
508.9
484.3
468.6
1987
7.2
163.5
295.8
368.8
458.4
548.3
516.1
469.8
474
1988
9.3
180.1
343.9
435.5
541.2
634.8
558.5
509.9
497.1
1989
9.5
206.5
391.8
531.7
640.2
704.8
638
541.9
543.2
1990
8.9
221.2
416.7
553.9
661.8
756.3
632.3
546.5
521.4
1991
12.7
280.6
518.4
669.7
775.1
886.6
725.2
601.9
591.4
1992
13.9
316.5
559
724.1
802.9
920.1
777.6
651.5
610.8
Bureau of Justice Statistics
18
Resource table

Number of
Full-time
Full-time
State
agencies
employees
sworn officers




All States
18,769
921,978
663,535




Alabama
432
14,389
9,767
Alaska
69
1,884
1,254
Arizona
130
16,828
10,088
Arkansas
360
7,958
5,819
California
524
103,967
69,134




Colorado
247
14,002
9,896
Connecticut
129
10,319
8,525
Delaware
45
2,134
1,660
District of Columbia
3
4,651
3,909
Florida
385
60,808
37,395




Bureau of Justice Statistics
19
Presentation table
Homicides of children under age 5

• Relationshipwith child

Percent of perpetrators


Father
32
Acquaintance
30
Mother
28
Other family
6
Stranger
4
Bureau of Justice Statistics
20

• Tables should --

• Show the data
• Reduce nondata ink
• Avoid puzzles
• Create a logical visual pattern
• Use the same standards for verbal elements we
  use in graphics

Bureau of Justice Statistics
21
Excess lines and marks make the data hard to see
State
Number of agencies
Full-time employees
Full-time sworn officers

All States ...................
18,769
921,978
663,535




Alabama ....................
432
14,389
9,767
Alaska ........................
69
1,884
1,254
Arizona ......................
130
16,828
10,088
Arkansas ...................
360
7,958
5,819
California ...................
524
103,967
69,134




Colorado ....................
247
14,002
9,896
Connecticut ...............
129
10,319
8,525
Delaware ....................
45
2,134
1,660
District of Columbia .
3
4,651
3,909
Florida .......................
385
60,808
37,395




Bureau of Justice Statistics
22
Without the lines the data are the focus of the
table
State
Number of agencies
Full-time employees
Full-time sworn officers

18,769
921,978
663,535
All States




432
14,389
9,767
Alabama
69
1,884
1,254
Alaska
130
16,828
10,088
Arizona.
360
7,958
5,819
Arkansas
524
103,967
69,134
California




247
14,002
9,896
Colorado
129
10,319
8,525
Connecticut
45
2,134
1,660
Delaware
3
4,651
3,909
District of Columbia
385
60,808
37,395
Florida




Bureau of Justice Statistics
23

• Keep decimal places to a minimum

• They imply false precision(often they are not
  significant)
• They pad the numbers making them bigger than they
  really are
• Use only as many decimal places as needed to
  break a tie

Bureau of Justice Statistics
24
With fewer decimal places, it is easier to read
This table has too many decimal places
Homicides of children under age 5
Homicides of children under age 5
Male perpetrators Male perpetrators Male
victim 67.8697 Male victim 68
Female victim 21.9989 Female
victim 22 Female perpetrators Female
perpetrators Male victim
7.7851 Male victim 8 Female
victim 2.3463 Female victim 2
The numbers were different enough that decimal
places were not needed
Bureau of Justice Statistics
25

• Presentation tables should --

• Reduce the data presented
• Not present everything to show that it all adds
  up
• Use only one of the two variables when
  presenting binary data

Bureau of Justice Statistics
26

• Order the content according to how the data will
  be used

• Presentation tables use rank order or another
  hierarchy -- Look to the numbers to carry
  out the ordering
• Resource tables use alphabetical or reference
  order

Bureau of Justice Statistics
27
Rank order in presentation tables reveals the
structure of the data
Homicides of children under age 5

• Relationshipwith child

Percent of perpetrators


Father
32
Acquaintance
30
Mother
28
Other family
6
Stranger
4
Bureau of Justice Statistics
28
Order resource tables alphabetically or in
another meaningful manner
State
Number of agencies
Full-time employees
Full-time sworn officers

18,769
921,978
663,535
All States




432
14,389
9,767
Alabama
69
1,884
1,254
Alaska
130
16,828
10,088
Arizona.
360
7,958
5,819
Arkansas
524
103,967
69,134
California




247
14,002
9,896
Colorado
129
10,319
8,525
Connecticut
45
2,134
1,660
Delaware
3
4,651
3,909
District of Columbia
385
60,808
37,395
Florida




Bureau of Justice Statistics
29

• Structure the table to --

• make patterns and exceptions obvious at a
  glance
• place numbers that are to be compared close to
  one another (in a column is best)
• place summary data where it will be most useful

