new {ESE cu

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ABSTRACT

THE EFFECT OF THE CONTENT OF
PREVIOUSLY ASKED QUESTIONS ON THE CONTENT OF
THE RESPONSE TO A STANDARD QUESTION

by John P. Calicchia

Although some of the variables that affect responses in
questionnaires have been from time to time theoretically defined and
discussed, relatively seldom has the effect of these same variables
been empirically investigated in an interview situation. The present
research, carried out in an industrial research organization, was de—
signed to explore the effect of one such variable, namely, the generality-
specificity of the content of questions asked.

It was the intention of this research to determine the extent
to which the content of a response was a function of the content of
previously asked questions, and the following hypothesis was made
concerning this relationship.

The content of previously asked questions would establish a mental
response set in the responding individual, and this mental re-
sponse set would influence the subject's response to a subsequent

question in such a way as to reflect the nature of the content of the
previously asked questions .

 

John P. Calicchia

Furthermore, it was of interest to test, in an interview
situation, Parten's assertion concerning the order in which general
and specific questions should be asked. She stated that if both general
and specific question were to be used in the same interview schedule,
the general ones should precede the specific ones, instead of the
opposite; otherwise, subjects tended to answer the general ones in
terms of the earlier specific ones.

To test the basic hypothesis and Parten's assertion, two
interview schedules were available for use. The two types of inter-
view schedules were semi-structured but differed in content. These
interviews consisted of a series of specific questions. While the
questions clearly indicated the idea or procedure to be discussed in
the answer, the subjects had complete freedom in the content and
manner of responding to each question.

One of the two interview schedules, the management
interview, contained six major questions intended to reflect the sub-
ject's knowledge of company's procedures. The questions in this
interview schedule were general in the sense that they asked for in-
formation about the management and accounting control procedures
which were known to at least some extent by everyone interviewed.
The second interveiw schedule, the project interview, contained ten
major questions intended to reflect the subjects's understanding of a

particular technical project or work area. The information asked

 

 

John P. Calicchia

for in this interview was specific to a relatively restricted part of the
company; such information would be known best by those working on
the particular project or work area. The generality-specificity of
the content in these questions made up the independent variables in
this study.

The last question in both interviews was standard. The
content of the response to this standard question constituted the de—
pendent variable. This question simply asked the subject to list the
changes perceived as instituted in the organization during the past
year.

Seven interviewers interviewed approximately half the
fifty-five engineers with only the management interview and approxi-
mately half the subjects with only the project interview.

A code book was developed to reflect the content of the
response to the standard question. Two raters independently judged
the generality—specificity of the categories in the code book. The
reliability of judgments between the two raters was found, and the
disagreements ironed out in discussion allowing for a common agree-
ment as to which categories in the code book were general and which
categories were specific.

The content of the responses to the standard question was
then analyzed, and frequency counts were obtained of the number of

general and specific units of information each response contained.

 

 

 

 

John P. Calicchia

for in this interview was specific to a relatively restricted part of the
company; such information would be known best by those working on
the particular project or work area. The generality-specificity of
the content in these questions made up the independent variables in
this study.

The last question in both interviews was standard. The
content of the response to this standard question constituted the de-
pendent variable. This question simply asked the subject to list the
changes perceived as instituted in the organization during the past
year.

Seven interviewers interviewed approximately half the
fifty-five engineers with only the management interview and approxi-
mately half the subjects with only the project interview.

A code book was developed to reflect the content of the
response to the standard question. Two raters independently judged
the generality-specificity of the categories in the code book. The
reliability of judgments between the two raters was found, and the
disagreements ironed out in discussion allowing for a common agree-
ment as to which categories in the code book were general and which
categories were specific.

The content of the responses to the standard question was
then analyzed, and frequency counts were obtained of the number of

general and specific units of information each response contained.

 

John P. Calicchia

for in this interview was specific to a relatively restricted part of the
company; such information would be known best by those working on
the particular project or work area. The generality-specificity of
the content in these questions made up the independent variables in
this study.

The last question in both interviews was standard. The
content of the response to this standard question constituted the de-
pendent variable. This question simply asked the subject to list the
changes perceived as instituted in the organization during the past
year.

Seven interviewers interviewed approximately half the
fifty-five engineers with only the management interview and approxi-
mately half the subjects with only the project interview.

A code book was developed to reflect the content of the
response to the standard question. Two raters independently judged
the generality-specificity of the categories in the code book. The
reliability of judgments between the two raters was found, and the
disagreements ironed out in discussion allowing for a common agree-
ment as to which categories in the code book were general and which
categories were specific.

The content of the responses to the standard question was
then analyzed, and frequency counts were obtained of the number of

general and specific units of information each response contained.

 

 

 

John P . Calicchia

for in this interview was specific to a relatively restricted part of the
company; such information would be known best by those working on
the particular project or work area. The generality-specificity of
the content in these questions made up the independent variables in
this study.

The last question in both interviews was standard. The
content of the response to this standard question constituted the de-
pendent variable. This question simply asked the subject to list the
changes perceived as instituted in the organization during the past
year.

Seven interviewers interviewed approximately half the
fifty-five engineers with only the management interview and approxi-
mately half the subjects with only the project interview.

A code book was developed to reflect the content of the
response to the standard question. Two raters independently judged
the generality-specificity of the categories in the code book. The
reliability of judgments between the two raters was found, and the
disagreements ironed out in discussion allowing for a common agree-
ment as to which categories in the code book were general and which
categories were specific.

The content of the responses to the standard question was
then analyzed, and frequency counts were obtained of the number of

general and specific units of information each response contained.

 

John P. Calicchia

for in this interview was specific to a relatively restricted part of the
company; such information would be known best by those working on
the particular project or work area. The generality-specificity of
the content in these questions made up the independent variables in
this study.

The last question in both interviews was standard. The
content of the response to this standard question constituted the de-
pendent variable. This question simply asked the subject to list the
changes perceived as instituted in the organization during the past
year.

Seven interviewers interviewed approximately half the
fifty—five engineers with only the management interview and approxi-
mately half the subjects with only the project interview.

A code book was developed to reflect the content of the
response to the standard question. Two raters independently judged
the generality-specificity of the categories in the code book. The
reliability of judgments between the two raters was found, and the
disagreements ironed out in discussion allowing for a common agree-
ment as to which categories in the code book were general and which
categories were specific.

The content of the responses to the standard question was
then analyzed, and frequency counts were obtained of the number of

general and specific units of information each response contained.

 

 

 

 

John P . Calicchia

for in this interview was specific to a relatively restricted part of the
company; such information would be known best by those working on
the particular project or work area. The generality-specificity of
the content in these questions made up the independent variables in
this study.

The last question in both interviews was standard. The
content of the response to this standard question constituted the de-
pendent variable. This question simply asked the subject to list the
changes perceived as instituted in the organization during the past
year.

Seven interviewers interviewed approximately half the
fifty-five engineers with only the management interview and approxi-
mately half the subjects with only the project interview.

A code book was developed to reflect the content of the
response to the standard question. Two raters independently judged
the generality-specificity of the categories in the code book. The
reliability of judgments between the two raters was found, and the
disagreements ironed out in discussion allowing for a common agree-
ment as to which categories in the code book were general and which
categories were specific.

The content of the responses to the standard question was
then analyzed, and frequency counts were obtained of the number of

general and specific units of information each response contained.

 

John P. Calicchia

It was hypothesized that the questions in the management
interview would create a general mental response set in the individual
and this general response set would influence the subject's response
to the standard question in such a way that its content also would be
general in nature, that is, the content of the response would not be
delineated and would refer to the whole organization. On the other
hand, the questions in the project interview would create a specific
mental response set, and would, therefore, influence the subject's
response to the standard question in such a way as to reflect this
specific mental response set, i.e. the content of the response would
be rather detailed and refer to a particular job and not to the entire
organization.

If the hypothesis were true, there should have been a
statistically significant greater proportion of general categories
counted in the responses to the standard question in the management
interview than counted for the project interviews, and conversely,
there should have been a statistically significantly greater proportion
of specific categories counted in the responses to the standard ques—
tion for the project interviews than counted for the management inter-
views.

The results of attempts to find inter-rater reliability
indicated that such reliability was relatively low. A series of "t” tests

were computed on the data obtained for each coder independently in

John P. Calicchia

order to test the hypothesis. The results of these "t" tests did not
support the hypothesis. These results, however, did refute Parten‘s
assertion, in an interview situation, that is, having specific questions
precede a general question did not necessarily induce the subjects to
answer the general question in terms of the earlier specific questions.
Although the findings concerning the basic hypothesis were
essentially negative, the possibility of a relationship between the con-
tent of the question and the content of the response was indicated by a.
trend in the data. The experimenter suggested that, perhaps, refine-
ments in the procedure used-—as mentioned in the discussion--may

help clarify the present uncertain relationship.

Approvefd—c/Jln—L. W. Dds-M

Date: 4" 15—" 6’6

 

 

THE EFFECT OF THE CONTENT OF

PREVIOUSLY ASKED QUESTIONS ON THE CONTENT OF

THE RESPONSE TO A STANDARD QUESTION

BY

John P. Calicchia

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Department of P5 ychology

1966

 

To my parents,

Louis and Lena Calicchia,
who have made this possible,
and to my brothers,
Joseph and Louis
to whom I wish much success

in their own endeavors.

 

ACKNOWLEDGMENTS

I would like to express deepest appreciation to my com—
mittee chairman, Dr. Frederic R. Wickert. His patience, under—
standing and stimulating ideas were of immeasurable importance in
the execution of this thesis.

Many thanks are also extended to Dr. James Phillips,
and Dr. D. M. Johnson for their helpful suggestions and criticisms.

Thanks are due to Wallace Berger and Richard Larsen,
graduate students, and my Aunt Margret Finelli for their aid in
coding and data analysis and to those friends, especially Morris

Spier, for their guidance and encouragement.

 

 

Y:‘mm---

 

 

 

TAB LE OF CONTENTS

ACKNOWLEDGMENTS .

LIST OF TABLES .

LIST OF APPENDICES .

INTRODUCTION

BACKGROUND OF THE PRESENT PROBLEM .

HYPOTHESIS
METHOD .
Subjects .
Materials
Procedure

RESULTS

Data Processing
Summary of Results

DISCUSSION .
SUMMARY AND CONCLUSIONS
REFERENCES . .

APPENDICES

iv

17
l7
17
18
23

23
36

39

53

57

59

 

LIST OF TAB LES

Inter-rater correlations . . . . . . . . . .

The percentages of agreement between coders on
each response for all the interviews . . . . . .

Mean difference, standard deviation of the difference,
n, degree of freedom and 't' between coders on cod-
ing the responses to the standard question in the
management interview, a) general, b) specific . .

Mean difference, standard deviation of the difference,
n, degree of freedom and 't' between coders on cod-
ing the responses to the standard question in the
project interviews, a)general, b) specific . . . .

Mean proportion, standard deviation of the proportion,
n, degree of freedom and 't' between the management
and project interviews for Coder l on the proportion
of general and specific categories found in the re-
sponses to the standard question . . . . . . .

Mean proportion, standard deviation of the proportion,
n, degree of freedom and 't' between the management
and project interviews for Coder 2 on the proportion
of general and specific categories found in the re-
sponses to the standard question . . . . . . .

Mean proportion, standard deviation of the proportion,
n, degree of freedom and 't' between the management
and project interviews for the master code on the pro-
portion of general and specific categories found in
coding the responses to the standard question . . .

Summarytable of results . . . . . . . . .

 

Page

26

27

30

30

32

32

34

38

 

 

 

 

LIST OF APPENDICES

Appendix Page
A. Management interview schedule . . . . . . . . 59
B. Project interview schedule . . . . . . . . . . 68
C. Code book for the standard question . . . . . . 73

V1.

Appendix
A .

B.

LIST OF APPENDICES

Management interview schedule
Project interview schedule .

Code book for the standard question

vi

Page
59
68

73

 

INTRODUCTION

This study dealt with mental response sets induced by a
verbal context in an interview situation.

In reviewing the psychological literature, it was noticed
that the hypothetical constructs of response style and response set
were frequently interchanged and misused. It was unfortunate that
this state of affairs existed because these two concepts are not syn-
onymous and do not refer to the same phenomena. (Rorer, 1965)

Response style, or response bias, according to Rorer,
usually referred to a tendency on the part of the individual to select
some response category or activity a disproportionate number of
times independently of the situational context, that is, neither the
nature, structure, nor content of the stimulus situation influenced
the response in any way. Phenomena such as "yea-saying" and con-
stantly choosing false in a true and false examination could be cited
as an example of the type of behavior usually referred to by the term
response style.

Response set, on the other hand, according to Rorer,
referred to the phenomenon in which a previous rather specific activity
determined or influenced the present behavior, that is, where the

nature, structure, or content of a previous or present stimulus situation

 

 

 

 

 

influenced the subjects' response in such a way that their responses

reflected the previous or present stimulus situation. In this type of
behavioral phenomenon the situational context was extremely important.
The response depended a great deal on the context of the situation.
Different contexts provoked different kinds of responses.

