Quality & Quantity 36: 391–409, 2002.
© 2002 Kluwer Academic Publishers. Printed in the Netherlands. 391
A Purposeful Approach to the Constant
Comparative Method in the Analysis of Qualitative
Utrecht University, Faculty of Social Sciences, Department of Methodology & Statistics, PO Box
80140, 3508 TC Utrecht, The Netherlands (E-mail: firstname.lastname@example.org)
Abstract. The constant comparative method (CCM) together with theoretical sampling constitute
the core of qualitative analysis in the grounded theory approach and in other types of qualitative
research. Since the application of the method remains rather unclear, researchers do not know how
to ‘go about’ the CCM in their research practice. This study contributes to a purposeful approach of
the CCM in order to systematize the analysis process and to increase the traceability and veriﬁcation
of the analyses. The step by step approach is derived from and illustrated with an empirical study
into the experience of multiple sclerosis (MS) by patients and their spousal care providers. In this
study ﬁve different steps were distinguished on the basis of four criteria: (1) the data involved and the
overall analysis activities, (2) the aim, (3) the results and (4) the questions asked. It is concluded that
systematization of qualitative analysis results from the researcher using a sound plan for conducting
CCM regarding these four aspects.
Key words: qualitative research, constant comparative method, grounded theory, qualitative analysis,
The constant comparative method (CCM) together with theoretical sampling con-
stitute the core of qualitative analysis in the grounded theory approach developed
by Glaser and Strauss (Glaser and Strauss, 1967; Strauss, 1987; Glaser, 1992).
Comparison is also the dominant principle of the analysis process in other tradi-
tions of qualitative research. All kinds of aids, such as memo writing, close reading
and rereading, coding, displays, data matrices and diagrams support the principle
of comparison. Many academic works have offered interpretations, explanations
and illustrations of grounded theory, as well as providing relevant techniques, pro-
cedures and rules of thumb (Strauss and Corbin, 1998; Wester, 1995; Denzin and
H. R. Boeije, Ph.D. works as an assistant professor in the Department of Methodology and
Statistics of the Faculty of Social Sciences at the University of Utrecht. She specializes in research
and education in the ﬁeld of qualitative methods.
392 HENNIE BOEIJE
Lincoln, 1994; Strauss, 1987; Chenitz and Swanson, 1986; Miles and Huberman,
Although much has been written about these methods and techniques, this is
yet not the case for the important CCM. Researchers often describe at great length
how their studies were carried out, but remain vague when it comes to giving an
account of the analysis. Issues such as the subject of the comparison, the phase of
the research in which it took place, the reason for the comparison and the results
of the comparison remain unclear. It is this lack of explication and account that re-
duces veriﬁcation and therefore the credibility of qualitative reports. A purposeful
approach that can be used by researchers will not only systematize their work, but
will also increase traceability when they describe how they used and implemented
the approach in their research practice.
This article aims to contribute to the development of such an approach for CCM.
Our own empirical research into couples coping with the chronic illness multiple
sclerosis is used to illustrate different types of comparison in a grounded theory
process. From this ‘logic-in-use’ we want to move towards a ‘reconstructed logic’,
that enables others to purposefully carry out their analysis. It is not so important
whether their studies involve interviews with dyads or with individuals. The point
of concern is the systematization of the analysis process. The following questions
1. How varied are the comparative moments according to which elements are
compared with each other?
2. What are the aims of the different types of comparison and to what results in
research do they lead?
3. How do different types of comparison relate to each other?
By answering these questions we hope to make the unclear process of constant
comparative analysis more transparent. In the next section, we start with the prin-
ciple of constant comparison and its place in qualitative analysis. Secondly, the
focus is on the lack of practical rules of how to ’go about’ constant comparison.
Then, the step by step approach is introduced and the different steps are described.
In the following section an example is given that illustrates each step and ﬁnally
we come up with a conclusion.