Bureau of Justice Statistics
30
This structure makes it difficult to compare the
same units
1990 1995 1996 ITEM Number
Percent Number Percent Number
Percent TYPES OF SUBSTANTIATED
MALTREATMENT/1 Victims,
total 690,658 (X) 970,285 (X) 969,018
(X) Neglect 338,770 49.1 507,015 52.3 500,032 5
1.6 Physical abuse 186,801 27 237,840 24.5 229
,332 23.7 Sexual abuse 119,506 17.3 122,964 12.7
119,397 12.3 Emotional maltreatment
45,621 6.6 42,051 4.3 55,473
5.7 Medical Neglect (NA) (NA) 28,541
2.9 5,758 2.7 Other and unknown
61,477 8.9 157,819 16.3 175,613 18.1
Bureau of Justice Statistics
31
A revision puts the numbers that are to be
compared close together
TYPES OF SUBSTANTIATED MALTREATMENT
/1 Number of victims Percent
of all victims 1990 1995 1996
1990 1995 1996 ------------------------------
-------- -------------------------------------- Vi
ctims, total 690,658 970,285 969,018
100.0 100.0 100.0 Neglect 338,770 507,015 500
,032 49.1 52.3 51.6 Physical
abuse 186,801 237,840 229,332 27.0 24.5
23.7 Sexual abuse 119,506 122,964 119,397
17.3 12.7 12.3 Emotional maltreatment
45,621 42,051 55,473 6.6 4.3
5.7 Medical Neglect ... 28,541 25,758
2.9 2.7 Other and unknown
61,477 157,819 175,613 8.9 16.3 18.1
Bureau of Justice Statistics
32
A better way to restructure this table is to make
the annual comparisons up and down
Emotional Victims, Physical Sexual
maltreat- Medical Other and Year
total Neglect abuse abuse ment Neglect unknown
Number 1990 690,658 338,770 186,801
119,506 45,621 ... 61,477 1995 970,285 507,
015 237,840 122,964 42,051 28,541 157,819 1996 96
9,018 500,032 229,332 119,397 55,473 25,758 175,61
3 Percent 1990 100.0 49.1 27.0
17.3 6.6 ... 8.9 1995 100.0 52.3 24.5 12.
7 4.3 2.9 16.3 1996 100.0 51.6 23.7 12.3 5.7 2.7
18.1
Bureau of Justice Statistics
33

• Notation should --

• be placed in the row labels and column headers
  if possible (otherwise they interfere with our
  ability to see the data)
• not be numeric characters so they are not
  confused with the numbers

Bureau of Justice Statistics
34
Notation in this table interferes with your
ability to compare the numbers
1990 1995 1996 ITEM Number
Percent Number Percent Number
Percent TYPES OF SUBSTANTIATED
MALTREATMENT/1 Victims,
total 690,658 (X) 970,285 (X) 969,018
(X) Neglect 338,770 49.1 507,015 52.3 500,032 5
1.6 Physical abuse 186,801 27 237,840 24.5 229
,332 23.7 Sexual abuse 119,506 17.3 122,964 12.7
119,397 12.3 Emotional maltreatment
45,621 6.6 42,051 4.3 55,473
5.7 Medical Neglect (NA) (NA) 28,541
2.9 5,758 2.7 Other and unknown
61,477 8.9 157,819 16.3 175,613 18.1
Bureau of Justice Statistics
35
Lighter notation conveys information without
interfering
TYPES OF SUBSTANTIATED MALTREATMENT
/1 Number of victims Percent
of all victims 1990 1995 1996
1990 1995 1996 ------------------------------
-------- -------------------------------------- Vi
ctims, total 690,658 970,285 969,018
100.0 100.0 100.0 Neglect 338,770 507,015 500
,032 49.1 52.3 51.6 Physical
abuse 186,801 237,840 229,332 27.0 24.5
23.7 Sexual abuse 119,506 122,964 119,397
17.3 12.7 12.3 Emotional maltreatment
45,621 42,051 55,473 6.6 4.3
5.7 Medical Neglect ... 28,541 25,758
... 2.9 2.7 Other and
unknown 61,477 157,819 175,613 8.9
16.3 18.1
Bureau of Justice Statistics
36
Avoid tables with many empty cells and dont use
notation for true 0s

Number
Number
Number
State
admitted
executed
removed




All States
--
--
--




Alabama
--
--
--
Alaska
--
--
--
Arizona
--
--
--
Arkansas
--
--
--
California
--
--
--




Colorado
--
--
--
Connecticut
--
--
--
Delaware
--
--
--
District of Columbia
--
--
--
Florida
--
--
--
-- see page 27.
Bureau of Justice Statistics
37

• Rows should --

• have consistent labeling from table to table
• indent the labels to show subdivisions
  (nesting)
• create patterns with a line or half line skip
  every fifth entry in long lists

Bureau of Justice Statistics
38
Label nesting shows relationships
Homicides of children under age 5
Male perpetrators Male victim 68
Female victim 22 Female
perpetrators Male victim 8
Female victim 2
39

• Columns should --

• contain the same unit of measure
• have labels that clearly identify what is
  measured (both numerator and denominator for
  rates or percentages)
• avoid repetitious categories by spacing and
  grouping

Bureau of Justice Statistics
40
Mixing units of analysis compares apples to
oranges (or dogs to cats)
Source Statistical Abstracts of the United States
Bureau of Justice Statistics
41
Avoid large gaps between the stub and first column

Full-time
Full-time
State
employees
sworn officers




All States
921,978
663,535




Alabama
14,389
9,767
Alaska
1,884
1,254
Arizona
16,828
10,088
Arkansas
7,958
5,819
California
103,967
69,134




Colorado
14,002
9,896
Connecticut
10,319
8,525
Delaware
2,134
1,660
District of Columbia
4,651
3,909
Florida
60,808
37,395



the grand canyon
Bureau of Justice Statistics
42
A few additional thoughts

• Use no more than 9 columns in a table
• Add clues regarding the unit of analysis next to
  the first number (, months, widgets, etc.)

Bureau of Justice Statistics
43
Make all verbal tasks easy to understand

• Do not use abbreviations
• Avoid acronyms
• Write all labels left to right
• Use proper grammar

Bureau of Justice Statistics
44

• Table titles should --

• briefly describe the data presented
• highlight the difference between this table and
  others in the first words
• be long enough to inform but short enough not to
  bore

Bureau of Justice Statistics