All Is Data

Barney G. Glaser, Ph.D., Hon. Ph.D.

……Although data is plural “is” sounds better

All is data” is a well known Glaser dictum. What does it
mean? It means that exactly what is going on in the research
scene is the data, what ever the source, whether interview,
observations, documents. It is not just what is being, how it is
being and the conditions of its being told, but all the data
surrounding what is being told. It means what is going on has to
be figured out exactly what it is for conceptualization, NOT
description. Data is always as good as far as it goes, and there is
always more to keep correcting the categories with more relevant

All is data is a grounded theory statement, not applicable to
QDA. Data is discovered for conceptualization to be what it is for
a theory. It is discovered by constant comparison which generates
a category and properties that vary it.

The data is what it is and the researcher collects, codes and
analyzes exactly what he has: whether baseline data, properline
data, vague data, interpreted data or conceptual data (see “Doing
GT”). There is no such thing as bias, or objective or subjective,
interpreted or misinterpreted, etc. It is what the researcher is
receiving (as a human being, which is inescapable). Data is what
the researcher is constantly comparing with tedium, to be sure,
as he generates categories and their properties. Remember again,
the product will be transcending abstraction, NOT accurate

While the QDA researcher may be disappointed with what
he is collecting, the GT researcher’s job is to analyze its
components, its type of data, and take a conceptual perspective on
it. Good as far as it goes means the GT researcher is always doing
a perspective on a perspective (data) with the goal of generating a
theory that resolves continually a main concern, which, as I have
said many times, accounts for the main action in a substantive

For the GT researcher the world is totally empirical. As he
collects data his job is to deal with exactly what is happening, not
what he would want to happen, not what his own interest would
wish the data to be. The data is not “truth” it is not “reality”. It is
exactly what is happening. The GT researcher has to be oriented
to each course of action having its own meaning. To be sure it
does. And once the GT researcher lets this meaning emerge and
sees the pattern, he/she will feel “sure” that this is what is going
on. This sureness can not be known beforehand. It emerges
conceptually through constant comparison.

That the data may not be reality or the truth, should not
disturb the GT researcher. He should keep in mind that, after all,
socially structured, vested fictions run the world, accurate
descriptions run a poor second. Thus data is what is occurring, it
is socially produced and it is up to the GT researcher to figure it
out, BECAUSE the participants are doing it, talking it, using it,
think it, are it, respond to it, offer it and so forth. It is going on
right in front of the GT researcher! For example, treating talk (an
interview) as data comprises not just what was said, but that the
talk was given, in a certain way, in a certain context, with a
certain endurance, in a culture, with talk story attached etc., etc.
The same reasoning applies to reticent talk, shyness, silence etc.

While the GT researcher is not concerned with forced
accuracy, he has the job of being honest in comparing exactly
what the data is on as many dimensions as possible. Again, not
want he wants the data to be, nor what he did not take the time
to figure out. This is part of the GT method brought out by
constant comparisons, which conceptually bring out the proper
distortions (fictions) inherent in all situations.

The “all is data” prospective requires many, many incidents
to compare and saturate categories to emerge the variations,
which others may term distortions. No one incident categories are
part of the GT method, as they may be in impressionistic QDA
(see “Doing GT”). The constant comparing of incident to incident
and incident to category and its properties is a must, to bring out
what data is actually going on. The one incident concept commits
a version of the ecological fallacy. A particularistic (one person)
source of a concept is generalized to a pattern applicable to many
people. An impressionistically grounded concept leads to a
grounded conjecture, not a grounded theory. It can be used with
the pattern rhetoric and a macerated sound bit approach to
saying GT has been generated, with no notion of the systematic
nature by which it should be generated.

In comparing incidents, in most researches this is dominated
by one type of data, usually an interview-observation in field
notes. This is not hard comparing; it just takes conceptualizing
thought. But uniformity of data is not required for comparing.
Diverse data from other emergent sources can be compared. By
diverse I mean whatever may come the GT researcher’s way
while theoretically sampling: documents and current statistics,
newspaper articles, questionnaire results, social structural and
interactional observations, interview, casual comments, global
and cultural statements, historical documents, whatever,
whatever as it bears on the categories.

How far the GT researcher goes depends on his theoretical
sampling, skill, maturity and the emergent. Most beginning GT
researchers are somewhat timid about their ability to do GT.
They, therefore, prefer the uniformity of data collection from a
site where the area of interest goes on. This frequent approach of
the beginner sets him/her up for description capture, since it
mirrors some QDA requirements about uniform data. However
the set up is a safe harbor until a true GT is generated; he can
fall back on a QDA description. It is not unwise for a start in GT
research, but not necessary!

How far the GT researcher goes in diversity of data to
compare is based on some judgment, but more so on what
emerges as the researcher goes to comparative groups. Diverse
data is a delimited occurrence. Anselm and I, in doing the dying
studies, as mature researchers, compared everything in sight to
bring out our awareness context theory.