2. The Principle of Constant Comparison
In the work of Glaser and Strauss constant comparison is important in developing
a theory that is grounded in the data. Tesch (1990) adopts this view when she calls
comparison the main intellectual activity that underlies all analysis in grounded
The main intellectual tool is comparison. The method of comparing and
contrasting is used for practically all intellectual tasks during analysis: form-
ing categories, establishing the boundaries of the categories, assigning the
segments to categories, summarizing the content of each category, ﬁnding
A PURPOSEFUL APPROACH TO CCM 393
negative evidence, etc. The goal is to discern conceptual similarities, to reﬁne
the discriminative power of categories, and to discover patterns. (Tesch, 1990:
By comparing, the researcher is able to do what is necessary to develop a the-
ory more or less inductively, namely categorizing, coding, delineating categories
and connecting them. Constant comparison goes hand in hand with theoretical
sampling. This principle implies that the researcher decides what data will be
gathered next and where to ﬁnd them on the basis of provisionary theoretical ideas.
In this way it is possible to answer questions that have arisen from the analysis of
and reﬂection on previous data. Such questions concern interpretations of phenom-
ena as well as boundaries of categories, assigning segments or ﬁnding relations
between categories. The data in hand are then analysed again and compared with
the new data. The units should be chosen with great care and in a way that enables
questions, new or otherwise, to be answered efﬁciently and effectively, thereby
allowing the process of analysis and in particular the comparative process to pro-
gress. The cycle of comparison and reﬂection on ‘old’ and ‘new’ material can be
repeated several times. It is only when new cases do not bring any new information
to light that categories can be described as saturated. This means that these cases
can be easily assigned to one of the already existent categories in the growing
Comparisons that are highly regarded increase the internal validity of the ﬁnd-
ings. One criterion for qualitative research is that the researcher tries to describe
and conceptualise the variety that exists within the subject under study. Variation
or range exists by the grace of comparison and looking for commonalities and
differences in behaviour, reasons, attitudes, perspectives and so on. Finally, con-
stant comparison is connected with external validity. When the sampling has been
conducted well in a reasonably homogeneous sample, there is a solid basis for
generalizing the concepts and the relations between them to units that were ab-
sent from the sample, but which represent the same phenomenon. The conceptual
model can even be transferred to different substantial ﬁelds that show similarities
with the original ﬁeld. Glaser and Strauss (1965) illustrate that the concept of
awareness context (what each participant knows about the identity of the other)
can explain interactions in international espionage and between hospital staff and
a dying patient.
3. How to Go About Constant Comparison
The literature does not make clear how one should ‘go about’ constant comparison,
nor does it address such issues as whether different types of comparison can be
distinguished. Morse and Field (1998: 130) state that each piece of data must be
compared with every other piece of relevant data. Although this might be true, it
is not a very practical rule, since the question of what a researcher judges to be
relevant remains unanswered. Like other authors, Schwandt (1997) in his Diction-
394 HENNIE BOEIJE
ary of Terms indicates the functions of CCM without describing the procedure for
carrying it out.
According to Strauss and Corbin (1998) the art of comparison has to do
with creative processes and with the interplay between data and researcher when
gathering and analysing data. Beside ‘close-in’ comparisons that serve the more
traditional functions mentioned above, they distinguish so-called ‘far-out’ com-
parisons. The latter are more extreme and make greater demands on creativity,
including as it does comparisons with metaphors and parables. This type is inten-
ded to explicate the inﬂuence of self-evident assumptions and interpretations on
the analysis and contributes to the formulation of alternative interpretations and
questions. Despite pointing out the aim of CCM, the application of the methods
and techniques remains rather unclear. There is more to this process than just com-
paring everything that crosses the researcher’s path. The lack of clear-cut questions
and a prior coding system means that data have to be produced by the researcher
in the course of the study. It is necessary to understand the ’production’ procedure,
which takes place primarily at the data analysis stage, in order to judge the value
of a study.
4. Towards a Step by Step Approach for Constant Comparison
From a thorough examination of our own experiences with qualitative analysis,
we have attempted to develop a procedure for the CCM. The approach described
consists of different steps that will be illustrated with reference to a study of couples
coping with the effects of multiple sclerosis or MS for short (Boeije et al., 1999). In
this study both partners from 20 married couples were interviewed. The research
questions deal with the patients’ individual lives, the lives of their spouses who
care for them and their relationships. Accordingly, the analysis level shifted from
the group of patients and the group of spouses to the couples.