The following discus sion focuses on some of the
variables that could be involved in creating mental response sets.

The literature revealed many studies which have investi-
gated the effect of mental sets on the content of the responses as in-
duced by questions in questionnaires. The exact nature of the process
by which such mental sets were induced was far from completely estab-
lished, but the results of some of these studies have identified a few of
the variables that were involved in creating these mental response sets.

The grammatical structure of questions was shown by Fries
( 1952) to be extremely important in influencing responses. He reported
that questions with the highest degree of structure, that is, the highest
degree of syntax, had the greatest influence on the content of the re-
sponses to these questions. Thus, the more implicit the sentence
structure was in indicating the exact information desired, the greater
was the extent to which the content of the response was congruent with
the information called for.

Cronbach ( 1946) has shown that response sets have the greatest
influence when the questions were ambiguous or unstructured. He re—

ported that if the question was structured so that one knew the answer

 

required, the response was made directly to the content of the question,

and response sets were not created. But, on the other hand, if the
content of the question did not indicate the topic to be discussed, the
content of the responses given to this question was a function of a re-
sponse set created by situational (preceding questions) and/or person-
ality (motivation) variables.

While there seemed to be quite a bit of data on how mental
response sets induced by the nature, structure, or content of questions
affected the responses to questions on questionnaires, the literature
revealed few, if any, studies investigating this same phenomenon in
the interview situation; rather most of the writers simple assumed
that these variables would have the same effect.

Selltiz, Johoda, Deutsch, and Cook (1964) did some hypothe-
sizing in this area, and stated that question content was very important.
They believed that the content of the response was likely to be influenced
by the content of the preceding questions. For example, if previously
asked questions created a favorable or unfavorable mental set toward
the topic considered, the response to the subsequent question would
reflect this attitude.

Parten ( 1950), another writer who was interested in both
questionnaires and interviews, stated that the question, or questions,
preceding a given question produced a "set" in the respondent and
caused him to reply very differently from the way he might have if the

preceding question or questions were different.

 

She also indicated that the order of the question was im-

portant because the content of the response could be influenced by the
question arrangement. She stated that if both general and specific
questions were to be used, the general ones should precede the specific;
otherwise, subjects were likely to answer the general ones in terms of
the earlier specific ones.

Parten cited as empirical evidence for this statement the
study performed by Link ( 1947) , one of the few people found who had
done research in this area. He investigated the effect of question ar-
rangement on the responses to two related questions. He reported that
the order of asking the questions is instrumental in determining the
response to the second of the two questions asked. For example, he
found that when the first question asked was ”Do you believe that workers
and unions have the right to strike when wages and working conditions
don't suit them ?”, and the second question asked was "Do you believe
that businessmen have the right to shut down their factories and stores
when labor conditions and profits don't suit them?“, 65 percent of the
5, 000 subjects said that the workers had the right to strike, but when
the order was reversed only 61 percent of the subjects held this same
view. The percentage who replied that business had the right to close
down their shops was also greater when question one preceded question
two. When question one came first, 52. 3 percent approved of the right

to close the establishment, but when question two was presented first,

 

46 percent believed in the employer's rights. Considering the fact that

the N was 5, 000, such numerically small differences in percentages
were statistically significant.

Comments have been made on the nature of the process
by which mental response sets could be established in terms of the
favorableness or atmosphere of the content of the questions (Selltiz
$11. , 1964) , the grammatical structure of the question (Fries, 1952) ,
and the order in which the questions were asked (Link, 1947; Parten,
1950) . No studies we r e found that investigated how the generality-
specificity of the content of the questions asked created a. mental re-
sponse set and thereby influenced the content of the response.

The purpose of the present study was to experimentally
define more clearly one of the variables that may affect the content of
the responses in an interview situation, namely the generality-specificity
of the content of the questions. More specifically, the present research
tried to determine, in an interview situation, to what extent the generality-
specificity of the content of a previously asked series of questions created
a mental response set that influenced the content of a response to a stan-
dard question that followed the series. The results of this investigation
would permit the experimenter to empirically deny or support Parten's
assertion in interview situations concerning the order in which general

and specific questions should be asked.

 

BACKGROUND OF THE PRESENT PROBLEM

In experimental studies which use an interviewing method
calling for complex verbal responses, a common methodological prob—
lem, often encountered but seldom either studied or controlled, is the
effect of the content of a preceding series of questions on the content
of the responses given to a standard question that follows the series.

An opportunity to study such an effect arose in connection
with a field study (Wickert, 1964-66) of changes in an industrial
research organization. The subjects in this field study were a group
of fifty-five engineers selected randomly from a larger group of
engineers . The subjects generally filled both a management and pro-
ject role in this research organization, that is, they were at the same
time line research superivsors and members of one or possibly two
project teams on which they may or may not have been project leader.
but were always at least contributing engineers.

This field study was conducted in two phases. The first
phase took place during a two —month period in 1964, and the second
phase, during the same two-month period one year later, in 1965.
During the time interval between the two phases the administration

was supposed to institute organizational changes in an attempt to

 

 

 

improve communication within the engineering organization. The

major change that was supposed to be induced was a series of improve-
ments in the managerial control and accounting procedures in use,
accompanied by a program of educating the engineers to make all
accounting and managerial procedures, both the old and the improved
ones, more meaningful and useful to them. Up to this point there was
some evidence that the engineers tended not to take the management
control procedures too seriously: they often gave the impression that
they thought that these procedures hindered their work, were unim-
portant, and not part of their job. They seemed to feel that their job
was to create and not to fill out papers. The engineers tended not to
realize or to want to understand the function of accounting and mana-
gerial control procedures. In response to this situation the adminis—
tration attempted to improve the managerial and accounting control
procedures and to teach the engineers the roles these control pro-
cedures played in the company.

One of the purposes of Wickert's study that directly con-
cerns the present study was to determine indirectly, through measur-
ing communications in the organization, to what degree, if any, the
administration had actually succeeded in improving the control pro-
cedures and in educating the engineers.

Then, in addition, by asking the subjects in a systematic
way in 1965 what changes they had perceived during the last year,

Wickert hoped to get supplementary information on the degree of

c

 

 

 

 

 

success the administration had in accomplishing their aim. This type

of questioning, that is, asking the subjects what they perceived as the
independent variable (changes induced by the administration), is seldom
if ever done. When it is done, according to Orne (1962) it is usually
done in a haphazard, unsystematic, and after-the-fact way. Usually,
when an experiment gave unexpected results, the experimenter then
attempted to discover what went wrong by asking the subjects what they
thought the experiment was all about. The problem, in this field study,
was in "manipulating" the experimental condition that defined the indepen-
dent variables; there were many cross-currents--often the case in
field studies--so that the experimenter was not sure what changes were
induced between phase one and phase two. He knew only what the ad—
ministration had told him. One example of such cross-currents was
that the accounting procedures were instituted originally under an
accounting-minded president, but during the time interval between
phase one and phase two the accounting minded president left the
company and was replaced by an engineering-minded president.
Furthermore, during the time interval between phase one and phase
two a massive layoff was instituted which was accompanied by a feel-
ing of uneasiness. The experimenter had no idea of the effect of such
changes on the independent variable. Since the subjects were know-
ledgeable and functionally involved in the company, the experimenter

assumed that these men would have been generally aware of changes

 

in the management control and accounting procedures and of any pro-

gram for educating the engineers in the use of these control procedures.
50, in order to determine what the subjects themselves perceived as
the independent variable, the experimenter asked these engineers in a
systematic way, as part of the study, what they perceived as the
changes in the company during the past year.

Still another aspect of the Wickert study needs to be
explained in order to make clear why the present study was done.
This field study found in a preliminary investigation that there were
two main types of communication in the company. One, which will be
called management communication, flowed in a vertical direction,
that is, up and down the hierarchical structure of the organization
and reflected the management control communications already referred
to as a concern of the company's management. In this type of communi-
cation, the engineers were supposed to tell management what they were
doing and how they were spending their time, while management was
to indicate what was expected of them. The other type of communica-
tion flowed in a more or less horizontal direction, that is, among
peers on a particular engineering project, and dealt with the technical
aspects of that project.

Since there were the two different types of communication,
tw0 different interview schedules had to be constructed. The two types

of interviews were semi-structured but differed in content. These

 

10

interviews consisted of a series of specific questions given in a stan-
dard, logical order; nevertheless, while the questions clearly indi-
cated the idea or procedure to be discussed in the answer, the subjects
were given quite complete freedom in the content and manner of their
response to each question. In other words, the interviewees were
free to respond to the questions in any way they saw fit. The content
of the questions in these interviews sampled the content of the two
types of communication, that is, the management interview, with its
nineteen interrelated questions, intended to reflect the subject's under—
standing of the company's management control procedures such as
forms used, managerial practices, etc. This interview asked for
general information known and used to at least some extent by all the
members of the organization. The project interview also included
nineteen interrelated questions intended to reflect the subject's under-
standing of a particular technical project or work area. The informa-
tion asked for in this interview was specific to a relatively restricted
part of the company, for example, technical breakthroughs on a given
project, or details of a newly developed process related to the project;
such information would be known best by those working on the particular
project or work area.

Since, as was mentioned earlier, many of the subjects had
both a "management" and ”project" role, it was assumed that they

would be equally familiar with the content in both interviews. The

cx-u-‘I- -.-_—w.\w;‘_; ' ’

 

'-=,. mmm’é-i; all." gag-.7 ..7: -4... - VJ

 

 

 

11

subjects were randomly assigned to one of three groups. One group
(management group) was given only the management interview; the
second group (project group) was given only the project interview;

and the third group (management—project group) was a control group
and got both the management and project interviews. The subjects
were divided into these groups because the main purpose of the Wickert
study was to determine which of the two types of communication was
flowing the more smoothly, the communication up and down the hier-
archical structure as compared with the communication among peers
on a particular project. Communication was to be measured by find—
ing out the extent to which immediate superior subordinate pairs in

the hierarchy for each type of communication responded in the same
way to the questions in the interview for their group. If agreement

in one group for one type of communication was statistically signifi-
cantly greater than in the other group for the other type of communi-
cation, the experimenter might conclude that the type of communication
in the former group was better than the type of communication in the
latter group.

In both of the two types of 1965 or second phase interviews,
the last question in the series of questions asked was the standard
question. This question was: "Think back on how things were in the
Engineering Division early last Spring-—about March of 1964. Now,

what do you see as the important, the major changes in the administration

 

 

 

 

 

12

and management--and the general climate of the division since then--
the few really big changes ? Please try to list these."

It was desired in the Wickert study to consolidate the
responses to this standard question asked immediately following both
the management and project interviews so that both sets of answers
could be used to indicate what the subjects perceived as the changes
introduced in the company during the past year.

In an exploratory look at the data it appeared that the con-
tent of the responses reflected the content of the two different immedi-
ately preceding series of questions. So that combining might be
justified, the purpose of the present research was to discover to what
degree the content of the series of immediately preceding questions
influenced the content of the responses given to the standard question
at the end of the series.

Since the management interview contained a series of
questions which asked for general (company-wide) information and
since the project interview contained a series of questions that asked
for, relatively specific information, it was hypothesized,as an alterna-
tive to the null--no—effect--hypothesis, that the two different series
of questions would establish two different mental sets and thatthese
would influence the content of the responses to the standard question.
The management interview would establish a general set by the nature

of its content and would tend to elicit non-detailed responses that deal

 

 

 

 

 

13

with the entire organization. On the other hand, the project interview
would establish a specific set and tend to elicit detailed responses that
were specific to a particular project or work area; it would neglect to
pick up changes that affected the entire organization.

If this alternative hypothesis was confirmed, it could be
said that the content of the responses was a function of the content of
the series of previously asked questions. An immediate practical
implication was that the responses to the standard question in the
Wickert study could not be conveniently combined but must be treated
separately. The content of the series of previously asked questions
for each of the groups was different, and consequently different mental
sets may have been established which in turn may have affected the
kind of answers given to the standard question. On the other hand,
if the alternative hypothesis was not confirmed, that is, the null
hypothesis was not rejected, the responses could be consolidated and
treated as one mass. The responses would not have been influenced
by the content of the precedingly asked questions. Judgment would
then have to be withheld regarding the extent to which the content
of a series of questions affects the content of the response to a
particular queStion that follows the series.

The long-range implication of confirming the hypothesis
would be that an experimenter using interviews containing a series

of questions under circimstances similar to those prevailing in this

 

 

 

 

     

14

study would have to consider the effect of the content of previously
asked questions on the content of the response to a question of
interest. If the null hypothesis was not rejected, judgment would
have to be withheld regarding the effect of the content of a series of
questions onthe content of the response to a particular question that
follows the series.