One frequent problem GT researchers have, especially
beginners, is wanting the truth in an area dealing with “delicate
matters”. However, they just get properline data or what they are
suppose to hear and see. There is no reason to be honest as a
participant, as revealment can easily bring troublesome
consequences. But properline data is the data, and it tells us
what is going on to prevent people from knowing what is actually
going on. The study can be seen conceptually as one of how the
public is handled based on concealment, nonrevealment or hidden
agendas. It may take infiltration as a participant-observer to find
out what is concealed. This becomes another study. Vague data,
interpreted data or conceptual data may result in the same study
of derailing revealment. They all obfuscate. They are normative
properties of spinning. The QDA quest for accuracy is out of place
in a GT research.

Data Worry, Data Doubt

The concern with accuracy of data, comes from and yields an
immense literature on data worry, data doubt and the need to
produce accurate descriptions. Data is always suspected as bias,
subjective not objective, untrue, poorly interpreted, bad or
contaminated and otherwise distorted and suspect. It is suspected
humbly by the researcher, who wants readers to believe it
anyway, and strongly by critics, who say it just is not what is.
This indictment is continuous, especially by arrogant quantitative
researchers. Strenuous corrections are always advised.

This critique is moot for GT as GT produces abstractions not
descriptions. Distortions are just more variables to conceptualize
and make a part of the analysis. Remember though, to be
relevant distortions must earn their way into the theory as they
are conceptualized. GT reduces a distortion to the theory
depending on what theory is emergent and how relevant it is to
the theory as another category or property of it. Multiple methods
of data collection are good for diversity of comparisons, but
multiple methods of data correction are not necessary. All is data
is “what is” as it comes “as is”.

At the end of this chapter I will deal with some of the
discussions of overconcern with distortions of data type, to show
how GT frees its researchers from this tyranny.

Transcending Worrisome Data

Data worry, data doubt as I said, permeates all methods of
QDA. It is a constant tyranny. All (almost) researchers, who
describe, work the polar opposite of feeling the data is condemned
to distortion and trying to legitimate it as objectified to the point
of being accurate and of being evidence. They admit the distortion
and insist on the legitimacy of their factual descriptions. They
give up positivism yet use its canons for data: accuracy,
objectivity, reproducability and so forth. They constantly argue
for acceptance of their somewhat wrong data as “OK” under the
circumstances, conditions and contest.

I have not seen one article that considers conceptualization
that considers abstraction from time, place and people that
considers this as freedom from the tyranny of normal distortion
by humans doing description. GT is a methodology that provides
this conceptual freedom in a highly scholarly and useful way. GT
makes the condemnation and preoccupation with accuracy and
evidence moot.

GT gives this freedom from distortion tyranny to data that is
sourced from interpretations, constructions, voice of participant,
personal experience of various kinds, culturally differential,
language differential, objective-subjective, value laden, behavior
vs. spoken, truth vs. properline, credibility of informants,
selective data collection, selective non-random sampling, multiple
versions of the truth, historical, biographical, gender bias,
varying interview and observation techniques. The reader may
think of more sources. To repeat in this context, all that GT does
is to rigorously generate conceptual hypotheses that get applied
with fit, relevance and workability (explanatory).

Since there is no literature on the freedom from accuracy
brought on by GT, I will consider here how GT frees up several
sources of worrisome data.

Hard Data

The elite, haloed data of science is hard data. It is the data of
the biological and natural sciences. It is hopefully what survey
researchers produce as quantified. It is bedrock data. By
comparisons qualitative data is called soft and demeaned as such,
implying that it is rife with subjectivity that make it inaccurate
and uncountable. It is seen as another form of contrived
journalism. It is seen as not being able to “sort fact from fiction”
one authors says, because it has “no rigorous, critical standards.”

Without arguing on these merits, GT brings freedom to this
invidious comparison. Categories and their properties apply fully
variable. They leave the time, place and people of hard data or
soft data. It is not truth that transcends; it is conceptualization!
The data for GT research is not seen as limited data needing
scientific support, as it is in QDA description. GT research data is
just “whatever” data to compare.


Another worry about QDA data is whether or not it is
reproducible by others. In this way, another researcher can come
back to the same action scene and reproduce the events described
by another to verify the descriptions. Or, would another
researcher find the same material. This is a check on whatever
bias may occur, so it will not be seen as particularistic to one

This is not relevant to the GT abstraction. It is conceptual
and applies to any action scene. Furthermore new (albeit
different) data from the same site simply get compared and
modify the theory. The indicators need not be reproducible, just
comparable and interchangeable for conceptualization.
Theoretical sampling also takes the researcher to site spreading,
which makes reproducing the same data for similar abstractions
virtually impossible.

