Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 101
APPENDIX C
CODING RESPONDENT PROTOCOLS
One of the major tasks involved in the study was coding re-
spondent protocols. In constructing the categories for this operation,
a number of factors had to be considered: Some investigators in this
field (20) have pointed to the lack of theory from which to derive hy-
potheses for testing. It was important therefore, to construct cate-
gories which would yield data which could be studied in an exploratory
fashion, as well as permitting the testing of the hypotheses. (This
problem is treated more fully elsewhere. (2, 21) At the same time,
a certain amount of descriptive data was necessary as a means of
orienting the readers and investigators to the actual sequence of
events. Finally, the categories had to be defined well enough to per-
mit coders to categorize respondent statements reliably. This last
s ometime s means that we must sacrifice p sychologically me aningful
distinctions (concerning feelings, emotions, etc. ~ for more reliable
categories dealing with easily observable characteristics.
The procedure used in constructing categoric s is explained
in the following paragraphs.
Two investigators who were completely familiar with the
kinds of information needed for fulfillment of the research constructed
the categories. The general areas to be coverer] were listed and a
definition of each written. The definitions were based in part on first-
har~d experience which one of the investigators had in Port Jervis, and
in part on a thorough examination of the ''key" interview protocols.
To avoid the introduction of bias into the final analysis, the respondent
protocols were not used in this step.
After the categories had beer defined, and various codes
developed under each category, a small sample, 10, of the respondent
protocols was taken to check the adequacy of the coding proceclure.
Each inve stigator independently coded five protocols and then compared
his listings with the other. Modifications were made and the next five
were coded. This procedure was continued using the same ten proto-
cols and a few of those from key interviews until agreement between
the investigators was high, and it was judged that the information re-
quired would be obtained with the categoric s .
101
OCR for page 102
At this point, two completely naive coders were trained.
They were given a general summary of the background of the study,
but were not exposed to the hypotheses, proposals, or any discussions
of the specific information needed. As is indicated on the sample
page in this Appendix entitled "Instructions for Coders, " each coder
was directed to familiarize himself completely with the set of codes.
Following this, each coded one of the ten protocols used in the pre-
ceding steps. The results were reviewed by one of the code con-
structors, and clarifications of various points were made. Next,
they independently coded four more of the s e protocols . They r econ-
ciled differences and then referred their work to the same investigator.
After coding the remaining five protocols, it became apparent that
further training was unnecessary, and the coding procedure began.
These ten protocols could not be thrown out of the sample, although
the special treatment given to them may have introduced some minimal
bias into the results.
The coders independently coded sets of five protocols. They
then met and reconciled all differences, at the same time keeping a
record of such differences so that the degree of initial agreement
could be determined. This procedure was continued until about half
of the protocols were coded. At this point it became evident that it
was uneconomical to complete such small sets of protocols before
each consultation; the size of the sets was then increased to ten and
later to fifteen.
The per cent agreement on each category is given in the
accompanying table. The percentages range from 81. 1 per cent to
100 per cent. Those with the highest reliability primarily concern
biographical data. No category can be considered unreliable.
It is important to note, however, that the reliability estimates
given in this table are based~upon the independent judgments of the two
coders. These figures are overestimates of the reliability of the
categories as they are written. The frequent meetings of the two
coders resulted in their reaching mutual understanding as to the mean-
ings of categories and statements in the protocols.
On the other hand, the figures listed are underestimates of
the reliability of the data analyzed. These data were based on the
pooled judgments of the coders, which are obviously more reliable
than the independent j udgment of either co de r .
102
OCR for page 103
TABLE A
PERCENT INITIAL AGREEMENT OF CODERS
A N = 146 - Total sample
B N = 111 - Total who heard false report
C N = 333 - Total for three messages*
PERCENT
RANK AGREEMENT
CATEGORY TITLE
NUMBER IN SAMPLE
.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
99.3
99.3
99.3
99.3
99.3
99.1
98.6
98.6
98.6
98.
98.2
97.9
97.9
97.3
97.3
96.6
Sex
Code Number
Code Number
Code Number
Socio-Economic Status
Automobile
Numbe r in Hous ehold
Heard False Report
T elephone
Number of Children
Length of Residence in the City
Flood Evacuation
Flood Victims Staying in Home
Attitude T o war ~ P ok c e D ep t.
Age
Disaster-Related Job
Fled as a Function of Rumor
Location When Heard False Report
Attitude T award Red Cro s s
Schooli ng
Previous Disaster Experience
Expo s ed to Previous Rumor s
Attitude Toward Fire Dept.
When Flood Victims Stayed in Home
A
A
A
A
A
A
A
A
A
A
A
A
A
B
A
A
A
B
B
A
A
B
B
A
Reliabilities for the three threat and denial messages were computed
together. Therefore the N = three times the number exposed to the
reports .