It was the hope of the experimenter that the results of
this and similar research would eventually be combined to form
"norms" which could be referred to by any experimenter working
with complex verbal material and wishing to have some idea for
advance planning, about factors that should be taken into account

because of their possible effects.

 

HYPOTHES ES

The present research was designed primarily to investi-
gate the extent to which the content of an immediately preceding series
of questions influenced the content of the responses given to a standard
question that followed the series.

The following general null hypothesis was given regarding
the intent of this research.

The content of the immediately preceding series of questions
would have relatively no effect on the content of the responses
given to a standard question that followed the series.

Upon rejection of this null hypothesis the following alterna-

tive hypothesis was suggested.
The content of the immediately preceding series of questions
would establish a mental response set in the responding individual
predisposing him to respond to the standardquestion in such a
way that the content of his response would reflect the content of
the preceding series of questions.

In terms of the nature of the present research the alterna-
tive hypothesis was broken down into three specific aspects to be tested,
and each aspect independently would have to be confirmed in order that
the null hypothesis may be rejected. These three aspects to be tested
were:

1. There would be no statistically significant difference

between the proportion of general and/or specific statements found for

15

 

 

 

 

 

 

16

the responses to the standard question for either the management or
project interviews. This statement assumed that, (a) there were just
a certain number of units of information contained in each statement
(response) and (b) these statements were accurately interpreted.

Z. A greater proportion of general statements (as defined
previously) would be found in the responses to the standard question
for the management interviews than in the responses to the standard
question for the project interviews. This statement assumed that a
general mental response set was established by the content of the
general questions in the management interview.

3. A greater proportion of specific (as previously defined)
statements would be found in the responses to the standard question
for the project interviews than in the responses to the standard ques-
tion for the management interviews. This statement assumed that a
specific mental response set was established by the content of the

specific questions in the project interview.

 

 

 

 

 

METHOD

Subjects
The subjects in this study were 55 engineers, all employees

of the same aerospace company in Michigan.

Materials

a) Interviews. --Two interview schedules, constructed for
the Wickert field study previously referred to, were utilized. Both of
the interview schedules, it will be recalled, had a similar structure
but differed in content. One interview (management interview) asked
about managerial control procedures while the other interview (project
interview) asked about some particular project the subject was work-
ing on. Both interview schedules contained nineteen questions. The
nineteen questions in the management interview schedule were divided
into six major areas, and the nineteen questions in the project inter-
view schedule were divided into ten major areas. The last question
in each interview was standard and asked: "Think back on how things
were in the Engineering Division early last Spring--about March of
1964. Now, what do you see as the important, the major changes in
the administration and management——and the general climate of the
division since then-—the few really big changes ? Please try to list

these.’l

 

 

 

 

18

b) "Code Book". --A "code book" for the standard ques—
tion was developed by categorizing the responses to this question.
Two male graduate students acted as judges and independently read
the responses made by the engineers to the standard question and de-
veloped categories, represented as codes in a code book, which ex—
pressed the unitary ideas that were found in the responses to the

standard que stion .

Procedure

A representative sample of subjects, each having generally
both a management and project role, were randomly assigned to one of
three groups: a management group (composed of thirty engineers who
received only the management interview); a project group (composed
of twenty—nine engineers who received only the project interview); and
a management-project group (composed of nine engineers who received
both interviews) .

Seven interviewers were randomly assigned to interview
the engineers in all three groups. The interviewers, all male graduate
students in psychology, took notes during these interviews and these
notes were recorded on a dictaphone within two hours after the interview.
The dictaphone belts were transcribed and these transcribed interviews
were then coded.

The "code book" for the standard question was developed
by two of the seven interviewers by categorizing the ideas expressed in

the response to this question.

 

 

 

 

 

19

After the code book had been developed, it was given to
two judges who had a good inderstanding of this project. These judges
independently read the code book and indicated which categories they
thought were general and which categories they thought were specific.

The instructions and rules by which such decisions were
made were as follows:

The criteria to be used in deciding which categories are
general and which categories are specific are the (a) Generality-
specificity of the statements. A general statement is one that
pertains to the total or whole organization and does not contain
any concrete example of changes perceived by the subjects. A
specific statement is one that presents concrete examples of the
changes perceived. (b) The referent of the statement, that is,
whether the statement referred to the entire organization or just
to his job, project or work area. These criteria were chosen
because it was hypothesized that there would be a difference in
the generality-specificity of the content of the responses to the
standard question for the management and project interviews.

The rules by which such decisions are to be made are as follows:

a) If the statement is general in the sense that it does not
present any concrete example of the perceived change,
mark it general. For example, "Things are better now. "

b) If the statement mentions a concrete change, mark it
specific. For example, "There is less paper work now. "

c) If the statement refers to the entire organization, mark
it general. For example, "The company is more
efficient. "

d) If the statement refers to the subject's perceived change
in status concerning his job, project or work area, mark
it specific. For example, "I've been promoted. "

In some categories both the elements of generality-specificity
and entire organization--job, project and work area are present.
The rule for deciding which categories are general and which are
specific for these categories are as follows:

a) If the statement refers to the total organization but is
general in the sense that it does not present any concrete
examples of the changes perceived, mark it general.

For example, "The climate of the company has improved. "

b) If the statement refers to the entire organization and gives
concrete examples, mark it specific. For example, “There
is more R and D money available in the company. "

 

 

 

20

c) If the statement refers to the individual's job, project,
or work area, but is general in the sense that it does
not contain any concrete perceived changes, mark it
general. For example, "Working conditions are better
in my department. "

d) If the statement refers to the individual's job and gives
concrete examples, mark it specific. For example,
"I've got a new department head."

The judgments of the two judges were correlated to see to
what extent there was reliability between these two judges on judging
the generality-specificity of the categories. A phi coefficient computed

on these data was .915.

 

 

 

 

 

 

.492 .508 1.000
N s .033 .426 .459
it
E G .459 .082 .501 4> = .915
G 5
Judge 1

According to Guilford, this would be a conservative estimate of the
correlation between the two raters since this phi~coefficient under—
estimates the Pearson r when p is not . 5 (Guilford, 1956) . The
categories disagreed on were ironed out, in discussion, allowing for
a common agreement as to which categories were general and which
were specific for the purposes of establishing one set of data for use
in the next research operation. The categories in the "code book"
were then marked general or specific according to this common

agreement .

 

 

 

21

Coder selection was based on an assumption underlying
the "classical" approach to content analysis. This assumption stated,
in essence, that data which has to be coded into categories reflecting
the content of the material can be analyzed by anyone without training
(Berelson, 1952) . Under this assumption the coders utilized happened
to be a male graduate student, twenty-two years old, and a housewife
of forty-five with two years of college education.

The materials needed to do the analysis were then given to
the two different coders, who then independently coded the responses
to the standard question for each interview. More specifically, these
coders marked on a code sheet that contained all the categories in the
"code book" the categories that they thought were expressed in the
responses for each of the subjects.

To prevent the coders from associating the responses
with the respondents and therefore with what type of interview they
received, numbers were assigned to each response which were then
taken out of the interview proper and copied on plain white paper.

The coders used these copied responses to code from and not the
original interview.

The instructions given the two coders were in part
written and in part verbal.

The written instructions were:

The coder should pay special attention to the generality or

specificity of the interviewee's statements that he is coding
because there is a plan to use the degree of generality or

 

22

specificity as an experimental variables on this question. If

an interviewee makes a vague, general, or global reference,

his reference should be coded in a general category; if, on the
other hand, he makes a specific statement, it should be coded
as specific. If, however, he starts off, for example, by
making a specific statement and then says that this is just an
instance of something broader, or if he makes a broad statement
and then gives specific examples, both general and specific
codes should be used.

In addition, the experimenter gave a verbal explanation
which consisted of essentially the same information contained in the
written instructions. These verbal instructions were relatively con-
sistent for both coders.

A master code, for which a third person arbitrated the
disagreements between coder one and coder two, was developed so
that there would be a single code expressing the ideas found in the

responses to the standard question. Independent records were pre-

served for all three sets of codings.

 

 

RESULTS

Data Processing

 

To determine to what extent there was agreement between
the two coders in the coding of the responses to the standard question,
the experimenter correlated (product—moment correlation):

a) the number of "general" categories used by each coder to
code each response to the standard question.

b) the number of "specific" categories used by each coder to
code each response to the standard question.

c) the "total number" of categories used by each coder to code
each response to the standard question (a + b = c) .

In addition, the experimenter obtained:

_ n( ADE)

a fa t —
a) percent geo greemen n +n

 

A B n (A QB)
nA + nB
b) average percentage of agreement : N

number of categories used by any
one coder.

Notation: n

N = total number of categories used by
both coders.

A = refers to the categories used by
coder L

B = refers to the categories used by
coder 2.

In order to test the hypothesis, independent 't' tests were

performed:

23

 

 

24

a) between the Management and Project interviews on the pro-
portion of ( 1) general categories found and (2) on the pro-
portion of specific categories found in the responses to the
standard question for coders 1 and 2 independently.

b) between the Management and Project interviews on the pro-
portion of ( 1) general categories found and (2) on the pro-
portion of specific categories found in the responses to the
standard question for the Master Code.

Correlated 't' tests were performed:

a) between coder 1 and coder 2 on the proportion of general
categories used to code the Management interviews and the
proportion of specific categories used to code the Manage-
ment interviews.

b) between coder 1 and coder 2 on the proportion of general
categories used to code the Project interviews and the pro-
portion of specific categories used to code the Project
interviews.

The process of content analysis requires a coder to evalu-
ate a response, or part of a response, from a personal point of view,
and to interpret each statement in terms of one or more of the cate-
gories presented in a code book. To establish reliability generally,
at least two coders should be used. The greater the degree to which
there is agreement between the two coders, the higher is the probability
that both coders interpret the statement using the same criterion as a
frame of reference. In this study the criterion was the generality
or specificity of the statements, and this frame of reference was
created by the instructions given to the coders. Thus, the more
reliable the two coders were in coding the response to the standard
question, the more likely these coders interpreted the same state-
ments as general or as specific and the more likely they would use
the same number of categories to code the responses. Therefore,

any significant statistical difference or lack of significant statistical

 

 

     

25

difference found between the management and project groups would have
a firmer statistical basis because of this reliability check. Without
such knowledge, it would not be known to what extent a found signifi-
cance or lack of significance found was due to a true difference or a lack of
a true difference between the populations or because the coders were
using different criteria to evaluate and interpret the responses.

The results may not be due to the generality or specificity of the
response, but rather to some unknown factors.

To determine how consistently the two coders were in
coding--to what extent they agreed in the coding of each response,
product moment correlations coefficients were computed. These
inter-rater reliabilities were performed on:

a) the "total number" of categories used by the coders in
coding each response to the standard question for all the
interviews.

b) the number of "general" categories used by the coders in
coding each response to the standard question for all the
interviews.

c) the number of "specific" categories used by the coders in
coding each response to the standard question for all the
interviews. I

The results of these calculations are shown in Table 1.
It can be seen by looking at this table that there was more agreement
between the coders with regard to the "total" number of categories

and the number of "specific" categories used than for the number of

"general" categories used by each coder in coding each response.

 

 

 

26

The table also shows that none of these correlation coefficients were
high. Therefore, there was relatively low agreement between coders
in the way they coded each response. The coders did not interpret
the responses in the same way. They saw different units of informa-

tion in the content of each statement.

 

 

 

Table l. --Inter —rater correlations.

r . 57 . 41 . 56
the total number of the number of the number of specific
categories used by general categories categories used by
Coder l and Coder used by Coder 1 Coder l and Coder 2
2 for coding each and Coder 2 in in coding each response
response for all the coding each re- for all the interviews.
interviews. sponse for all the

interviews.
p<-01 p<.01 p<.01

An alternative procedure for determining the reliability of
the two coders was to calculate the percentage of agreement between
the coders for each response for all interviews. Using the TAT
(Atkinson, 1958) method of obtaining percentage of agreement, the
number of categories used by both coders in coding the response for
each interview was divided by the sum of the total number of categories
used by each of the two coders independently for each interview. These
percentages are reported in Table 2.

An inspection of this table will reveal that out of 67 re-
sponses there Were 7 responses in which there was absolutely no

agreement between coders and the highest percentage of agreement

 

 

 

 

 

27

 

 

.N. .NN we? usofioohmm mo omdusounmm ammuoxfiq.