Reproducibility is also irrelevant for GT, because new data,
whether the same or not just get compared to develop new
properties of categories that saturate the categories. Saturation is
that no new interchangeable indicators emerge to add more
properties to the categories, which are of relevance. If another
theory is developed from the same site since highly different data
are found, then we just have another GT, that’s all. Each
grounded theory with its general implications stands on its own.
For example one can find supernormalizing in similar data that
yielded cutting back: in data on heart attack victims and sports
injuries. They are a paired dichotomy, related by being at
different ends of the same continuum but standing on their own.
Each was a different dissertation.

If the incidents to compare are events that could never be
reproduced, that GT is even more freeing and powerful, since it
conceptualizes the underlying patterns in seemingly different
events. An example is social protest events, any one which cannot
be reproduced but many certainly can be compared for generating
a GT of non political social protests.

Selection Bias

Selection bias refers to non-random sampling for the purpose
of consciously or not so consciously getting a preferred distortion
in the data. Deliberate selection bias over — represents cases at
one end of the distribution. This then confirms a preferred
hypothesis or yields a preferred description, or it just produces
preferred distortions with no purpose or an external purpose. The
larger the sampling of people, cases, units or event selectivity the
more general the factual distortions appear, but they still are
distorted by bias.

Again GT conceptualization by constant comparison will
show the underlying bias as it emerges — for bias is just another
variable. It is an independent variable which varies other
variables in the preferred direction. GT conceptualization frees us
from the tyranny of selection bias’ affect on data and the tyranny
of explaining it away.

Theoretical sampling is diametrically opposed to selection
bias. The theoretical sample is chosen for theoretical saturation of
categories. Theoretical sampling directs selection for a theoretical
purpose. If bias creeps in, then it will surface as another category
by constant comparison and saturation.

Writing Authority

The implied claim of all writing in social science is that the
description must be correct if it is published. The writer is
assumed to be a competent observer who reports accurately on
self and others. Publishing legitimizes descriptive writings. It is
the mystique. Except for a few critiques, which most people do
not read, publication certifies authoritativeness. Thus publication
certifies distortions of the objective in descriptions. It is hard for
most readers to appraise the research techniques used in a paper
or monograph. We trust the editors who review the writing before
publication and they may, and usually so, favor the bias in
descriptions. This is just the way it is and always will be.

However, this authoritative attribution is moot for GT. GT
categories and their properties are fully variable as applied. The
applier is not wedded to any position as emergent fit goes on, as
modification goes on and as generalizing is accomplished when
applying the theory. It is the conceptions (categories and their
properties) in writings that grab the reader and who knows how
they will in fact use them. If the conceptual statements are
reduced to description, the power of variable fit and relevance is
lost. This reduction however does not ruin the abstract ideas and
the abstract promise of GT.

Personal Predilections

There are several ways that personal predilections distort
descriptions. To mention several which all overlap, they are
prejudices, value laden, ethics, ideology, psychological blocks,
spinning distruths, preconceptions that are unchangeable,
oversimplification, axes to grind in favor of a position, human
limits by culture, education, etc., differential perceptions and
interpretations, intense reflexivity and self accounting as a
researcher proceeds, an immaculate description need no matter
how the data falls, methodological entrapment on how to process
knowing, romantizing, exaggeration and so forth. I am sure the
reader can add to this list. These are all quite human and
inescapable totally. They are just able to be limited as much as
possible in effecting the description.

GT rescues us from this personal biasing whatever the
source. Once the theory is generated it, as we saw above, has
total conceptual endurance and generality however someone may
distort its use. Cultivating, supernormalizing, credentializing,
desisting residual self, pluralistic dialoguing, cutting back, client
controlling, enhancing creativity and so many others grounded in
GT research just exist abstractly. Personal predilection variables
simply emerge as variable in the theory and are put into relief by
constant comparing and theoretical saturation. They are not
discounted if relevant, there is no need to explain away the
concept based for example on a particular value. The distorting
effect of the bias must earn its way by working as part of the GT.

A last note on personal experience as data. From the point of
view of GT, there is no such thing. It is just another interview
and/or observation to be compared into the theory by constant
comparison. It is no more. By comparing we then see how
particularistic, eccentric, idiosyncratic, normative, structured
etc., the experience is. This applies to journals on “lived
experience”, biographies or whatever the source however rich it
may appear. GT research is after conceptualization of social —
psychological patterns, NOT of individual patterns whether
described or conceptualized.

No matter how apparently insightful and objective an
individual account may appear of an area of interest, it is still
just individual. It is just a set of incidents to compare to many
others from other people. Learned informants may be valuable in
a research, but they must constantly be checked on by other data;
must saturate by theoretical sampling of many other participant
incidents. Personal experience research is just that, personal
experience, not social psychological research.

In sum, GT is on an abstract level. It is bias neutral and fully
variable. And to be sure it can be applied with bias by an
individual who wants to issue it out, whether manifestly or
latently, awaredly or unawaredly.