103
OCR for page 104
PERCEN T
RANK AGREE MENT
CATEGORY TITLE
NUMBER IN SAMPLE
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
96. 4
95.5
94.6
94. 6
94. 0
93.8
93. 7
93. 7
93. 4
92. 8
91. 9
91.9
91.0
91. 0
91. 0
91. 0
90.1
90.1
89. 8
89.2
89.2
89.2
89.2
89.2
88.3
88.3
88.3
88.3
87. 4
87. 4
86. 5
85. 6
85. 6
84.7
81.1
Mean Reliabilitv = 93. ~ nor ren1:
104
Group with Re spondent - Strange r s
Denial Communi c ation
Activity when Heard False Report
Total Denial Sources
Threat Content
Damage by Flood
Group with Respondent - Farnily
Attitude T o war d O the r Or our s
Source anti
'_ ., ~
1
Channel of Threat
Group worn Re spondent - Neighbo r s
Time Re spondent Heard Fal se Repo rt
Communication with Family
Estimated Time Until Inundation
C onfir mation Channel
Re actions of Othe r s
Attitude Toward Civil Defense
Speed of Flight
Attitude Toward Unspecified Groups
~_ ~. ~. ~
-O
~ource en cl ~nanne1 ot tne L'enla
Belief in the False Report
Stated Personal Danger
Attempts to As sist Community
No. of Denial Sources to Belief
Total Denials Heard
Location of Family During Threat
Estimated Danger to Property
Things Taken in Flight
Denial Action
Confirmation Source
Content of the Denial s
Advice to Others
F amity Re sponsibilitie s
Nature of Flight
Number of Confirmation Attempts
Denial Source Leading to Belief
B
B
B
B
C
A
B
C
B
B
B
D
B
B
B
B
C
B
B
B
B
B
B
B
B
B
B
C
B
B
B
B
B
Median Reliability = 93. 9 per cent
OCR for page 105
In addition to the sample codes in the Appendix, the reader
will find a copy of the coding sheets used to record the coders' judg-
ments. The use of a separate code street has two primaryadvantages:
(1) It permits the coding operations of each coder to take place inde-
pendently. The practice of recording codes on the protocols obviously
has an influence on the judgments of subsequent coders. (2) This
format speeds up the punching operation when the codes are to be re-
corded on IBM or similar cards. The punch operator can quickly
and accurately punch and verify codes recorded in thi s fashion.
ANALY ZING THE KEY INTERVIE WS
Because of the varied nature of the key interviews, and be-
cause a different type of information was desired, it was impos sible
to use the same set clef codes for their analysis as for the respondent
protocols. Basically, we wanted an accurate description of the actions
of officials and semi-officials during the period of the false report,
rather than the testing of hypotheses. Data of this sort do notlend
themselves well to the latter type of treatment.
One of the category constructors and one of the coders, who
was by now sophisticated with respect to the study ant} its objectives,
worked cooperatively in pulling the information from the key protocols.
First, lists of actions taken in various centers were made by the coder.
These were checked for accuracy by the category constructor. Then
communications charts were constructed and these were also checked.
In some instances, data were missing from the protocols, while in
others factual disagreements between respondents in the same center
were noted. In such cases, both investigators agreed to avoid infer-
ences and to report only those data which could be supported by state-
~ents in the protocols.
Finally, descriptions of the action were written along with
the conclusions drawn from them, and checked for accuracy by the
second investigator as well as by a third investigator who had not
taken part in the phases described above.
105
OCR for page 106
SAMPLE PAGES FROM
INSTRUCTIONS FOR CODERS
Introduction
In the pages which follow, you will find a complete set of in-
structions for coding protocols from Operation DAM. It is extremely
important that you follow these instructions very carefully. If you have
any questions concerning the nature of the various categories, please
consult with either Dr. Danzig or Dr. Thayer.
In order to complete your assignment you will need the follow-
ing equipment: (1) this set of instructions for coding, (2) the set of
interview protocol s as signed to you, (3) a copy of the interview s chedule,
and (4) a set of coding sheets. After studying the instructions for coding
carefully and settling the answers to any questions, glance over one or
two of the protocols so that you may become familiar with the material
with which you will work. Next, select your first protocol for coding
and follow the instructions below. Please do not mark the protocols
themselves in any way. The only place you are to make any notations
is on the coding sheets themselves.
The adequacy of this study as a piece of research will, to a
large extent, be determined by the coding job which you are doing. All
hypotheses will be tested with the coded information that you will supply.
The conclusions reached and recommendations made will also be based
primarily on your work. It is evident, therefore, that your job is a
most important one and must be performed with care.
Coding Categories
-
You are to code the face-sheet data first. After this you will
code the information contained in the rest of the protocols.
You will note in the category definitions given below that there
is a uniformity throughout with respect to coding the response ''none"
and coding where no information is given. In any column where the re-
sponse is "none" or its equivalent (the category is not applicable), you
will always code zero (0~. In any column where there is no response;
i. e., no information is given, you will code "X. " In other words, "X''
is equivalent to missing data; i. e., it looks like the interviewer should
have included this information but failed to Lo so. Please make sure
that you follow this procedure in all cases as it will simplify the com-
puting and interpreting job.