The texts that resulted from transcription of the interviews provided the input
for the analysis process used to make sense of the data and to reconstruct the
perspectives of the groups being studied. The analysis consisted of two activit-
ies, namely fragmenting and connecting (Dey, 1993). Both strands are necessary
and keep each other in equilibrium. The former emphasizes the separate themes
which emerge during the interview and focuses on an individual ordering process
which is relevant to the research questions. The process of fragmenting lifts the
coded pieces out of the context of the interview as a whole. This process of code-
and-retrieve was conducted using the WinMAX computer program for qualitative
analysis (Kuckartz, 1998). The latter activity accentuates the context and richness
of the data as the interview parts were interpreted as a whole and the pieces of one
case were connected (Sivesind, 1999).
From this empirical study a ﬁve-step analysis procedure was derived. We em-
phasize again that the number of steps as such is not important, because that
A PURPOSEFUL APPROACH TO CCM 395
depends on the kind of material that is involved. As our study involved dyads we
ended up with ﬁve steps:
1. Comparison within a single interview.
2. Comparison between interviews within the same group.
3. Comparison of interviews from different groups.
4. Comparison in pairs at the level of the couple.
5. Comparing couples.
It emerged that the comparisons made differed on four criteria, namely (1) the
data or material involved and the overall analysis activities; (2) the aim; (3) the
questions asked and (4) the results. The sections describing the ﬁve steps all follow
the structure of these criteria. Each step is summarized in Table I.
4.1. STEP 1: COMPARISON WITHIN A SINGLE INTERVIEW
At the start of the research the comparison is conducted within one interview. In
the process of open coding, every passage of the interview is studied to determine
what exactly has been said and to label each passage with an adequate code. By
comparing different parts of the interview, the consistency of the interview as a
whole is examined. For example, clariﬁcation is needed if an interviewee says that
he has no difﬁculty with the physical dependency caused by MS in one part of the
interview but indicates elsewhere in the interview that feeling dependent causes
If one fragment is given the label ‘dependence’, the researcher studies the inter-
view for other fragments that should be given the same code. If a reference is made
to the same category more than once in the course of an interview, the fragments
relating to this category are compared in order to ﬁnd out whether new informa-
tion about this category is given or whether the same information is repeated. The
fragments are then subjected to further comparison to ﬁnd out what they have in
common, how they differ, in what context the interviewee made the remarks and
which dimensions or aspects of dependence are highlighted.
The aim of this internal comparison in the context of the open coding process, is to
develop categories and to label them with the most appropriate codes. In this way
it is possible to formulate the core message of the interview with the codes that are
attached to it and to understand the interview including any difﬁculties, highlights
and inconsistencies. It represents an attempt to interpret the parts of the interview
in the context of the entire story as it has been told to us by the interviewee.
396 HENNIE BOEIJE
Table I. Different steps of the constant comparative analysis procedure in keywords
Type of comparison and Analysis activities Aim Questions Results
1. Comparison within a Open coding; Develop categories What is the core message of the interview? Summary of the interview;
single interview summarizing core of the interview; understanding How are different fragments related? Is the interview Provisionalcodes (code tree);
ﬁnding consensus on interpretation consistent? Are there contradictions? Conceptual proﬁle;
of fragments. What do fragments with the same code have in Extended memos.
2. Comparison between Axial coding; Conceptualization of the Is A talking about the same as B? What do both Expansion of code words until all
interviews within the formulating criteria for comparing subject interviews reveal about the category? relevant themes are covered;
same group that is interviews; produce a typology What combinations of concepts occur? What Description of concepts;
persons who share the hypothesizing about patterns and interpretations exist for this? Criteria for comparing interviews;
same experience types. What are the similarities and differences between Clusters of interviews (typology).
What criteria underlie this comparison?
3. Comparison of interviews Triangulating data sources. Complete the picture What does group 1 say about certain themes and what Veriﬁcation of provisional
from groups with enrich the information does group 2 haveto say about the same themes? knowledge of interviewees from
different perspectives but What themes appear in group 1 but not in group 2 group 1;
involved with the subject and vice versa? Additional information;
under study Why do they see things similarly or differently? Memos.
What nuances, details or new information does group
2 supply about group 1?