One major worry in QDA research is a different take on
personal predilection. It is seen as the researcher being a
constructionist or a constructor of data. Listen to this author. “A
strategy of inquiry comprises a bundle of skills, assumptions and
practices that researchers employ as they move from their
paradigm to the empirical world. Strategies of inquiry put
paradigms of interpretation into motion. At the same time,
strategies of inquiry connect the researcher to specific methods of
collecting and analyzing empirical materials. Most methods rely
on interviewing, observing and document analysis. Qualitative
research is endlessly creative and interpretive. The researcher
does not just leave the field with mountains of empirical
materials and then easily write up his of her findings. Qualitative
interpretations are constructed.”

The point is clear in this orientation, data is constructed
with interpretations. This QDA approach is clear, volumes of data
are collected according to a preconceived framework. Writing up
this mountain of data is not easy even with a framework. The
QDA researcher constructs first the data and then the write up.
The resulting descriptions therefore suffer from all the personal
distortions mentioned above and more I discuss briefly below.

Listen to this author: “Transformations in data occur when
ignorance and misapprehensions give way to more informed
insights by means of a dialectical interaction. Knowledge consists
of those constructions about which there is relative consensus
among those competent to interpret the substance of the
construction. Multiple knowledges can coexist, and these
constructions are subject to continuous revision with changes
most likely to occur when relatively different constructions are
brought into juxtapositions in a dialectical context. Knowledge
accumulates by a process of accretion with each fact serving as a
kind of building block which adds to the growing edifice of
knowledge based on more informed and sophisticated
constructions which progress over time.”

GT research, again, transcends this constructionist
cumulative approach. Properly done, GT research delimits the
data collected and provides the emerging framework for analysis.
This limits construction and then the ensuing conceptual
analysis, as I have already said, abstractly reveals the biasing
categories as well as the other categories. What is going on, goes
on abstractly no matter the take of the constructions and
interpretations based on more and more building blocks yielding
more accurate meaning. More and more data for GT just leads to
more constant comparison which leads to modifications and
generalizability. This worrisome take of constructions on data
collection and analysis for description and the accumulation of
knowledge are moot for GT conceptual generation. All is data,
and GT is good as far as it goes, and can always go further to
more theoretical completeness for a substantive area and then on
to formal theory based on comparisons of data from many groups.

For example, a theory on becoming a nurse can be put on a
more formal level of becoming a professional by comparing the
substantive theory to data from theoretically sampling of many
professional schools on the relevant categories. The accumulation
of knowledge for GT is not more precise data for description, it is
more abstract theory or formal theory (see “Status Passage”,
“Organizational Careers” and “More Grounded Theory
Methodology,” all by Glaser.)

Another example is that a basic social process like default
remodeling in corporate banking takeovers goes on, and is
neutral to the distortions of data. In fact the whole study may be
on the proper distortion of a basic social process, say on the
properline that the takeover is cooperative. But the process is
neutral to bias and a fully variable abstraction to be applied to all
corporate takeovers with more comparisons based on theoretical

Lastly, a constructionist’s assumption is that constructs are
constitutive of being human. Constructs are thought objects that
try to grasp social reality which get more accurate as constructs
get more sophisticated. By the same token GT conceptualizations
are constitutive of being human also and GT becomes more
verified as categories saturate and the theoretical completeness is
reached. Fit and work and relevance is the goal of GT to strive
for, not accuracy.

Verbal-Actual Behavior

Accuracy of description is big on this source of bias between
telling about behavior as seeing it. Given the popularity of
interviewing in QDA, some authors worry: “The mistake is to
treat the verbal formulations of subjects as an appropriate
substitute for the observation of actual behavior”. Or “Such
studies may suffer from the gap between beliefs and action,
between what people say and what they do.” Or, “We only get into
difficulties if we treat patients’ responses as standing in a one-toone
account with what happened in the actual consultation.
Exaggerated stories give vent to thoughts which had gone on
unvoiced at the time of the consultation.” Or, “stories, no matter
their political spin, represent part of the substantive theory that
will be generated. It is not that the participant is untruthful, it is
rather that he does not wish to tell the whole story since this
might put the story and him in a bad light.”

This is a perennial problem in accessing the distortion of a
QDA description that is based on telling behavior much more
than observing it. Participants properline to strangers with ease.
The GT research method treats all as data and the study is on
what is given to the researcher, verbally and actually, not an
abstract view of accuracy, but an abstract theory of what is going

A preponderance of verbal data yielding a GT theory is
merely modified by constant comparisons when observation or
documents become part of the theoretically sampled for data.
Grounded theory is good as far as it goes, remember! It may
slowly be modified by observations bringing out more properties
closer to reality, but never neglecting the reality of the
participants’ view, while being sure that a gap is maintained
between verbal and actual.

For the GT researcher the issue is all is data, whether
private or public, revealed or concealed, words or behavior. It all
bears on generating concepts for the core variable or basic social
process being studied. The talk is data, the behavior is data and
the gap between them and its sources is data. The emergent
theory should sort it out conceptually. What actually goes on is all
of it.