 

 

 

 

9:. we or: S. 0.3 en ii 2
m.mm $ o.o~ cm 0.8. mm i: 3
m .3 8 e .3 3. wen mm o 3
-- me o 3. o am «.2 E
0.2 we m .2 5. 92 on 93 2
7mm 3 or: 3. 3.: am mgm .2
o.m~ no cg: 2. 0.3 mm 0.3. 2
9mm 3 ed... 3 i: S. 0.8. 2
~.- 8 m .5 2. N.~N em 22 0
9.: am o.o~ me 0 mm ~.- w
o.o~ mm v.3 2. mien 3 mix 5
er: 3. o.o~ ow mifl mm Nam e
o .8 em. 0 a... 9%.. NM v.3 m
53 mm 0 mm o.o~ an 9%. e
o.m~ em 93 S 0.8 om m .3 m
73. mm O on o.o~ S 93 N
«12 mm ms; mm or: m: 58 z
“:08 acme “Gog have
-wonwr. nonfisz drum/w Honfidz ..oonm< HoQESZ noonwaw nonasz
mo as BSKEBQH «0 as BOT/HQEH we as BUT/Hound «0 one Beau/nous;
.mgwwxrhmuflm

9.3 Zn HOW uncommon gone no N popoO pad H nopoO nook/won. Emcfioonwd mo eunucoonom 9.33. .N 3me

 

 

 

 

28

between coders, 50 percent, occurred in one responSe. The average
percentage of agreement was 22. 7 percent for each response. The
results of these calculations also indicated that there was little agree-
. ment between coders in coding each response.

A series of correlated 't' tests were performed to see
whether any significant statistical differences could be found between
the coders in coding the management interviews general and coding the
project interviews specific, that is, between the proportion of general
categories used by Coder 1 vs. the proportion of general categories
used by Coder 2 in coding the management interviews, and the propor-
tion of specific categories used by Coder 1 vs. the proportion of specific
categories used by Coder 2 in coding the responses to the standard ques-
tion in the set of project interviews.

Assuming that there were just a certain number of units of
information contained in each response and that the coders interpreted
the responses from the same frame of reference, that is, from the
generality or specificity of the statements as defined previously, and
in accordance with the alternative hypothesis, no statistically signifi-
cant difference would be found between the proportion of general and the
proportion of specific categories used by each coder to code the responses
to the standard question for both the management and project set of inter-

views .

 

 

 

 

29

Tables 3 and 4 give the results of these tests. As indicated
by the tables, significant statistical differences were found between
coders for coding the management interviews ( 1) general and (2) specific
at the .01 confidence level, and for coding the project interviews ( 1)
general and (2) specific at the . 01 confidence level. In three of the
cases the mean proportion of categories used to code ( 1) the manage-
ment interviews general (2) the management interviews specific, and
(3) the project interviews general was greater for Coder 2 than for
Coder 1. Coder 1 had a higher mean proportion than Coder 2 for coding
the project interviews specific.

Since neither the reliability coefficients nor the percen-
tages of agreement were high, and since the 't' test performed on the
number of units of information contained in the responses indicated
relatively little agreement, it was decided that in order to test the null
hypothesis at all, independent two -tailed 't' tests would have to be com-
puted on the data for each coder independently.

If the mental sets hypothesized were established, and if
they affected the content of the responses in the predicted direction,
there should be a significant difference between the management and
project interviews in both the proportion of general categories used
and in the proportion of specific categories used by each coder. More
specifically, statistically significant differences should be found between
the management and project interviews in the proportion of general

categories used and in the proportion of specific categories used.

 

 

 

 

30

Table 3. --Mean difference, standard deviation of the difference, 11,

degree of freedom, and correlated 't' between Coder 1 and Coder 2 on

coding the responses to the standard question in the management inter-
views. a) General, b) Specific

 

 

Mean Standard Deviation Degree of
Difference of the Difference 11 Freedom 't'

 

a) General

 

 

 

 

Coder 1
vs. Coder 2 .259 .026 33 32 9.181
P < .01
b) Specific
Coder 1
vs. Coder 2 . 249 .024 33 32 10. 375
p < .01

Table 4. --Mean difference, standard deviation of the difference, n,

degree of dreedom, and correlated 't' between Coder 1 and Coder 2 on

coding the responses to the standard question in the project interviews.
a) General, b) Specific

 

 

Mean Standard Deviation Degree of
Difference of the Difference n Freedom 't'

 

a) General

 

 

 

Coder 1
vs. Coder 2 . 254 .033 34 33 8.194
p < . 01
b) Specific
Coder 1
vs. Coder 2 . 240 .023 34 33 9. 230

 

p<.01

 

 

 

According to the alternative hypothesis, there should be a greater pro-

portion of general categories used in coding the management interviews
than in coding the project interviews, and, conversely, there should be
a greater proportion of specific categories used in coding the project
interviews than in coding the management interviews.

The results of the two —tailed independent 't' tests performed
to test this aspect of the hypothesis are shown in Tables 5 and 6. For
Coder 1 (Table 5) , no statistically significant difference was found
between the proportion of general categories used to code the responses
to the standard question for the management interviews and the proportion
of general categories used to code the responses to the standard question
for the project interviews. Neither was there any significant difference
found between the proportion of specific categories used to code the
responses to. the standard question between the management and project
interviews. This aspect of the alternative hypothesis was not supported
by the data obtained for Coder l.

The results for Coder 2 are shown in Table 6. Here, also,
no statistically significant difference was found between the management
and project interview in the proportion of general categories used to
code the responses to the standard question in each group, but there was
a statistically significant difference at the 1 percent level between the
management and project interviews in the proportion of specific cate-
gories used. The mean proportion of specific categories used in the

management interviews was greater than the mean proportion of specific

 

 

    

 
     
 
   
 

I: 9‘”
I - #0
I: '8

.

 

  

32

Table 5. --Mean proportion, standard deviation of the proportion, n,

degree of freedom, and 't' between the management and project inter-

views for Coder 1 on the proportion of general and specific categories
counted in the responses to the standard question.

 

 

 

 

 

 

 

Mean Standard Deviation Degree of
Proportion of the Proportion 11 Freedom 't'
General
Management . 520 . 061
. 1
vs. Project .517 .054 67 65 05
Specific
Management . 478 . 061
. 1
vs. Project .469 .054 67 65 0 7

 

 

Table 6. -—Mean proportion, standard deviation of the proportion, n,

degree of freedom, and 't' between the management and project inter-

views for Coder 2 on the proportion of general and specific categories
counted in the responses to the standard question.

 

 

 

 

 

 

Mean Standard Deviation Degree of
Proportion of the Proportion n Freedom 't'
General
Management . 541 . 122
vs. Project .634 .074 67 65 Laos
Specific
Management . 556 . 122
vs. Project .364 .082 67 65 2'667

 

p<.01

 

 

 

 

33

categories used in the project interviews. This was not in accord with
the null hypothesis nor with the alternative hypothesis. If any difference
should be found according to the alternative hypothesis, it should be the
converse.

The results obtained for both coders did not allow for
rejection of the null hypothesis and judgment had to be withheld as to
what extent the content of a series of previously asked questions in-
fluenced the content of the responses to a following question.

Since an index of the changes perceived by the subjects
was needed for Wickert's study, an attempt was made to develop a
master code by having an arbitrator reconcile the disagreements be-
tween the coders in coding each response to the standard question.

This was done, and the present experimenter then assumed the master
code to be another possible criterion on which to test the hypotheses.

Two-tailed independent 't' tests were performed on the
data to see if the responses to the standard question for the management
group differed from the response to the standard question for the pro-
ject group in the proportion of general categories found in these
responses, and conversely, to determine if the responses to the
standard question for the project group differed from the responses to
the standard question for the management group in the proportion of
specific categories found (Table 7). A 't' of 5.839, statistically
significant at the . 01 level, was found between the set of management

and project interviews for the proportion of general categories found.

 

 

 

 

34

In accordance with the alternative hypothesis a greater proportion of
general categories was used in coding the management interviews than
was used in coding the project interviews. This difference, since
everything else in the two groups was the same, was attributable to the
general set established by the content of the questions previously asked.
The mean proportion of general categories found in the response to the
standard question for the management group was greater (statistically
significant) than the mean proportion of general categories found in the
responses to the standard question for the project interviews.

Table 7. --Mean proportion, standard deviation of the proportion, n,
degree of freedom and 't' between the management and project groups

for the Master Code on the proportion of general and specific categories
found in the responses to the standard question.

 

 

 

 

 

 

 

Mean Standard Deviation Degree of
Proportion of the Proportion n Freedom 't'
General
Management . 655 . 053
vs. Project . 328 .048 66 64 5'839
p < . 01
Specific
Management . 436 . 045
66 4 1.
vs. Project .539 .050 6 981
p < . 01

A 't' of l. 981, statistically significant at the . 01 confidence
level, was found between the management and project sets of interviews

for the proportion of specific categories used. The significant 't' was

l . I .

4-1."..- - nu

 

 

- .4 .1. .. l , . . : . . o- - ,3
i, \\ tut. L. .. . vo... a, .. . . . i.e.!hll: finial!) 17., .
.¢. .- ... I. .5 Inhfcl . . . , .LJV. vidal.) Iii. , ii i,
v , in I.l . .7... .I. D , i It ,
. , f i F v , .

 

 

 

35

also in accord with alternative hypothesis, that is, a greater proportion
of specific categories was used in coding the project interviews because
of the specific set established by the specific questions that were
previously asked. The results from the master code supported the al-
ternative hypothesis, that is, using these results as a criterion, the

null hypothesis could be rejected.

 

 

 

 

SUMMARY OF THE RESULTS

1. The results of the correlated 't' tests between Coder 1
and Coder 2 on

a) the proportion of general categories found in the responses
to the standard question for the management group,

b) the proportion of specific categories found in the responses
to the standard question for the management group,

c) the proportion of general categories found in the responses
to the standard question for the project group,

d) the proportion of specific categories found in the responses
to the standard question for the project group

were as follows: There was a statistically significant difference found
for all cases. In cases a, b, and c, Coder 2 had a higher mean pro-
portion than Coder 1 and in case d Coder 1 had a higher mean proportion
than Coder 2. According to the position taken by the experimenter re-
garding the number of units of information contained in each response,
the coders would have used the same proportion of categories in their
coding if they were coding reliably. Thus, there was little, if any,
agreement between coders.

2. The results of 't' tests computed on the data obtained
from each coder independently and used in comparing the management
and project groups for the proportion of general categories and the

proportion of specific categories used where: (a) No statistically

36

.. £3“: ":5"

 

 

 

37

significant difference was found for Coder l in the proportion of general
categories found in both the management and project interviews, nor
was there any difference in the proportion of specific categories found
in the responses to the standard question for both the management pro—
ject groups. According to the alternative hypothesis a statistically sig-
nificant difference would have been found—-a greater proportion of
general categories would have been found in the responses to the stan-
dard question for the management group, and a greater proportion of
specific categories would have been found in the responses to the stan-
dard question for the project group. (b) No statistically significant
difference was found for Coder 2 in the proportion of general categories
used to code both the management and project interviews, but there was
a statistically significant difference found for the proportion of specific
categories used to code both the management and project interviews.
Coder 2 used a greater proportion of specific categories to code the
management interviews than to code the project interviews. This result
was not in accord with the null hypothesis, nor with the alternative
hypothesis. In fact, it was the converse.

None of the results obtained from treating the data of the
coders independently supported the alternative hypothesis.

The results of the 't' tests performed on the master code
between the management and project interviews with regard to, (a) the
proportion of general categories, and (b) the proportion of specific

categories found in the responses to the standard question were both

 

 

 

38

significant. A greater proportion of general categories was found in
the management interviews than was found in the project interviews,
and a greater proportion of specific categories was found in the project
interviews than was found in the management interviews. These results
rejected the null hypothesis and supported the alternative hypothesis.
Since none of the results obtained on the tests performed
on the two coders independently supported the alternative hypothesis
and the reliability of the master code could not be computed from the
data collected, the experimenter reserved the right to withhold judg-
ment as to what extent the content of a series of previously asked
questions had on the content of the responses to a standard question
that followed the series, but wished to comment that, in his opinion,

such a trend was present in the data.

Table 8. --Summary table of results.

 

 

 

 

 

 

 

 

Management Group Project Group
General Specific General Specific
It! ltl It! ltl
Coder 1 vs.
. 1 . . .
Coder 2 9 81 10 375 8 194 9 230
p<.01 p<.01 p<.01 p<.01
Coder l Coder 2
General Specific General Specific
It! It! ltl It!
Managerilent .051 .017 1. 208 2.667
vs. Pr0ject
p < . 01
Master Code
General Specific
ltl ltl
Management 5. 839 1. 981

vs . Project

 

p<.01 p<.01

 

 

DISCUSSION

From the general observation of the conditions under
which this study was made, it may be that the failure to support a
number of the aspects of the hypothesis was due to some of the tech-
niques and methodology used rather than to inadequacies of the hypo-
thesis.

Keeping within the realm of the findings, however, the
major, overall conclusion drawn from the investigation was that
judgment should be withheld as to what extent the content of a series
of previously asked questions affected the content of the response to a
standard question that followed the series. The content of these ques-
tions may have had an effect, but because of some other variables
that may have been operating or because the nature of the effect was
not the one hypothesized, it was not possible to detect it. Because of
these possibilities, it would not be justifiable to conclude that the
content of the question had "no appreciable" effect on the content of
the response to the standard question, but rather, to say judgment
was suspended with regard to this question.