4. Comparison in pairs of Selecting themes from open coding Conceptualizationof What is the relationship like from both perspectives? (Conceptual) proﬁle of relationship;
interviews with two that concern the relationship; relationship issues Are ther e contradictions/agreementsbetween them? Extended memos;
partners belonging to a summarizing the relationship; understanding of the What are the central issues and how are they Inventory of central issues.
couple ﬁnding consensus on the interaction between partners resolved?
5. Comparing interviews Finding criteria to compare couples; Find criteria for mutual What are the typical differences between couples A Criteria for comparing couples;
with several couples hypothesizing about patterns and comparison and B? What is the possible reason for this? Clusters of relationships (typology).
types. produce a typology On which criteria can couples be compared?
What patterns exist in the relationships of couples
that experience this phenomenon?
A PURPOSEFUL APPROACH TO CCM 397
Important questions regarding comparison in this ﬁrst phase of analysis are:
−Which codes are used to label the categories in this particular interview? What
characteristics do fragments with the same code have in common?
−What is the core message of this interviewee?
−Is the storyline consistent? Are there any expressions that are contradictory?
How are all the fragments related?
The initial analysis generates a number of results. The ﬁrst result is a summary of
each interview. The second is a list of provisional codes (a code tree) which is the
beginning of the process of conceptualization. The third result is the distillation of
the interview into an inventory of provisional codes or a conceptual proﬁle. The
fourth result consists of memos which describe the analysis process.
4.2. STEP 2: COMPARISON BETWEEN INTERVIEWS WITHIN THE SAME GROUP
All new interviews conducted are treated as described above. Memos, codes and
codings increase as a result. As soon as more than one interview has been conduc-
ted, the interviews are compared. The comparison in this step is between interviews
within the same group, which means persons who share the same experience, in our
example persons who are having to cope with MS. The ﬁrst ﬁve interviews were
selected with a view to exploring the subject and obtaining a variety of research
subjects. Once more information had been obtained, patients were selected more
carefully in order to answer the questions raised by the comparison process. The
selection therefore became more closely tied to theoretical ideas and hypotheses
which of course were provisional and needed to be veriﬁed in other cases. Firstly, it
is important to compare fragments from different interviews that the researcher has
interpreted as dealing with the same theme and that have been given the same code
(axial coding). By making an inventory of characteristics of each category it was
possible to describe and even deﬁne some concepts, for instance what dependency
means in the context of MS. In the research, these ‘themes’ function as criteria for
the systematic comparison of the interviews. By comparing it becomes evident that
some interviews can be grouped together because they are similar with regard to
In many cases, some codes are combined with other codes and form a pattern.
Other interviews do not ﬁt this pattern. It is therefore important to look for patterns
or, in other words, for combinations of categories or codes. The analysis revealed
that interviewees coped with dependency in very different ways. To ﬁnd out why,
to identify the conditions that apply and the consequences that result, it is necessary
398 HENNIE BOEIJE
to compare interviews as a whole. This analysis was mainly done in the meetings
of the research team, where a typology was constructed.
The aim of this step is to further develop the conceptualization of the subject. To
this end, axial coding is used. This means searching for indicators and character-
istics for each concept in order to deﬁne that concept. A second aim is to discover
the combinations of codes which exist. This produces clusters or a typology. In our
example a typology of people with MS who deal with their disease in a particular
In this phase the following questions are of importance:
−Is interviewee A talking about the same category as B? What do both
interviews tell us about the category?
−What are the similarities and differences between interviews A, B, C ...?
−What are the criteria underlying this comparison?
−What combinations of codes/concepts occur? What interpretations exist for
This step results in an extension of the amount of codes (the code tree) until no
more codes are needed to cover all the various, relevant themes contained in the
interviews. The relevant parts of the interviews are those that say something about
the research questions. When all the coded segments that belong to a given code are
put together, it becomes possible to discover all the relevant characteristics of this
concept in a substantial ﬁeld and to describe the concept. The researcher then goes
about identifying the criteria on which some interviews differ from others. Together
these criteria constitute the dimensions on which a typology can be constructed.
These different dimensions are mostly governed by patterns or combinations of
codes. These combinations form proﬁles, clusters or types (see for the use of
typologies Layder, 1993: 137).