This verbal-behavioral data juncture goes on frequently in
business studies, where properlining is the norm, since
concealing information is so important. Descriptions may be
inaccurate. But GT conceptual hypotheses can be quite relevant
and subsequent data does not correct so much as it just generates
modifications. For GT the quest for “what is really going on,” is
just that, in full abstraction; it is not actual behavior.

Gender Bias

The history of feminist thought has been brought out clearly
in feminist qualitative research. As Virginia Olesen says,
“Whatever the qualitative research style, and whether or not selfconsciously
defined as feminist, these many voices share the
outlook that it is important to center and make problematic
women’s diverse situations and the institutions and frames that
influence those situations, and then to refer the examination of
that problematic to theoretical, policy or action frame works in
the interest of realizing social justice for women.”

Unless focused directly on women or a woman’s problem,
male gender dominance in QDA research, according to feminists,
still exerts control. However male oriented, influence frameworks
as well as a healthy white female middle class dominant
frameworks are crumbling in the face of steadily incisive
criticisms of women of color, this world feminists, disabled
women, and lesbian women. Now a substantial body of feminist
research ranges over the full span of qualitative methods. Gender
dominance favoring males is diminishing in qualitative

While important in QDA research, gender bias is irrelevant
for GT research. GT research is gender neutral as it is for all face
sheet variables. Gender like race, color, ethnicity, age, education,
religions, culture, social class and so forth must earn its way into
a GT as a relevant variable. When relevant in a GT, gender will
emerge as a variant of the dependent variable.

For example in pluralistic dialoguing gender is only relevant
if the males on the team either defer to “take charge” ladies
(nurses) or dominate their stereotypes. It did not emerge.
Competence was a major factor and there is no gender monopoly
over competence. In another study of same sex (lesbian)
marriages, we found that the emergent division of labor between
the couple was the same for heterosexual marriages. Gender was
not an issue. In another study of woman’s health clinics, gender
did emerge as relevant in women relating to a female doctor
sensitive to woman’s problems. Judith Wuest’s article:
“Precarious Ordering: A Theory of Women’s Caring” focus’s on the
demand on women to provide care. This bias focus misses the
whole literature on entrapment care of disabled, elderly, infants,
etc. which falls on spouses of either gender or family members of
either gender. The demand for care, which is inescapable is
gender neutral.

Thus in GT research when gender emerges as relevant, fine,
bring it into the theory, but not until then. Yet the feminists will
try to issue-out from the start a neutral conceptual analysis to
being a woman’s issue when it is general or gender neutral. The
issuing-out comes after generating a conceptually neutral GT
theory when it is applied and the issue can be turned many ways,
for example, to color, economic class, type illness, age (elderly)
gender or what. Gender has no priority on issues. The over-focus
on gender may be fine for routine QDA feminist researchers. It
has no place in generating GT which is abstract from people.

The tacit assumption of male tone, male dominance, male
focus, is empirical in the data and neutralized in the
conceptualization of the GT. Reverse dominance is just as
relevant. So beginning bias is another preconceived forcing of the

Value Laden Inquiry

There is probably no such research producing value free
objective description, which transcends opinion or bias. Even
hard science has its own soft work when it turns to interpreting
hard findings by conjecture seeded with values. These
interpretations are typical descriptive material only more or less
grounded in other research conjecturally related to the findings.
Only GT abstractions are value neutral. They can be applied
based on any value.

There are several sources of value laden research which
almost by definition ignores accuracy in favor of taking a position.
Value laden inquiry comes from a particular world view, religion,
a political stance, a gender stance, ethics, a hard-line
methodological view, a granting organization with a cause, etc.
This distortion of accuracy is not of concern to the QDA
researcher doing the research. It is his job to distort properly. It
is only of concern to the reader judging its merit. The reader tries
to judge who owns the knowledge coming from the research
sufficiently well which allows the researcher to cast it with
appropriate values.

This distortion overlaps with much of what I have said
above. The value which distorts soon emerges as a category when
applying GT methodology to the same data for generating a
theory. It may be a value of the participant or the researcher. No
matter, its impact in the GT will be seen abstractly, however
unchanged the QDA description. For example in a study of
disavowing credit among ward nurses, it soon becomes apparent
that the disavowing was a value based on the authority structure
in medicine. The nurse, whatever her achievement, should defer
credit to the doctor in charge.

By the same token the researcher may add his/her values to
interpretations of the data which distort it. Again GT being value
neutral will begin to show the value as a category in analyzing
the same data by GT methodology. This was seen clearly in a GT
study of discharging patients in a private psychiatric residence. A
previous study described the discharging based on curing very
troublesome patients. Curing was successful and of great value. A
GT study explained the discharging as getting rid of troublesome,
incurable patients and keeping the successfully treated in order
to maintain a calm sentimental order in the facility which made
the‘ facility appear as successful.