(Sample codes are reproduced on the following two pages.
106
OCR for page 107
COLUMN
NUMBER
CATEGORY DEFINITION
SEX: On the coding sheet in the space opposite column 1
indicate the sex of the respondent as follows:
1. Male
2. Female
RESPONDENT CODE NUMBER: Opposite column 2 write the
first digit of the three-digit code number given on the face
she et.
RESPONDENT CODE NUMBER:
Write the second digit.
4. RESPONDENT CODE NUMBER: Write late third digit.
5.
6. SES:
28.
AGE: Code the respondent age as follows:
2.
3.
4.
1. 20
~1
26
31
or under
to 25
to 30
to 35
5. 36 to 40
6. 41 to 50
7. 51 to 60
8. 61 anti over
1. A
2. B
3. C
4. D
SOURCE AND CHANNEL OF THREAT MESSAGE -- FIRST
MESSAGE: For columns 28 through 30 the same coding will
be used. You will note that two major distinctions are being
made with respect to the source and channel of the threat
and denial messages. First, we are distinguishing between
official and unofficial sour ce s . Se coed, we are di stingui sh-
ing among various types of official sources: impersonal,
central, and personal official. The code "Impersonal Offi-
cial" refers to such sources as the fire-truck loudspeaker,
the radio, or television. These sources are obviously
107
.
OCR for page 108
official, yet the respondents cannot interact with such sour-
ces. They must accept or reject the information supplied
and cannot ask for clarification or additional information.
The code "Central Official" refers to a source at a commu-
nications center where the respondent has a chance to inter-
act with such a source. Such interaction may take place
either in a face-to-face conversation or through the tele-
phone. There are two basic characteristics of such a source;
(1) He is at a communications center and as a function of
this may be perceived by the respondent as having up-
to-ciate information.
(2) The re spondent has the opportunity to interact with him
and ask for reassurance, further clarification, or addi-
tional information.
The code "Personal Official" refers to a source that is offi-
cial but is not at a communications center.
NOTE: In counting threat messages, do not count any after
the respondent heard the denial message and believed the
denial.
All columns 28 through 30 would therefore be coded as follows:
1. Impersonal official -- fire-truck, loudspeaker
2. Impersonal official -- radio or television
3. Central official -- word of mouth communication fro
civil defense, police, fire, etc., at a communica
tions center .
4. Central official -- telephone source from civil defense,
police, fire, etc., at a communications center.
5. Personal Official -- policeman, fireman, civil defense,
etc., on the street and removed from a communica
tion s c ente r .
6 . Unofficial - - source is a relative .
7. Unofficial -- source is a neighbor or friend.
g . Unofficial - - source is a pas s er -by unknown to re spondent.
108
OCR for page 109
OPERATION DAM - CODE SHEET
Column Column
Number Code Category Number Code Category
9.
10.
11.
12.
13.
14.
Sex
Code Number
Code Number
Code Number
Age
6. SES
7. Car
Telephone
Disaster Job
No. in Household
No. of Children
Length Residence
S chooling
Flood Evacuation
1 5. Damage by Flood
16. Victims With
17. Victims When
lS. Previous Experience
19. Fled_ Rumor
20. He ard Report
21 . L`o cation
22. Activity
3 . Re sponsibilitie s
24. Group With - Family
25. Group With - Neighbors
26. Group With - Strangers
27. Where Family
28. Threat S&C - 1st
.
29. Threat S&C - 2nd
30. Threat S&C - 3rd
.
31. Threat Content - 1st
32 . Threat Content - 2nd
33. Threat Content - 3rd
34. Time Heard Report
35. Belief in Report
36. Est. Danger - Personal
-
37. Est. Danger - Property
38. Time Until Flood
39. Nature of Flight
40. Confirmation Source
Institute for Research in Human Relations
109
41. Confirmation Channel
42. No. Confirmation Attempts
43. As sist Community
44. Communicate with Family
45. Things Taken With
46. Speed of Flight
47. Reactions of Others
48. Denial S&C - 1st
49. Denial S&C - 2nd
50. Denial S&C - 3rd
51. Denial Source-Belief
52. No. Denial Sources-Belief
-
53. TotalDenials
54. Total Denial Sources
55. Denial Content - 1st
56. 1st (continued)
57. Denial Content - 2nd
58. 2nd (continued)
59. Denial Content - 3rd
60. 3rd (continued)
61. Denial Action
62. Denial Communication
63. Previous Rumors
.
64. Attitude re CD
6 5. Attitude re Police
66. Attitude re Fire
6 7. Attitude re Red Cro s s
68. Attitude re Other Group s
-
69. Attitude re Unspecified
7 0. Advice to Othe r s
7 1 . Sirens and Noi se
72.
73.
74.
75.
76.
77.
78.
79.
80.
OCR for page 110
Representative terms from entire chapter:
respondent protocols