4.3. STEP 3: COMPARISON OF INTERVIEWS FROM DIFFERENT GROUPS
It is important to give data triangulation a central place in qualitative analysis
(Kimchi et al., 1991). In this third step, interviews from two different groups are
compared with regard to the experience of a speciﬁc phenomenon. In the MS study
this was done by interviewing the spouses of the MS patients. The interviews with
the spouses were conducted in the same way as the interviews with the patients,
A PURPOSEFUL APPROACH TO CCM 399
with a view to obtaining additional information from the perspective of the care
providing spouses as one of our research questions was concerned with this topic.
The interviews with the spouses were also used to arrive at a better understanding
of the patients’ perspectives and strategies. The care providing spouses were asked
to give their opinions on how their partners were dealing with their illness and the
answers were compared to those of the patients themselves. Spouses sometimes
gave different examples than the patients and touched on themes that placed the
patients’ experiences in a broader context.
The aim of comparing the interviews of persons with a speciﬁc experience with the
interviews of others who are involved but are not undergoing the experience them-
selves (e.g., patients and their spouses) is to complete the picture already obtained
and to enrich the information on the ﬁrst group, i.e., the patients. A comparison
between the interview of a patient and his or her spouse can also be used to validate
the patient’s story either conﬁrming the story or casting doubt on it.
The important questions posed in this step are:
−What does group 1 say about certain themes and what does group 2 have to
say about the same themes?
−Which themes appear in one group but not in the other group and vice versa?
−Why do both groups view issues similarly or differently?
−What nuances, additional detail or new information does the other group
supply about the group of our interest?
This comparison does not really provide a new directory of code words or a new
order. This step is aimed at deepening the insights and completing the information
about the group of patients which has been the focus of our interest until this point.
Naturally, new memos are made.
4.4. STEP 4: COMPARISON IN PAIRS AT THE LEVEL OF THE COUPLE
The comparison in this phase takes place in pairs. The difference between this step
and the preceding ones is the level of analysis. It concerns both partners belonging
to a couple so the level is dyadic. It may be clear that such a step can also involve
other dyads, like a parent and a child, a teacher and a pupil or an employer and
an employee. Accordingly, the codes from the open coding process in the ﬁrst
400 HENNIE BOEIJE
step are examined for relationship issues. In fact, open coding took place on rela-
tionship issues at that stage and memos were made about the interpretation of the
The interaction between the two individuals can be reconstructed from what
both partners say about each other and about their relationship. The comparison
produces insights into similarities and differences in perspective, how issues are
solved, how communication takes place, agreement or disagreement on relevant
themes, and so on. These ﬁndings with respect to content are conceptualized in the
same way as the themes found to be relevant in steps one and two. The relationship
seen from two different angles is more complex than the experience of one person.
So it is possible that saturation will not be reached until more couples are selected.
The aim of comparison in pairs is to ﬁnd information about issues concerning the
couple from both perspectives. This step aims at a greater knowledge about the
interactions between the partners and the constitution of the relationship.
The questions asked are:
−What is the relationship like from both perspectives? How can it be typiﬁed
−Which codes are used to cover the core issues?
−Are there contradictions between partners or do they agree with each other on
−What are the central issues the couple has to cope with and how are they
The results are comparable with those of the ﬁrst step. The ﬁrst result is a concep-
tual summary of the relationship. The second result provides a clue to the central
issues that couples have to deal with in a speciﬁc situation, i.e., facing a chronic
illness. These central issues can lead to criteria on which to compare couples. The
third result constitutes extended memos concentrating on the dyadic level instead
of the individual level.
4.5. STEP 5: COMPARING COUPLES
The ﬁnal stage of comparison is between couples who share the same experience.
This also takes place on the couple level and not on the individual level. Compar-
ison between couples can begin as soon as more than one couple is involved in the
A PURPOSEFUL APPROACH TO CCM 401
research. The comparison in this step is the most complex of all. First, it implies
interviews from two different perspectives, namely individuals who are involved
but do not share exactly the same experience, i.e., one of them ill and the other
providing care. Second, an image of the relationship emerges from what the two
individuals relate. Accordingly it is a form of aggregation but one that takes account
of the two stories being told. Third, even a comparison between only two couples
involves at least four interviews. This complicates the intellectual overview.