The Voice

QDA research, once legitimized many years back, was
revolutionary on the dimension that AT LAST it gave the
participants the chance to be heard as they wanted to be heard,
not as some preconceived questions would steer them. The
participants voice was listened too! This, of course, ushered in the
problem of worrisome distortion in reporting the voice to others.
Could the researcher listen at all or listen well enough, was he
patient enough, was he listening with bias, was he listening with
cognitive disjunctures (speaking past each other), could the
researcher handle the lack of uniformity of listening to many
voices (when giving up the uniformity of preconceived questions)
and so forth. We have seen above how “voice” data could be
distorted in so many ways.

The QDA researcher listens to the data in terms of a
preconceived framework to help him organize it. The GT
researcher is in seventh heaven listening to what ever is voiced as
he constantly compares to discover the participants main concern
and the core category and sub categories. The GT researcher
knows that latent patterns exist and will emerge complete with
conceptualized distortions as variables relevant to a GT
explaining the ongoing action. The constant comparing will
correct and precise the meaning involved even as it abstracts the
relevant categories. In contrast the QDA can only hope his preresearch
devised framework will work to yield a not too distorted

Now-a-days QDA is in fashion and problems are studied
from the perspective of the participants. QDA types will vary on
their approach to the voice: one phenomenologist said “more
precisely phenomenological research where the primacy of the
subjective experience of the participant takes precedence over the
interpretation of the researcher.”

Thus from a history not being heard research has swung to
intense in depth listening to the voice. Critiques of how the
“voice” is listened to actually often stifle the listening in the
service of getting it right. Efforts at objective reporting
description yield to subjective recounts. Objective pretence is
given up. Writing authoritativeness is used in its stead: if it is in
writing it must be close to accuracy. Verisimilitude is used: truth
and proof are the same for tests of objectification. The canons of
objective, natural science are still trying to be adhered to. Also
these QDA researchers feel strongly that the researcher owes it to
his participants to tell their story accurately. And also the QDA
researcher owes it to the participants to tell them what is being

This struggle over how the voice is heard and related in
research writings is non-existent for the GT researcher. It is all
just data to be conceptualized for theory however the data comes.
And, as I have said above, it is never advisable or owed for that
the GT researcher tells the participants their conceptualized
main concern and its continual resolving. The only exception to
this is in carefully applied consultations or in action research,
when substantive theory is applied for use rather soon after
discovery. Remember GT generates categories’ labeling patterns,
which is merely about what is going on, not for or against and not
for corrective action. People disappear into these patterns which
abstract their behavior. GT is not the participant’s voice, it is the
patterns of behavior that the voices of many indicate. These
patterns fit, work and are relevant to the behavior the voices try
to represent.

Put another way, GT substantive theory is not a privileged
reporting of the participants voice because the GT researcher was
“let in” to confidences. It is a substantive theory explaining voice
productions. Participants often recognize the power of the
conceptual grab of GT in making sense of the many voices in an
action scene when GT is used as consultation. I cannot emphasize
enough that the current unleashing of participants’ voices is just
more data for GT research. That’s all! GT is voice distortion free.

Whose voice is a worrisome distortion? In “cooperative
inquiry” which sounds a bit like participant-observation, the
researcher and the participant have deep extended mutual
sharing conversations. Interviewing is minimal, in favor of
sharing. Thus is the voice a composite of both researcher and
participant or more the researcher or the participant. Whose
interests are being voiced? Has the study subtlety been switched
to the researcher’s concerns? Maybe, but if so, then that is what
the GT research is studying! It becomes a study of subtle
researcher manoeuvring participants, awaredly or unawaredly,
and becomes part of the literature on researcher impacting
changes on the research scene and its participants.

As educated people, researchers can easily have this
distorting effect on less educated people. In cases of this nature,
the GT researcher could clearly not know the problem he will
study beforehand, and will probably be quite surprised at what
emerges. He is likely to find that cooperative inquiry produces
properlining, patronizing, unaware projections, cultural and
economic dominations and consensus collusions in support of
sharing rather than saying it as it is. Listening correctly is
subverted by sharing and story competitions.

In contrast, with the preconceived framework to sort out
conversations, the QDA researcher may get what he wants from
the conversational data. However distorted it may be, he can
force it with self determination and an appeal to conviviality.
Closeness to the participant does not insure accuracy. In fact one
author felt the closer the relationship the more unreliable the
data. None of this applies to GT Research which theoretically
samples with a myriad of interview relationships, lengths and
styles in one research. And then treats all as data.

Cooperative inquiry data will be a challenge to the GT
researcher doing secondary analysis of it. He is likely to find
himself in the midst of a methodology research, rather than
substantive area. Managed researcher research may be more
relevant than the substantive material. It will be like the study of
questionnaire response sets in survey research, when responses
do not make sense.

Another word about action research is in order. In this case
the GT researcher helps the participants do their own and/or
share in his research going on in order for them to use
purposively the resolving patterns of their main concern. This
consciousness in the midst of action has its draw backs. It is hard
for the untrained to have this research ability and perceptivereflexive
ability. The tendency is for participants to become
defensive over fictions and ineptitudes and to become self-serving
in lieu of being fully objective in applications. Also preconceptions
easily run ahead of the research, as this is how they think. The
need is for the corrective power of a consultant, who puts good
judgment and limits on use of the substantive theory. The action
oriented participants will perforce be descriptively bias, while the
GT researcher is abstractly neutral. This conflict will likely be
resolved for the former, since GT neutrality will be unacceptable.