From the point of view of the larger study from which
this study originated, this conclusion permitted the combination of the
perceived changes from the management and project groups. The

39

 

 

40

combined index could be used to indicate the changes perceived by the
individuals during the past year as was originally intended, instead of
listing the changes reported by each group independently.

Moreover, Parten's assertion that the general questions
should precede the specific questions or else the subjects would tend
to answer the general questions in terms of the specific questions pre-
viously asked, did not seem plausible in light of the present research.
If this statement were correct, a greater proportion of specific cate-
gories should be found in the responses to the standard question--which
was relatively general-- for the project interviews since it was preceded
by specific questions than should be found in the responses to the stan-
dard question for the management interview which was preceded by
general questions. As was mentioned before, this was not found in the
present research, but was indicated by the trends in the Master Code
results.

Furthermore, the specific conclusions to be made on the
basis of these findings must be considered tentative in view of the
study's limitation (which shall be duscussed) and because of the posi-
tive trends indicated by some of the specific findings. These results
indicated the desirability of further research.

The most meaningful results obtained in this research were
those gained from the use of the Master Code. If these results were
the sole criteria on which to judge the hypothesis, that is, disregarding

the data of the two independent coders, there would have been more

 

 

 

 

 

 

41

evidence in favor of the alternative hypothesis than against it. The
overall conclusion would have been a confirmation of the alternative
hypothesis. But, since the master code was kind of a compromise
on disagreed categories between coders, and no attempts do deter-
mine its reliability were made, and since none of the results ob—
tained from the data of the two independent coders supported the
alternative hypothesis, the experimenter felt that such a conclusion
would not be justified. Nevertheless, the findings from the master
code were, in the experimenter's opinion, indicative of what was
actually happening, and if some of the technical and methodological
limitations of the study could be reconciled, positive results might
be found.

In the following paragraphs, the experimenter discusses
some of the, study's limitations and speculates as to why negative
results were found.

Probably, the main contributing factor which influenced
the results that were obtained in this experiment was the relatively
low reliability between the two coders. Some of the factors that may
have contributed to this relatively low reliability and /or may have
influenced the outcome of this experiment regardless of the reliability
found will be duscussed in the following paragraphs.

One such factor was the inequality of previous knowledge
held by the coders regarding the project. One of the coders, Coder 1,

knew quite a bit about the larger study, the nature of the company, its

 

 

 

 

 

 

 

42

products and terminology, and the nature of the present research. He
was an interviewer for the larger study. These facts, themselves,
may explain the great discrepancy between the coders in coding each
response. The other coder, Coder 2, was completely naive, and often
complained that some of the technical jargon used in the responses was
baffling and that she could not determine whether or not certain units
of information were general or specific, as defined by the code book.
She did not know when a change reported was a change that affected the
entire organization or just the respondent's job.

The ability to code seemed to be a function of the previous
knowledge and experience the coders had with the subject matter. This
finding contradicted Berelson's statement regarding the asset of not
having to train coders when using the "classical" approach to content
analysis.

It was also possible that because of this previous knowledge,
each of these coders was using a different frame of reference in their
coding, Coder 1, his previous knowledge and Coder 2, her intuition.
Perhaps, this was why their data were so unreliable.

It was not the experimenter's opinion that the coders should
be equally knowledgeable with reference to the content to be analyzed,
but, rather, each should be as much of an "expert" as possible con-
cerning the operations of the company in question. In this way the

coders would know what a particular unit of information in a particular

 

 

 

 

43

response referred to and would not have had to speculate as to the cate-
gory in which it should be coded.

The coder should not only have knowledge of the content of
the responses, but he should also be trained in coding methodology.

It has been shown time and time again that one of the greatest factors
contributing to the unreliability of the coders was lack of training
(Marsden, 1964) . Training entails, essentially, the establishment of
a working knowledge of the code book, thereby eliminating any confusion
or mistakes that might arise because of failure to understand the basis
on which the code book was devised. Lack of training often results in
erroneous coding.

The next major factor influencing the outcome of the experi-
ment was the code book itself. Maier, fl. ( 1961) commented that
the relatively unstructured interview was a valuable data-gathering
technique which could elicit each individual's complete, yet, uninfluenced
response without the limitations of a structured interview, and provided
an indication of their importance, but these advantages were purchased
only at the expense of certain concurrent difficulties in analysis.

One of these problems, according to Wispe (1954) , was
category derivation. In a study such as this, i. e. , a study concerned
with the examination and interpretation of interview data, Wispe would
hold that the categorization process was crucial, since all subsequent
results depended upon the categories into which the responses were

analyzed. Empirically derived categories, that is, categories that

 

 

   

44

were relatively inductively derived from the data to be analyzed, must
not reflect every shade of meaning. Categories that did were pushed to
their limit. They precluded analysis and were psychologicallymeaningless
because the coder was reduced tolisting isolated segments of a psycho-
logical process from which no generalizations could be drawn.

It was interesting to note that one of the coders reported
that the categories were divided into so many minute subcategories that
she had difficulty discriminating between them. She went on to say that
maybe the people who composed the code book knew the difference
between some of the categories, but she did not. Thus, when she was
faced with a unit of information that could belong to one or more of the
categories, she arbitrarily put it in any of them because of not knowing
the difference between these seemingly identical categories.

To add to the confusion of the coders, oftentimes general
categories were mixed in with specific categories which, in turn,
were all subcategories under a larger specific category and vice versa
(see Appendix C) . This confusion might have provoked the coders to
put a response, or part of a response, into the first category it fit re-
gardless of whether the category was marked general or specific or
whether the coder thought the statement was general or specific. If
this kind of behavior occurred, the whole purpose of the research was

defeated.

 

 

 

 

 

45

Ideally, the categories should be wide enough to be profit-
able and comprehensive, but not so broad as to become psychologically
meaningless; nor should they be so narrow to be unanalyzable (Wispe,
1954) .

Furthermore, it was not known by the experimenter to what
extent the categories reflected the generality or specificity of the con-
tent of the responses. If these dimensions were not reflected in the
code book, it would be impossible for the coders to code the responses
in any meaningful way with regard to these dimensions. If this had been
the case, it was understandable why the predicted result was not obtained.

Since placing a response in a general or specific category
was a subjective cognitive judgment, a plausible explanation for the
negative results may be the fact that the coders did not agree with the
raters who originally decided which categories were general or specific,
and thus, may not have abided by the guide lines established by the ex-
perimenter. Thus, what was observed may have been a purely phenome-
nological process.

It was also possible that a coder "set" may have been estab-
lished as a result of coding the responses. After coding a number of
similar responses into the same categories the coders may have been
predisposed to use these categories, and therefore, tended to use them
even when the data called for different categories. Along this same
line, and also, contaminating the results, was the possibility that the

previous experience of Coder 1 may have acted as a bias predisposing

 

 

 

46

him to react to the responses using the impressions be obtained from
his interviewing experience.

Finally, the responses to the standard question not only
contained information called for by the question but also the subject's
comments. What were analyzed Were not changes per se, but also the
subject's opinions and evaluative comments about the perceived changes.
The analysis of the information not called for by the question--opinions
--may have counteracted any effect the mental sets established by the
content of the questions had on the pertinent parts of the response.

At the expense of repeating myself, if the coders were
trained and were experts regarding this research, thus minimizing
such contamination, they would have been able to differentiate more
easily the responses of interest and disregard the rest.

In concluding this discussion concerning the results of
this research, some mention should be made of the possible psycho—
logical processes underlying the behavior displayed in this study, and
these may help explain why negative results were found.

An article entitled "The question-answer process--a
conceptualization and some derived hypotheses for empirical examina-
tion" by Getzels ( 1954) attempted to explain what was happening in the
question-answer process. Parts of the author's explanation may be

appropriate to this research and shed some light on its findings.

 

 

 

47

The following paragraphs contain what was essentially
Getzel's understanding of Bruner's conceptualization of the psycho-
logical processes going on in an individual when he responded to a
visual stimulus, reinterpreted to fit this study.

Little energy has been used for asking what exactly was
represented in a respondent's answer to a question. How much of
the answer reported the individual's spontaneous reactions, how much
his ego-enhancing defences, how much the requirement of the particu-
lar situation within which the questions are asked?

While investigators in other fields of psychological func-
tioning were focusing their experiments on the process intervening
between a stimulus and observed response, investigators in the field
of opinion and attitudes were content, for the most part, to focus
their research on the observed response, i. e. , the answer to the
question, without probing into the underlying process represented by
the observed response. Getzels stated that there was no a_p_ riori
reason why the process of responding to verbal stimuli should be con-
sidered any simpler, for example, than the process of responding to
a visual stimulus. Indeed, since the former was more likely to in-
volve social interaction, the expectation could have been that it was
more complex.

The model for a preliminary formulation of the que stion-
answer process could be borrowed from Bruner's recent theoretical

and experimental work in the field of perception. A reaction to a

 

 

 

48

perceptual stimulus was conceived of as involving a three-step cycle.
Perceiving, as formulated by Bruner, began with an already existing
"hypothesis. " The assumption was that individuals were never ran-
domly set or eingestellt but were inevitably prepared for seeing,
hearing, smelling, tasting, some particular thing or class of things.
The second analytic step in the perceiving process was the input of
information from the environment. The third step was a checking or
confirming procedure. The new information was congruent with the
operating hypothesis or was in varying degrees incongruent. If con-
firmation did not occur, the hypothesis shifted in a direction partly
determined by internal factors and partly by the information factors
in the unsuccessful information checking cycle.

If an individual was ready to answer a question at all, he
was ready to answer it in a given way-~he had a hypothesis. However,
it could not be assumed that the question-answer sequence was a push-
button affair; ask a question and an answer would automatically come
out. Any question was embedded in a context--the topic of which was
under investigation here--but, what was usually not considered was
that the answer was embedded in an answerer. The relationship
between question and answer was not uni—dimensional but multi-
dimensional.

Using Bruner's formulation, a question tripped off a per-
sonal hypothesis. This was the first step of the question—answer

process. But it was an internal reaction. It could sometimes become

 

 

 

 

 

49

manifested in a blush, a stammer, or block or so-called "spontaneous
reaction" or an impulsive remark, but before it was ordinarily made
verbal, two other factors entered into consideration: the situation of
the questioning and the motivations of the individuals in the particular
situation.

A question was not a stimulus in a vacuum; it had a situa-
tional context. A question asked by a therapist as to how one felt toward
his family was one thing, by a neighbor another, by a stranger yet
another. There were social expectations defining what was appropriate
behavior and social norms defining responses that were congruent with
appropriate behavior. The personal hypothesis tripped off as the im-
mediate reaction to the question was checked against the requirement of
the situation. Appropriate behavior in the case of threatening objects
of inquiry could have permitted revelation of the personal hypothesis,
or may have required censorship and distortion on the personal hypothesis,
or may have required censorship and distortion of the personal hypothesis
in light of the situational requirements in which the question was embedded.
The second step then in the question-answer process was the assimilation
of the question-stimulus to the situation context of the checking of the
initially aroused personal hypothesis against the requirement of the
situation. The third step was the formulation of a response to the ques-
tion that would have facilitated or at least not threatened the respondent's

adjustment of personal needs to situational demands.

 

 

P. L It - , 1L1 1 113
)1-) i - ll ...... - 51.11 11-: 1- - -t- 1.0.99.7... )1- may 0... ,1- V

 

 

 

50

There was no reason to assume that the mere act of asking
an individual a question obligated him to divest himself of his self
interest and needs for ego-enhancement, interpersonal acceptance,
and conformity to group norms.

The practiced clinician, psychiatrist, and personnel worker
has learned that what an interviewee said about himself or about a con-
troversial issue could be as much motivated by needs other than giving
an accurate account of his feelings as by those for giving such an aCcount.
In a clash between accuracy of verbal report and self-interest adjustment,
the former would by no means inevitably have been the victor. Exposing
the personal hypothesis tripped off by the stimulus -question could have
aided or hindered the subject's adjustment to situational requirements.
The internal response was likely to be defended and censored. If certain
kinds of perceptual reactions could be explained as distortion of informa-
tion-stimuli by personal hypothesis, certain kinds of responses to ques-
tions Could equally have been described as distortions of personal
hypotheses by situational stimuli.

The personal hypothesis tripped off by the question--what I
are the changes you perceived during the past year--in the present
research could have been, management is trying to find out something,
what is it ?, what do they expect me to say? Not knowing the interviewer,
not knowing if the interviewer would report who said what to manage-

ment, and knowing that the company recently had a large layoff, these

 

 

 

51

men responded in a way they thought was expected of them. In this
way they would not be threatened--no trouble with management.

It was the experimenter's opinion that the nature of the
content of the question could have easily provoked the controversial issue
of the serious layoff which took place about the same time as the
interviews. It could have aroused strong motivation on the part of
the subjects to avoid any trouble for fear of reprisal, and this strong
motivation overrode and hid any effect the content of the previously
asked questions had.