The aim is to further conceptualize the issues concerning the relationship and to
ﬁnd criteria for the mutual comparison of couples. These criteria follow from the
core issues couples apparently have to deal with and serve as some kind of a coding
scheme for the analysis of new couples. Finally, a model or pattern of relationships
can be developed by comparing the couples.
The important questions in this step are:
−What are the typical differences between couples A and B? What is the
possible reason for this? To what is it connected?
−On which criteria can couples be compared?
−What patterns exist in the relationships of couples that experience this
The result of this analysis step is an inventory of criteria for comparing the
couples. Another result is a pattern of couples. This pattern is based on ﬁnd-
ing the dimensions that can serve as criteria for discriminating between different
5. An Illustration of the Step by Step Approach
In this section the step by step approach is exempliﬁed with our empirical research
into people with multiple sclerosis (MS). Although we could have used several
couples to illustrate our point, we often use a single couple (4). We believe that
once the line of an interview is understood, the example is easier to comprehend.
The comparative activities in the ﬁrst step are conducted within one interview.
Therefore we start with looking at a fragment from an interview with a male patient
designated as P4. This man obviously has aims left that are worthwhile and he
organizes his life as well as he can in order to reach his goals. He has many people
helping him and who even take his place whenever he must withdraw. In this text
segment he tells us about his dependency caused by MS and how he and his family
402 HENNIE BOEIJE
deal with that. After presenting this segment some codes and codings are presen-
ted. When the different paragraphs within the segment are compared it becomes
clear that some deal with the same theme and are therefore labelled with the same
code, for example ‘coping strategy’ or ‘dependency’. One paragraph or even one
sentence can have different codes attached to it. Sometimes codes constitute two
elements: a main category and a speciﬁcation of that category. For instance ‘coping
strategy: ms not the upper hand’.
Section of an Interview with P4 on September 27th, 1997
P(atient): Euh, I have it [multiple sclerosis] now for twelve years, as far as
I know thus. And I have the progressive type, so it has gone very fast and I
also became dependent very fast. Well, how do I deal with it? I don’t know,
in an ordinary way I think. From the beginning and after we went through
the ﬁrst blow, we have said MS will never have the upper hand. Three people
sit around this table and there are four chairs and one chair is for MS. And
more place it is not going to get. And so it is, we live with MS, but as much
as possible we live our own lives. We go on holiday independently, weekends
away, or days, etc. But dependent on my physical situation of course. For my
wife it must be as free as possible and then it is a burden anyway. Because she
is constantly busy with me.
I(nterviewer): What do you mean when you say there are four chairs and one
of them is for MS?
P: MS has been given a place in our lives, but MS is not the head of our table.
We noticed that in some families MS rules their lives. And we don’t need
that. I am a member of committees and I go to work for two days a week.
Although I have someone to replace me when I can’t go, because I am not
such a dreamer that I think that I can always do everything. I must have the
possibility to withdraw without feeling guilty. But with everything I promise
to others I have to accept that I have to confer with my family whether I can
promise to do it or not because one of my family members has to accompany
I: What is that like for you?
P: Well, for me it is not that hard, because we are very open about it in this
family and also to outsiders. We understand each other very well. And my
family sometimes puts me in my place when I ask too much of them.
Some of the codes that are attached to these paragraphs are:
MS has been given a place in our lives, but MS is not the head of our table.
We noticed that in some families MS rules their lives. And we don’t need that.
CODES: integration; coping strategy: MS not the upper hand
But with everything I promise to others I have to accept that I have to confer
with my family whether I can promise to do it or not because one of my
A PURPOSEFUL APPROACH TO CCM 403
family members has to accompany me.
CODES: dependency; family: confer
Well, for me it is not that hard, because we are very open about it in this
family and also to outsiders. We understand each other very well. And my
family sometimes puts me in my place when I ask too much of them.
CODES: family: open communication; self-critical
Well, how do I deal with it? I don’t know, in an ordinary way I think. From
the beginning and after we went through the ﬁrst blow, we have said MS will
never have the upper hand. Three people sit around this table and there are
four chairs and one chair is for MS. And more place it is not going to get.