In sum, the goal of understanding the complex world of lived
experience from the point of view of those who live it is laudable.
The world of lived reality and situation-specific meanings that
constitute the general object of investigation (the “voice”) is
thought to be constructed by social actors. The aim of attending
carefully to the details, complexity and situated meanings of the
everyday life world can be achieved through a variety of QDA
methods. But all offer merely the inquirer’s construction of the
constructions and meaning interpretations of the participants one
studies, nothing more.

Remember, that no matter how different these offerings of
data on the “voice”, for the GT researcher it is just data to be
submitted to the abstracting of the GT methodology by the
constant comparative method. The QDA struggle with worrisome
distortions stops at the start of GT research, as distortions of
voice are the data. Remember also, that it is just as human to
conceptualize to this freedom from time, place and people as it is
to somewhat bias descriptions with the wrestle between objective
and subjective that takes so many forms. Facts become sacred for
the written moment and concepts become easily modified by the
next comparative.


GT relies on constant comparing of incidents to generate
categories. Constant comparing saturates categories no matter
what the data bias, nor how complete any one interview is. The
data one gets is all the data there is. Field notes, observation
notes and interview notes are sufficient for GT. They are part of
the delimiting aspects of GT data collection.

GT does not produce findings or facts as I have said over and
over; it produces conceptual hypotheses. Therefore taping for
evidential accuracy in descriptions is an unnecessary waste of
time and resources, and restricts theoretical sampling for concept
generation. I have written at length about this in “Doing GT”.

Thus these statements by one well-known grounded theorist
are just plain wrong. “Open-ended interactive telephone
interviews were conducted to determine the participants
experience of bereavement. All interviews were tape recorded and
transcribed verbatim. All interviews were immediately
transcribed to ensure the accuracy of the transcription.”

Capturing everything by taping in the service of evidentiary
accuracy does not reduce data bias. It may increase it, since the
best way for a participant to tell a story is to exaggerate it or to
engage in extreme relativism to make a point. The GT researcher
need only collect data theoretically and let bias emerge. He does
not need full coverage per interview or evidence for incidents.

GT research teachers and students learn this lesson
gradually with experience. Listen to this teacher from Portugal:
“My students used to tape and transcribe the interviews. They
give me the transcriptions and we discuss it in the group. Not
taping and just doing some notes is tempting. We would gain time
and above all we could start doing theoretical sampling.
Something we could never do before.”

And Wendy Guthrie says: “Some months into the research, I
was concerned that I might be missing important pieces of data
by relying on memory. I also feared the delayed write-up of field
notes might cause me to forget vital information. To guard
against these dangers I decided to take a tape recorder with me
on my visits. I soon realized that this was not the way forward
and discontinued my trial recording. One reason I would not
recommend taping is that it gives a false sense of security.”
In short, taping does not do away with bias, it might increase
it and after all is said, bias is not an issue for GT
conceptualization. Complete interviews are not a source of
accuracy and both are not needed for GT.

Poor Instrumentation

To be sure poor instrumentation is a source of worrisome
data. “All is data” is good as far as it goes. GT in abstracting from
data can pick up bias from poor data collection up to a point. I
have never seen this point reached, but it can be logically. I have
seen unusable data from surveys, when people hired to
administer questionnaires, just fill them in themselves to save
time and make money.

GT can be used on any data collected by a myriad of
methods, so the GT researcher must always be wary of severe
collection problems and truly fake data. Overly preconceived
frames in a collection method may severely miss the point. This
seldom happens. The GT researcher can always ask the question:
“Does the sensitivity of the data selection and data collection
method match the needs of the research.” The answer is mostly
“yes” for GT research. The net balance of data collected which is
used for secondary analysis is empirical, and becomes part of the

Whether secondary analysis or original research GT
researchers seem to prefer qualitative data. They prefer deeper
understanding data on small samples as opposed to large sample
surveys. Thus virtually all of us use interviews, observations, and
documents. Some tape record and many do field notes or a
combination of both. It is important to always keep
instrumentation in mind. There are many forms of interviews in
theoretical sampling. They vary from long interviews to quick
question shots on the fly. This lack of uniformity does not mean
shoddy collection since it not a pretested questionnaire guide.
BUT a disastrous interview now and then should be noted in the
incident, and is probably part of the data. The interview fits like
the emerging theory generated from it. A poor interview should
be evaluated in this light, when and if necessary. A poor
interview will be picked up by constant comparison with solid
collected data.

The same reasoning applies to taking field notes. Most are
written either on site or soon after. Waiting a few days to write
field notes does usually deter good notes. But not necessarily.
However, deterred field notes also occur on an association basis
later on and can be quite good. Constant appraisal of data
collection methods is necessary but cursorily easy. Aberrations
from emergent patterns are picked up by constant comparison.