Thus, a question may have not been responded to in a
vacuum, but rather tended to be assimilated to the situation of the
question and the questioner. The nature of the response may have
been a function of not only the content of the question but also the
situational stimuli. Perhaps, if the layoff had not been so near and
so threatening to many of the subjects, the standard question would
not have aroused uneasiness as noticed by the experimenter, and the
effect of the content of the previously asked question might have been
shown.

In conclusion, the experimenter wishes to express the
opinion that the preceding highly theoretical explanation was only one
of a number of possible explanations of the behavior displayed in this
study. The behavior may have been simply due to the fact that the
information called for by the standard question required the subjects

to recall events of the past year while the rest of the questions in the

 

 

 

 

 

 

52

interviews required the subjects to think of on-going processes. This
shift in frame of reference might have caused the apparent shift in set

noticed by the experimenter.

 

 

 

 

I
*“N: - .. ':'_:l-:'

SUMMAR Y AND CONC LUSIONS

Although some of the variables that affect responses in
questionnaires have been from time to time theoretically defined and
discussed, relatively seldom has the effect of these same variables
been empirically investigated in an interview situation. The present
research, carried out in an industrial research organization, was
designed to explore the effect of one such variable, namely, the
generality-specificity of the content of questions asked.

It was the intention of this research to determine the ex-
tent to which the content of a response was a function of the content of
previously asked questions, and the following hypothesis was made
concerning this relationship.

The content of previously asked questions would establish a
mental response set in the responding individual, and this men-
tal response set would influence the subject's response to a
subsequent question in such a way as to reflect the nature of the
content of the previously asked questions.

Furthermore, it was of interest to test, in an interview
situation, Parten's assertion concerning the order in which general
and specific questions should be asked. She stated that if both general

and specific questions were to be used in the same interview schedule,

the general ones should precede the specific ones, instead of the

53

 

 

 

 

 

54

opposite; otherwise, subjects tended to answer the general ones in
terms of the earlier specific ones.

To test the basic hypothesis and Parten's assertion, two
interview schedules were available for use. The two types of inter-
view schedules were semi-structured but differed in content. These
interviews consisted of a series of specific questions. While the
questions clearly indicated the idea or procedure to be discussed in
the answer, the subjects had complete freedom in the content and
manner of their response to each question.

One of the two interview schedules, the management inter-
view, contained six major questions intended to reflect the subject's
knowledge of company procedures. The questions in this interview
schedule were general in the sense that they asked for information
about the management and accounting control procedures which were
known to at least some extent by everyone interviewed. The second
interview schedule, the project interview, contained ten major ques-
tions intended to reflect the subject's understanding of a particular
technical project or work area. The information asked for in this
interview was specific to a relatively restricted part of the company;
such information would be known best by those working on the par-
ticular project or work area. The generality-specificity of the con-
tent of the questions made up the independent variable in this study.

The content of the response to the standard question constituted the

 

 

 

 

55

dependent variable. This question simply asked the subject to list the
changes perceived as instituted in the organization during the past
year.

Seven interviewers interviewed approximately half the
fifty-five engineers with only the management interview and approxi-
mately half the subjects with only the project interview.

A code book was developed to reflect the content of the
responses to the standard question. Two raters independently judged
the generality-specificity of the categories in the code book. The
reliability of judgments between the two raters was obtained and
disagreements ironed out in discussion allowing for a common agree-
ment as to which categories in the code book were general and which
were specific.

The content of the responses to the standard question was
then analyzed, and frequency counts were obtained of the number of
general and specific units of unformation each response contained.

It was hypothesized that the questions in the management
interview would create a general mental response set in the individuals,
and this general response set would influence the subject's response to
the standard question in such a way that its content also would be general
in nature, that is, the content of the response would not be delineated
and would refer to the whole organization. On the other hand, the ques-
tions in the project interview would create a specific mental response

set, and would therefore, influence the subject's response to the standard

 

 

 

56

question in such a way as to reflect this specific mental response set,
i.e. , the content of the response would be rather detailed and refer
to a particular job and not to the entire organization.

If the hypothesis were true, there should have been a
statistically significantly greater proportion of general categories
counted in the responses to the standard question for the management
interview than counted for the project interview, and, conversely,
there should have been a statistically significantly greater proportion
of specific categories counted in the responsesto the standard ques-
tion for the project interview than counted for the management interview.

The results of attempts to find inter—rater reliability indi-
cated that such reliability was relatively low. A series of 't' tests
were computed on the data obtained for each coder independently in
order to test the hypothesis. These results did not support the hypo-
thesis. These results, however, did refute Parten's assertion, in an
interview situation, that is, having specific questions precede a general
question did not induce the subjects to answer the general questions in
terms of the earlier specific questions.

Although the findings concerning the basic hypothesis were
essentially negative, the possibility of a relationship between the con-
tent of the question and the content of the response was to be indicated
by a trend in the data. The experimenter suggested that, perhaps,
refinements in the procedure used--as mentioned in the discussion—-

may help clarify the present uncertain relationship.

 

 

 

 

REFERENCES

Atkinson, J. W. Motives in Fantasy, Action, and Society: A Method
of Assessment and StudJ. Princeton, N. J.: Van Nos-
trand, 1958, 238.

 

Berelson, B. Content Analysis in Communications Research. Glencoe,
111.: Free Press, 1952, 24.

Cronbach, L. I. "Response sets and test validity." Educ.psychol. Measmt,
1946, _(_>_, 475-494.

 

Fries, C. C. The Structure of English. N. Y.: Harcourt and Brace,
1952, 124-127.

 

Getzels, J. W. "The question-answer process." Publ. Opin.quart.,
1954, E, 80-91.

Guilford, J. P. Fundamental Statistics in Psychology and Education.
N. Y.: Mc Graw Hill Book Co., 1956, 335.

Link, H. C. "The psychological barometer of public attitude. " J. appl.
Psychol., 1947, a, 129-138.

Maier, N. R. F., Hoffman, L. R., Hooven, J. J., fRead, W. H.
"Superior subordinate communication in management. "
Amer. Mgnt. assn., Res. Stud. 1961, 5_2_, 12.

 

Marsden, G. "Content analysis studies of therapeutic interviews. "

Psychol. Bull. 1965, 6_3, 298—322.

Orne, M. T. "On the social psychology of the psychological experi-
ment." Amer. Psychologist, 1962, 1_7, 776-783.

Parten, M. Surveys, Polls and Samples. New York: Harper and
Bros., 1950, 213-216.

Rorer, L. G. "The great response style myth." Psychol. Bull.,
1965, 63, 129-157.

57

 

 

 

 

58

Selltiz, C. Jahoda, M., Deutsch M., $Cook, S. W. [Research Methods
in Social Relations. N. Y.: Holt, Rinehart, Winston, 1964,
572.

 

 

Wickert, F. R. Interim Report, Communication Research Project,
Unpublished Manuscript, 1965.

Wispe, L. G., and Thayer, P. W. "Some methodological problems in
the analysis of the unstructured interview. " Pub. Opin.
quart., 1954, _1_8_, 223-227.

 

 

 

 

 

. .II. ‘5...

v. I 1.1 . .I‘v )
Lia-in. - R41. .111 I 1 1
u 1 (

 

 

- .
1 M1 - . . . it). .31 1- - - . 1 - -
.. . 1.0.1....- $1.11 1- whhl.-twm.n1.w:uwu - .51111 1.5-1. - l i

 

L Si Project ‘ January 1;, 1965

Add to all Management and Project Interview Guides this question
(this question may be omitted only if the interviewee had been previously
interviewed on a management or project interview during the 1965 round
of interviews) (this question will be number 6 for management interviews
and 10 for project interviews):

 

6. or 10. Think back on how things were in the Engineering Division early
last Spring - about March of 1961;. Now, what do you see as the important,
the major changes in the administration and management - and—The general
climate - of the Division since then - the few really big changes? Please
try to list these. (Instructions to the interviewer: do your best to
keep the interviewee from reminiscing and rambling; try to limit him to
listing the truly major changes he observed.)

 

 

 

 

 

 

 

January 13, 1965'

Peyc‘. 0103'; C67

1963’ ljr‘.'.1.2:"e,ent Interview IntreC'. i'ct'f on

(go's: I. .1. :f'il‘DiiiCC 33.:"='.‘rcel[‘. Then 563': I I2'10101‘e to t 11: Iri'l'.‘.-13;ou today
out tuc . 0.12.113 control : .d coordinati 0.1 53's ten - the co.. :..'.1’-_cation
system - 0.." Lute .131‘."incerinr3; DiVisim and close Cl3,r related '3): rte of ‘11s COLDU.‘
here in Grand ;;apids. You 11.13; have mean t'-:. 01173.. '1 there sate catestions when
we were 12 before: whether 3,'ou have moon or Int, we are asking the same

Questio 1.. of a 1e'):-‘esen'tativc say-3.10 of meagre-stunt people (not as a check on

;'3':‘ co: sisoe;.c .3-' ‘.mt)-::- a. are; of "tak'n3; the teqporature" 0'2‘.‘ the co..r.'1unication
s;.te::2 .. ere after a gooc. 5deal of tine has elapsed and quite a few changes have

taken 913 CO .

(new) I. B. "Type of dopz:r"':~:c nt"'5)cc_>:':cs "*3,ij of 'or: anization". Few (1965')
use these c..tergeries: (1) Vice resident, ...r1..e=-ri. ,(2) Technol r r;

(3) mod-.1603 (‘L-) Seivices ;(f§,- 3.3:. . .

azese 0113' if 3'ou fin that it is desirable to dose.

 

 

 

i.
"'

 

2 . . .' g V.- u. " 1 .' .. ‘1 r z. , . .', ,...,.
.1’1 - a ’ -"- I. .. “7.? 31.33%, Axe-J
.~. y' a ~‘. w~1 r tw{.~,' 1. .‘ ,~~-» ~ ,.. -. i 13",“? -\',V~'3 "".“,
-. ~~ I - - - .3 f It 0 . - . Q . tl . 3‘ o " #AI 0 h , A
. I! .. - -' t. ‘ 3%.“ .. 3 . ..- -'1 .Q' 3. a ‘2 "’1 -.A1‘-"- Au-H‘ ix rvvr 7 n I
\ti‘r bu..L\J ‘3 ‘v k‘ a 94.4 . . ...-. - .4 3 I . ,.L (16‘ - 1...‘. \ 1'55..‘¢. (.n I). up a.
I
9 ”3.3.4 3.-- ' -3. . .1 ....3 ..’.’ r1'?' w, .. ‘1 -.. a , ,.. ,.'.,.-l..' ......, D ‘t‘ f‘-.‘...‘,,
L. I . u . .: 1 ‘. . ' .u I
~C‘d J .bl‘. ”fifi"¢h .'-’ ula 4..‘-.-, V a .I 3 '. .‘-‘ UL.‘.I""~ '-'\'L "-|) Q. ~ -'.¢ "-,.‘._-an a .

1.1- . L'. ...... -r- u. 1. N..- «w.-. 9” -
1.3.0:? .. .10.}! ‘ a...‘ inn/3""L‘ " Q. --"‘l 1.31 t.” "4 “1’73

leIsfen or “a fut 5L fii“ltlon .xv. '3, ":14“t"ent axfierwr
p.21 ;_2, 3‘1"), :. 3"...0'3 1.‘ .7 '1? (,1."'?;'./1‘.‘.=‘. 33.:- -‘.-:‘ $93.15;"). 3".“2'51Ct.
..1?'“‘ ~rL 'dt {iifigfizte 0:1: gViTe ‘ “A 1"fit (U’SY‘7Y
), sy..:; g:';ci wit, C?i77fl“ht or ?fir3“-ré dhwsrtfioit.

31.1.": '. “ St) 0

,4

‘

' “Lo r“.\,>q 3~I~J .. p ..| Jn- Io¢< - . 1.- ' I 3‘, ‘ I ..-
‘1‘} “.1". ‘..‘3 4.54." '.".1,f~q-.13‘r ‘-~.‘, .‘4 Cw}. ”Lin 1:] 1.0 Q- ..3‘. :"')J 30.0

' ‘ ' ~' 1 3 .. A \so.\nn , n a - u r . O ' ‘1' '
mien “.0ifi vita 110 Var1eze Cu...‘lk,1031 et¢,3fi taro; H v.10a

4'“ l , - .- r‘ 1 ,‘. .5 a . u A; ,.J I'. o. v .1. . 3 .
5&3 it (1 ‘J"! C" 3. «31"..21..‘..’3. 213‘," W"! ~’,'.‘y 1’390 t-‘1'33 if}? 33r'37v33. .3792”?