Taping interviews can have failures or interim failures.
Forgetting to turn on the recorder or poor microphones can wreck
data. Tape recorders more than interviews evoke properline data:
people tell interviewers what they think the interviewer is
suppose to hear. People do not like to worry about negative
feedback from recorded conversations. For many QDA
researchers this is poor instrument data, because their quest is
accurate truths. For the GT researcher properline data is fine. It
generates a relevant theory about the fictions of social life.
Properline data is a very frequent form of data, which many QDA
researchers deny or dislike or take as truth unknowingly.

GT is a theory of method with a view of collection techniques
attached. The primary view is “all is data” no matter how
collected, which itself may become part of the theory. Collection
mix is not relevant until it becomes so. A GT researcher doing
original research should choose the best fit collection method. Fit
will be a combination of substantive, diplomatic and that allowed
by authorities and/or informal culture-structure.

Fact orientation makes instrumentation far more important
for QDA research than conceptual orientation of GT research.
QDA researchers of different orientations dialogue about the best
collection method for claiming accuracy. GT research is collection
method neutral, because it can conceptualize any data, which
makes it neutral to most poor instrumentation. GT is a general
methodology usable on any data, and it is up to the researcher to
figure out exactly what the data is.

In sum, the many aspects of GT methodology and its
conceptualizing of data, resolve and render moot these many
sources of bias descriptions. GT is bias neutral when done
properly and when “all is data”. I am sure the reader can think of
more sources of bias on the descriptive level that become
neutralized by various aspects of GT methodology which
abstracts from time, place and people. The biases mentioned
above overlap conceptually but are seen as empirically very
separate sources of worrisome data that must be dealt with to
some extent.

QDA is a multimethod area of research. No matter what the
QDA method used, the research typically begins with the
worrisome discussion of the type of data, nature of the data and
how its reality is socially constructed with meaning and
interpretation. The supposition is that the research will yield a
description with “some” kind of accuracy. The QDA researcher
justifies why the type of qualitative data is appropriate to study
his research problem to ward off critique.

A socially constituted meaning is framed by the preconceived
notions of a grand theorist of what accounts for social order. The
forcing framework turns research to conjecture however subtle. A
professional interest is elaborated. It all sounds so sensible, and
to some degree it is, however worrisome the distortions. The aim
is not to provide causal explanations of patterned behavior, but to
fully describe how members recognize, describe, explain and
account for the order of their everyday problems. Descriptions can
become unending when forcing an overdue type of qualitative
method’s data into a framework that does not delimit the
descriptions however bounded the unit and the data collection

When “all is data” the above QDA quests and efforts do not
matter. What matters is exactly what the data is, and then its
conceptualization by GT methodology. There is no need to justify
a type of data in relation to a problem. As we have seen from
above chapters, it all emerges to generate a theory. Accuracy or
data distortion is not the issue. Figuring out what the data is and
then conceptualizing it is the issue. This occurs with constant
comparisons yielding categories. Thus GT is an open
methodology, which yields conceptions of patterns from any kind
of The theoretical generation and integrations of these
conceptions is detailed at length in my other books. The tyrannies
of factual description do not occur in GT conceptualization.

Historically there has been a heavy emphasis on
quantification in science. Those sciences that lend themselves
especially well to quantification are generally known as “hard”.
Those less quantifiable sciences, particularly the social sciences
are referred to as “soft” denoting imprecision and lack of
dependability. In the last few decades this positivistic position
and its beliefs have come under scrutiny: the quantification
methods miss the meaning given by the individuals studied. Ergo
the growth of several qualitative methods have been employed to
contribute this meaning.

No single qualitative method can grasp the subtle variation
in ongoing human experience. As a result qualitative researchers
deploy a wide range of interpretive qualitative methods and
always seek better ways to make more understandable the worlds
of experience that are now being studied. These QDA researches
over the years have led to the realization and conclusion that
there is no value-free inquiry for studying human experience.
Researchers now struggle to develop situational and transcending
methods that apply to any qualitative research to reduce the level
of worrisome distortion. They wish to record their own
observations and interviews accurately while still uncovering the
meanings their subjects bring to their action. Their success is
asymptotic description accuracy: that is closer and closer to
accuracy. The QDA methods compete in their struggle with more
and more sophisticated and informed methods. The history of
social research is on their side.

Then in 1967 along comes grounded theory, born out of
descriptive qualitative research. After a few decades of maturing,
it has become clear that GT has left behind the struggles of
worrisome qualitative data. By its emergent abstraction or
conceptualization of the data to generate a theory that, when
applied, explains the very data the QDA researchers are trying to
describe. GT transcends QDA research by rendering so many of
its sticky struggles moot and thereby presents the next historical
stage in social psychological research. It is a highly rigorous
abstracting methodology which grounds theory in what is going
on whatever the data.