. 1
I
3.,
M}
c 7'
Pa
0
V .
.- I
(,3.
‘4
3,.
x J
*C
.33;
-
2
(it
,. o
i.
F‘
.4
H
to:
5*
H
1"
P“
L.
t“)
a:
“I
‘L'
I
’1
.1
J
\fl
'—
A
'1
’9
r:-
.1.)
":3
‘
I
4
Q.
Q
I ‘0
'1
I
p o

‘1 11 .- I .‘- f‘ ‘ .' d\ -
i “we!" “:2: QLF’E Q‘s a ‘05; J 1".

A H ". ‘ 1 -,. ' r l J .. 3 , '13.. ' ',,~ .‘3. 3, W' ‘E. '31 _ , 3 .
"us/t1 T. 4553 18 t0 OJEELPL (.fit‘. in vi‘hc '2 ~r111'j «1’31? " «'1‘ ‘211‘

'- -‘.u'. -‘ \‘. ta ' h. ’- "3,- .v‘ w. .o- ‘ n ‘ 3'!"
3"} “4LV'A.::4-Qd 3 C")¥i-H‘~lHlC-‘.VLC."5 .“ _ .‘.s"‘.'.‘ 3.2.x 3.435.251?) 337 L‘V'T'LL'? ".

. .,. 1 v ' r _.\ H. -1. ‘ ‘ — . n as.“ » c— . ~ ‘
.é. )‘fflwfnL. 3 0 0"3 .JZLL-i"! {LL-4'15: b3. .2: . 333.} over he? Fol-P1P

U J

'..

.‘ to ’ C“‘ ' I. ‘ V m '- ‘pi- e: » .fi ‘ ' 1'" o ' u - y- ‘ :11 v- -.-‘ k‘ ‘
a «J to ‘t‘»1'1‘4‘IU-Lr‘ttg’ .93. 4.! 3"‘4 J“ 1.1.‘9113.L, ‘5‘! An arr-J, b3"oru " Tr“rk‘t1“n

u
.‘v..‘.”' ‘
0"“. Len. :
-. - C 2 . ‘ . ' . C . o .1 ’
0'- 3, r ""‘|. . .. a» «v .q‘ 93‘ , - ,‘ ... . 3 u ‘
-J.In: { ‘..J :3) X“ ‘2ul.1 CL..‘ ’A ‘ l .n. 13) - p‘l,’. rlfl": r3?‘)‘!q‘:£ {:3 '0-4 A
~ .' 1 . ‘ A q '5 ' I r " 3‘ 4| 8
~ ‘ ‘-. D - ,Q . ' ~ ’ o \I' I- V‘
.J t I nth-i-“ / rl‘b ;\ J “v. 74.. t, n. J u v x. ‘1'3 OJOL‘Qk t J ‘3

 

 

 

 

+_'_ 12;. 5:: yam rvc-nnt racy ’2Z‘il"’,".":‘ in t‘m‘ ii!’,';.’ in", foster 8773
rrru;r$ + rta ouwr,rwun hf r‘m vrn‘rctn grfi pr) rufis. (Th6
int-r=;nc.s nay uccfl :0 ¥n 1:; 1 to thin‘ in n-itlve tmrvs at
:.13 “miwt; ha 71; ”jri in 51‘? in t.7k zbawt obztnclqa. He
may ha tcld Luz. that tn Le Ct‘vs r~ t.) 1.2 Text, nastinr the
chrtzéKwa th.t yr???“t pas i“1/ oi iiwxa '3? nAv profacLs 111
gr; fi’? fr:: reankn thw aLt ”L on o? arl cjxvlrcing tnnra
a “rw in a 7‘=Ltiwn to r. J Law f?”'?3‘ 41cm}. Tlexua list
’5‘ L'. I. '
2. flow .3 iivu Baa oven for'u1 tut, CFO may sf inw“lng at n21t
nap wna anxt 15 Lo :77 L531 3 10;j “r31 n? tiln‘ini anfl plafinlng
goes in: ostaaiizhlnq m ”01 bra} ct ¢f_;r0jram. So one staminrd
p;L:¢rn is folioxtd; tan mac. d. wr&s an Ci"' '¢iutfifi3o Earn 18.
d. quwstion 7:? t9!) 19’2”: mlml 1’) 53436 your i ”"z‘i'x: ions in £313
brnad ar:u:
2.1 Tho OCR
2.11 Ehat is an €73 “or? srncificaily. uhafi has the ’73
:ngnt in profiacts with which yfiu are E filliar? (Tha fbllouing
u
qx~stlons may he naud 1! thn a» w~r$ to th9m do not coma cut in
an warin; 2.11. Cthcrwisa Ld%y :3; he cmitLu%.) 2.12 To what ‘
extent docs an SR3 in ynur experiancu caver ;11 rw1evant 1mes?
2.13 To what mxteut .ru thn suatwnrnts a“! figurna quoted in an
35& of valze to the ;rnjcct an;;n~ur as ignnfi :9 work an thg SCR?
6....“
——*“”"”’F”4.3 . Cane a new projfict or pfOfiram has been establluh‘d, one way to
Izwk at khat hoquws nnxt 13 thgt : Jul 4?: o' it work ‘nclns.
,*rz ar , brpfl.1;”, L’Lfl'. t o KZ'Z tH.L ran ‘Tfit hug a?
cvn,::)lla n1 (liiLfl‘ an} cgoriinwnxg: v.Jh fira'zet:

/ n

) C35Lr711iNZ thm Frfijict 35 a projact Hui v

) Duntf011133 the

 

 

 

:vrj fit Ln r~“,1;~n to ct: ? rrn“vtu «3 , r. a” f vnt'vit; of
:u gr 3; .LL 1 M1‘yin ?u‘ ? .h~“”;“j ‘iVi: A. ‘Oilnsln ave
ue~rvv 1 \}::-.;.:.;::;'..'; l... ..'-r ?‘ ;. , f'e mm ‘L 'rr9.1r:zi:e:'1a in bath mf'
thaw Sl‘md avast:
3.1:1‘:3‘.I'.'.:1‘1 "L : :rv-‘w .48 .1. pmj'ict. 3.11 Elwin?- ni‘ it'tz'rt,
V435 Ln) T‘A is u.f~tt;va in .a.g.wr the Vrv'v:L. n.1nfinr to con-
trnl tha pro; at c a "re ct. List h:: .w'"s.

3.12 'ncn 107 ccqt r“ “N"? $11.3 15 «:‘finiully izportant,
71-1. 12:. .5 r. 7‘"":':‘” .. 1:. :Y‘ C": ._' To in) It'vrm (hurt: if .3" ‘ion
:1;t: an JLJ *rnj at. n: g @ a’d*" thv.t affix: tho pwrt~r;, or
at. r r': fic”4 h. :i_hz r"‘1 tfl for kgdq.
3.? ii"; .. WI... "'w prey-Cu in mlz‘ir‘r‘. 1.3 c- :v‘: 112‘) how. ."‘.1
ER” ...31 yrl*23 aflnut .3; ‘Tdthly ‘1~L::‘7;n‘ :372045t ixLWQ, .
crunsxicr‘ 5.721;, 1.4:: 'i‘wxwacart” OTC”? it 3 .3 5:912 ‘3: to ‘QL'L‘3, ‘tells
the uslal? G'T't ‘i'lnwar. (To a. war fig u"; ctéd to 3.2; 50 on
to 3.21.) 3.21 103;“ llrt tie hwrfi?itr a 3. ‘"ct enginner gets
fro. ha H2rth}y 'n.in:arin° 50*wnast far a profact to u31ch he in
3:;ignni. (If yuu are not a prcjvct awgimmcr y ursclr. thin of
505+ pro’:ct englnnfir with Hint} uo*? gnu av; mast ?1w11i...
3..2 lease‘ ist .a bnutfits a rrafnct 6351 near gets from Lbs

fsrncaat picture (Th3 fianthly

an a gawk? r of projects).

3.23 low thin: in L3~xs a? the

you are nut a dapurt”tnvt nanajar ywu
..H. nr with u‘t é wzrk you 35‘ 'ngt
”Tim Lac :“ “:5. '1.- 1"; an” -1;~'~
‘P'c:..‘.,;t-:;.L (111:) ".31.‘..)int‘1m” L.“ ‘J 1"

Traginfivring Voracast that is

Havar

~11 L .., - . -is‘w‘ Luv)
17:“ ‘ «3‘37" ‘1“L“-’-Y:t‘i""
“I‘M: J: ' 7‘1").

 

:r' inmrviou. {it-Lrwlauar'u my!» an‘. 1. --~

 

‘ ' '., _,. .... ' .. . -. L. -. 4“,. I
3.3 XLJ'V'JL Y.:.J 312.! '>.". ‘ . 1 _ . ' .. .- ... wring -'.s‘v‘.L.‘.iOfi

. ._»4 ~ . . .~, .1. x . . ., ‘. ‘7 r: "J‘ .’ . .-' ‘3
brhiuv .‘uJLt -. .rl .. 3] \..;j . .‘ . . ‘ . . .' .1. 1:- . ‘L ..' \ UHln‘l‘fl'lng
_ ' .: »A. . ' .. ' .‘. ..r . I“. - » »

$23.. 23.. ...7 u.-- .. i... .) _3. ..: '. .....,r z p“: dypr‘oxi'ithd

- ‘ v ’ II ‘ - “ .. - . -..y I .n -. ‘ I'~ '3
. 3hr. ‘ .3 t .J u.‘:..’.‘. .10.”... .3 w .7 .3 :3. .mngr) r313... .31}.
, .l ,_ \ _ ‘
. M... t... (\ :n't 1131mm t. 7 :
«"l.

 

   

C 'I t. J . '3.)

1. 3*.)_" vi. .1..." "‘2 234'; (“0 V“ at. Lr'.’ 'J 7-, '.‘.. 1 cr~:“r.'.':'

1-14;! 13.31 c m Tl‘t-‘i. .0 :3. .LL .5.

. ctr-.2 :11 r 4." :, <" ‘ r3 ' 'L :3 V‘tWfl" .,'”o

13- -. . 5‘. .. -. at. c 1“. Lu» 3;: >‘:.: '. 7 '- 1" “mfrf‘ts
and 151‘.) 1717;). -l.{;;;o 214nm to (*0 3.11.53. t3.» runny-m m1". of (1) 'V‘.‘ “’10
is i:.‘.;‘.'.'_7’.-::;‘..; :.;.‘; a: an} !-'<:'..’.v.37r. :\.‘ E(. ”)0 ~- .,.'.".'»‘-.‘.‘.q.‘.::.
1‘011“:;ri:'._: :.:3: (35'- 7'1 :13 (‘3 r; '.'--.»'.'z ';.\':'3. a; 5.. . 3. .s a“ 111.3.) '1.-:":‘_3.-’;,
:"..'.'.;..; ‘ 5.1 :..’.;.1, LT» ; ~-7 ,2 1’; cu 2.5190

   

absurmd £2.31. .33. irrl'vi'! 1.1'3 v:v"< 31.1 '.":'~:.t..3 um Pant :. 35.34435

‘. : - . v.‘ :. ." - _,. nu . ,.. .fi .
to ma std-21;}. (1....54.10.1210...) .-..i int-vain 5.2 I. you we a

, ‘.. ..
n.2,

 

Fix? M. .1. ‘~".‘£‘ (or 1“ ymx 9.2“ nut, LUZ/.74 in mms 0.4." thx:

.
V .-. .. . .. -\ . ... .. .. n. .‘W .' .
(japan; . 1.2. “,a.‘ t: m u ..:. do; an: met. .- ..¢:i—-{‘). co
I

 

hwy: -,-- vnu (13.9 ): (1) would 113:? to {‘331'92‘3 to
(usable for yummalf‘ (him IE) but w 10‘: at rm’imnt you (“I") has

to r. ”12' up 1. .1. the for Li doc‘m to 23; .321-3 (2) sMWfifl‘

L‘. ; .3 11m far :1 63.3136 rmxn‘vvv- than m M .“vy dvmiaien wwrmlf

 

: .- ~.~

3a.). 5.21 Ion-z. 1 31’. the mu-um taut earn-Jr to you 3mimr
-.‘:~cvo. 5.22 (acct, 1m t'n ..=‘.‘..3.‘.‘J 2‘1", ~‘.".:'.;!' to you. man”?

'11..) ‘1‘}3‘1'3.

 

3‘.‘ t‘m intrrvi'w).

 

 

 

 

 

January 13, 1965

Pathology 8-”)?

1965? Project Interview Introduction

.222); .-.. __;2troc’c.2ctio:2. Introduce your: 13, Then M2-

( ”L‘u :
I an t:.l]:i2 .::Ltl2 you todcy because of your association J t‘2 I1 Toject
(that p: 3;} cl; it is you 31122362 12 vc found out from the list in 21 . Ibfiin; cr's
office). (If an "old" Arno‘jc -ct (a >1‘ogoct 0: 152312 ;' in 19611,) sq, " '021 22a 22013
d‘scussed this same pro 'ect with us so .0 ti. :e 5250. ..’c're faint; over the 55.2.20
quostio: s c.‘;:in becaus e thcre 12‘s boon a m: t ciorl of change since we were

J.

‘ne~e beiore. '..'e are still co 100.. 2eci 2rit‘2. commnications, and talking with you
I. lps no ‘-: out now commnications take pl: 0 here." If t1.i: is a "new" (0.1965
ojcct) p: o j‘occ, you may need only aorta of t‘. :is introduction - or even none

2 41

010120) T zen go on to part I. 23. in the old guide.

 

 

II.

 

‘

Interview Guide - Projocts-ter*n—ou1-3t_g-tine Pusstions

A. Introduction. Introduce yourself. Ihon say: We are talking with

you today (or this time) because of your association vith Project .

 

2. Idcntification information: flame of the interviewee, level in the

 

Division (designate one: division or assistant division manager,
dopartt4nt manager (indicate line or staff). section monaner (indicate line
or stu€f), project enclneor). typo of department (desivnate one: above
department (usually division staff). system department, component or hard-
ware department. service department).

C. Role in project: (designate one: project leader, immodiate supervisor
of irojwct louder, an assigned member of the project, a "cognizant" person
relative to the projoct).

Q

Intarviow proper. The interviewer says: We should like to ask you some
rather systematic questions about this project and its conduct and control
as part of our study of the communications system of the Engineering
Division.

Your answerSto questions will not be identified in any reports
preyirod either for publication or general distribution. Instead, your
answers will be combined with those of others. Reports will therefore
take the form of statistical summaries.

1. that is your idea of the purpose of thisiéréjgctl 1.1 from the point
of view of the company, the division, and the department. 1.2 from
a professional. scientific point of View?

2. To the best of your kntrloige, how did this projoct got started? Where
did the ideas come from that contributed to its dcv lopnent to date?
What do you see as the ooyinnoring, custo:er, marketing and management

climate that brought about this project?

 

 

 

3.

4.

5.

6

7.

8

9.

 

~2-

:.l aunt do you sow as your rolo with rcsrpct to this project?
r3.2“flbout-vhatfproportiofi'6?“?ourrworkin3:day.havc,you.spent‘onrthis
QTOf‘ct firth: the p3:t*nonth?~~3.3 “fihout what Proportion do»you
—expaot—to—speni«durinz~tha~comin3~month? 3.u what other persons or
groups are involved in this project? 3.5 Fhat are their respective
roles?

h.1 What in your opinion is counany interest as well as division
interest in this project? h.2 In what concrete, specific ways have
thoy boon showinc thoir interest in this project? “.3 What is
custohwr intornst in this project, as you soe it -- what is its
nurknting potontial? a.“ how well do company mama: cont and customers
agree on the projéct's potential. do you think?

5.1 Hhat in your opinion was the most significant technical broak-
through on this project to date? 5.2 Why was this breakthrough so
significant?

6.1 What, if any. is now the most important romaining technical
,obstacle or block on this project? Please explain. 6.2 What in
your opinion is needed to overcome this obstacle or block?

7.1 How does this projnct stand with regard to its established
schedule? 7.2 How do you account for this?

8.1 How do you think the technical progress on this project compares
with the dollars spent to date on it? (Make your judgment in the
light of the various projects with which you are familiar.) 8.2 How
do you account for this?

Lfliitional remarks ani observations not falling under any of the above.

III. Gloss too interview. (Interviewor's name and the date of the interview).

12-19-63

 

 

 

 

       

.

.QI.‘,0\¥..1!... [IDS]...
1...? v. 4.. .

‘1 . 7 u 1.3 u.

v .3. . -V- -1

 

     

1.-
.L- ....

 

 

91.7....71 .t H-..KMED..M.§..W.W.¢W..F....WK... ., . ...1 -......H!.r.._x.. .3!

l.-. ...3........ unfit. ,-3....3..m-3..m..n.-3[3.$.53..i.r..... . «M.,?

 

' A,

.\. n .1 -. I 2.3 .. ..3 3.. | ... . . 1.... .. . . U

.lll‘. 1“... , , .r ’ a .{u‘ .\. 1V4, .3E.|3 . r 3. .v . . .. .Q.1..\«I..dm:.v mt,r..333|r 3 . 4 . .. 33.3 . L . ..

33 3 33. . - .. .1.\ h3N 4.3... on. ......3.3.1.o.-.|3... .3L.3.3..H.IM...3\3. . . 3.".- .. . 1.... yo. {3131.3 . 3.
3. 3344.-. 3 . 34:1... 33 33: 3. II . . 3 I-3-3.IV3.333.. .|.I 3 . v 3

 

 

Code Book

Management (Q. 6) and Project (Q. 10) Interviews
1965 Series
Q. 6 or 10. The major changes in the administration and management--
and the general climate-~of the Division between about March 1964
and the present (mid-January to early March l965)--an attempt to list

the few really big changes.

Soocial instructiong to thg coder: The interviewees. you will

 

find. tended to ramble and hardly limited themselves to the few big
changes. as asked for above. Moreover, their remarks not only range ”“
from very specific to very general but often include a causality or
time dimension that makes coding especially difficult. For example,
regarding the causality or time dimension. they will talk about an
event or influence that led to another event, to another. etc. Your
task in coding is going to be to concentrate on the events that hap-
pened and not on causes or effects. even though the events are spaced
out in time by the interviewee. You are to concern yourself with
coding events-that-happened, regardless of their sequence. It is
realized that we will lose some information through adopting this
policy. However, any coding is a simplifying process in which some
information is lost. we are mainly concerned, then, that you look
for events-that-happened.

More than in coding other questions, you are going to have to
consider the context in which remarks were made and use considerable
ingenuity in working out in which one of the limited number of categories

given below you think a remark should be coded. We must take some

 

Code Book Q. 6 or 10 -2— April 28.41965

chances here on reliability. You are asked to think a remark through
carefully. reinterpret the words somewhat if necessary, and come out
with coding the remark in that category or categories it best fits.
Several categories may be needed to reflect the eVents expressed even
in one short but information-heavy sentence.

The purpose of these special instructions is to help get reliable.
usable coding accomplished despite the wide-ranging answers to this
very unstructured question.

Another ggution: the coder should pay special attention to the
generality or specificity of the interviewee's statements that he is
coding because there is a plan to use the degree of generality or spec-
ificity as an experimental variable on this question. If an inter-
viewee makes a vague, general, global reference. his reference should
be coded in a general category; if. on the other hand, he makes a
specific statement, it should be coded as specific. If, however, he
starts off. for example, by making a specific statement and then says
that this is Just an instance of something broader, or if he makes a
broad statement and then gives specific examples, both general and
specific codes should be used.

The coder must code at least one code under 1. and one under 2.

 

Code Book (Q. 6 or 10) ‘3' April 28, 1965

Change(s) in the (usually economic or business) climate or en-
vironment in which the company exists and/or in the engineering
Division. (Changes coded under 1. should be differentiated
from tnose coded under 2. Whereas 1.-type changes are changes
in the surrounding climate, 2.-type changes are instituted
changes--man has intervened and done something or taken some
action.)

1.1 The climate has improved (March 1965 is better than March
196“).

E; 1.11 The work load is up, the company is in a growth phase.
the future looks better, there is more stability, more
backlog, there was a market swing (in our favor). the
pro-election political stall is over, etc.

€5 1.12 Money is easier to get, there is more R. & D. money.
now a manager can find funds to allocate as needs arise,
etc.

:3 1.13 The company is more aggressive and confident, is going
out after more business, puts more emphasis on selling,
has become a sharper competitor. etc.

f{>1.14 Others

1.2 There has been little or no change in climate (although some
say that there has been).

1.3 No comments on climate.

Change(s) in how the company and/or the Engineering Division
functions; instituted rather than climate changes, or changes
that Just happened.

p 2.1 Beneficial changes

#3. 2.11 Management-instituted beneficial changes
ibfi 2.111 Beneficial changes in formal organization

2.1111 Fewer layers of management; fewer persons

reporting to the Vice-President for

9; engineering, the Vice-Bresiuent for
Lngineering is easier to get to.

 

 

 

 

Code book (Q. 6 or 10) -h- April 28, 1965

£1 2.1112 The three divisions

9) 2.1113 Engineering administration is cen-
tralized (with professional accountant
heading it up).

5; 2.1114 Functional and departmental reorgani-
zations: departments with like fUnc-
tions are better grouped; more project-
oriented task groups and less emphasis
on departments; preliminary technology
is removed from departments and is now
in its own special area; smaller units
have been merged into larger ones (e.g.,
A.G.n. into the engineering Division).

:3 2.112 Beneficial but less formal. more general organ—
ization changes: better definiton of goals,
the company is run less like a product line
and more like a business, more attention is
paid to engineering, more centralized control
with work loads better estimated, engineering
is no longer such a step-child, more engineer- '
ing-mindedness. engineers more trusted and
given more say-so. more cost-mindedness.

f; 2.113 Better administration of dollars, more realistic
budgeting.

/t3 2.114 Better personnel assignments; technical person-
nel are assigned to technical areas.

/fl3 2.115 The new president-~some reasons why he is bet--
ter and some excuses for the immediate past
president.

#3 2.110 Others
7t3 2.12 Changes that just happened--sometimes possibly or in
part traceable to management actions.

‘ ,El 2.121 Greater efficiency: less redundancy, less
duplication. less paper work

76A 2.122 ImprOVed inter-organizational relationships;
less friction.

2.1221 Better relations between Engineering
and Marketing, Manufacturing, and other
6; important company organizations.

 

 

 

Code Book (Q. 6 or 10) -5- April 28, 1965

A

:3 2.1222 The newly-instituted. lower-level
meetings on the OCh's

53 2.1223 Master-projects control
5 2. 122a Others

2.123 Better teamwork: the company and/or the
dngineering Division operates more as a single
unit rather than as a series of aspirate units,
is closer knit (after the layoffs).

2.124 Better quality personnel: technical proficiency
is improved and have better engineers now; only
better quality personnel is left after the lay-
off; much deadwood was cleaned out in the lay-
off.

2.125 Better morale among company and/or Engineering
Division personnel. (Morale is better than
last year when it was Very low.) Greater se-
curity after the layoff (since only good quality
personnel remain). '

9 2.126 Can spend more time with customers

5 2.127 MP1.“ [Au-~{{-(m’u7.nzm,’punk.)

2. MW 0’5”

Little or no change took place, or, if changes did take
place. they have had little or no effect.

2.21

2.22

2.25

2.26

Efficiency has remained about the same.

There have been many changes (from general to specific
and any mixture), but they have had no effect.

There has been no change in morale; it is Just the
same-~this. despite all the stories and rumors.

There has been no essential change in cost conscious-
ness or management mindednese.

There have been few changes; if there were any, I don't
know what they are; I really don't know what's going
on; if there were changes, I wasn't affected; there
were no real changes (in working conditions or com-

pany policy).

Other

h

 

 

 

Code nook (Q. 6 or 10) -e- April 28, 1965

/Ek 2.3 Changes for the worse that have occurred.
[is 2.31 The company is generally less efficient now.
.i3‘ 2.32 More detailed changes that have hurt.
6 2. 321 The layoff

.5; 2.3211 We are left short-handed; we lost good
men that we can't realace; we lost good
men in the manpower reduction; it is
hard to get new prospects to come to
Grand Rapids; there isn‘t enough money
for raises for the present staff since
the company has to pay more to new man
to get them to come here.

5;) 2.3212 Morale is a problem, we lost faith in
management, we lost security, we are
more insecure, employee interest is low,
we can't look forward to retirement,
layoffs were badly handled (they told
us they were over and tnen would lay-
off more the next day; we weren't told
why--the layoffs just kept coming.)

53 2.322 Organizational structure changes; have moved
from technical to procedural mindedness; there
is more formality in the Engineering Division;

management is less available now; management
is less decisive now.

S 2.323 There is less cost consciousness now.
/C5- 2.324 Undesirable personnel policy changes
5; 2.3241 Forced Vacation times are not liked

{3 2.3242 Lack of money for pay increases (no
reference to the layoff)

S 2. 3243 0 there

[A 2.325 Others

#21 254' Changes as such were mentioned but with no comment on their
good or bad effects on the company or on morale.

§> 2.41 The new building next door-—the accountants are closer.

f) 2.42 Personnel policy changes. e.g., the LEAP program was
discontinued

 

w" ~_,

4. _-~'.-“'

 

 

 

Code Book (Q. 6 or 10) -7- April 28, 1965

3 2.1+3
[1 2A4

S 2.16

E) 2.u6

Specific personnel reassignments were named.

Changes took place in more or less routine adminis-
trative procedures.

Changes took place in the company's business. e.g.,
product mix changes, more systems and less components
emphasis.

Others

 

 

 

 

 

 

   

M'TITI'I‘flLlfllfljlfilfllflfijlflflflflilflfljflyliflflm”