CODES: coping strategy: MS not the upper hand; normalizing
In order to extend the comparative analysis to other interviews in the same group,
step 2, new patients were selected. At the start of our study these were selected
in such a way that a relatively heterogeneous population was constituted. For in-
stance, both male and female patients were selected, because it was thought that
the question of whether the wife or the husband falls ill or becomes a care provider
might make a difference in terms of family roles. Among the important themes to
emerge were adaptation to the disease, integration into daily life and the perceived
support of the spouse. As a consequence we wanted to select patients who differed
with respect to these themes.
As an example we show some fragments of an interview with another male
patient, P25, who shows a sharp contrast with P4. P25 can not adapt his life to MS
in any way. Everything that made his life worthwhile is lost and he is unable to
make sense of his ‘crippled and inactive’ life. Parts of his story touch on themes
that are also present in the interview with P4. As a consequence they have been
given the same codes, but they are brought up by P25 in a very different context
Some Codes and Coded Segments of Interview Excepts P25
But I try to set that MS aside, so that it doesn’t become a milestone. But that is
difﬁcult. Every day I start at the same level and every day I make the transfer
from bed to chair. I always was an active person and now I cannot even walk
and I am tired all the time. I have trouble putting MS aside, because I am
confronted with it every time. Sometimes I am inclined to do nothing at all
anymore. I feel rebellious and the dominant feeling about the whole situation
is powerlessness. Yes, it’s a disaster because there is nothing that helps and I
CODES: coping strategy: to put MS away; dependency; rebellion; powerless-
404 HENNIE BOEIJE
Table II. Two perspectives of people with multiple sclerosis
Criteria MS will never have the I want the consequences of MS
for comparison upper hand (P4) to pass me by (P25)
Feelings about MS Conﬁdent, realistic Bitterness, powerless
Identity Integration of illness Loss of former identity
Coping strategies Normalizing, managing Hiding, banishing
Dealing with the family Openness, support Isolation, distraction
Dealing with help Organizing, control Rebellion, resistance
My wife is tired. I can see that and I almost feel guilty about that. But I
get rebellious from that. Every time something should be done here in the
house, it is she who has to do it. I have to ask everything and she always has
to consider me. All that asking, she must do everything and I have to disturb
her so often. And we dispute about that because she is not ill, I am ill. And I
have to deal with all that. It is not only the physical complaints, but this illness
involves many mental problems.
CODES: family: disputes; relationship: burden; rebellion; guilt
I talk as less as possible about my illness. Sometimes I do, but not every
day. My wife wants me to talk more about it, but it does not solve anything.
CODES: relationship: communication; coping strategy: to put MS away
The comparison of both interviews made clear which problems MS caused and
what the illness meant to the lives of the various participants in the study. Different
perspectives of MS-patients were reconstructed of which two are displayed in Table
II. The criteria on which the interviews are compared and grouped together are
displayed left in the table. Patient 4 ﬁts the perspective ‘MS will never have the
upper hand’ and patient 25 ﬁts the perspective ‘I want the consequences to pass me
by’. Of course more patients ﬁt into each of these patterns.
In step 3 of the constant comparative analysis interviews from two different
groups are compared with regard to the experience of a speciﬁc phenomenon. In
the MS study this took place by interviewing the spouses of the MS-patients. As
an example we show two fragments of patient 4 (P4) and his spouse (S4) who talk
about the same event. It is not necessarily a comparison between two partners, in
fact all spouses can add information to the group of patients for the sake of trian-
gulation. The important thing here is that by adding the spouses’ information we
got a deeper understanding of the patient’s adaptation to illness and deterioration
in the context of the adaptation process of the family as a whole.
A PURPOSEFUL APPROACH TO CCM 405
Fragment of interview P4
I need a chair in the shower to take a bath. This chair had to be changed. And
then a physiotherapist came to look at my situation and he said you don’t get
a chair, you need a stretcher. And I said no, I am not up to a stretcher yet. But
my wife nodded in agreement. My wife sees my problems earlier than I do. I
sit on that chair and I shake and my wife catches me and I don’t even notice.
I think that all ill people see themselves as better than their family members
Fragment of interview S4
Yes, he did not want a stretcher in the shower. He thought that a chair would
do. But a physiotherapist visited us and he said that a chair was not adequate
and that my husband almost fell from it. Most of the times my husband is
the one who accepts his deterioration earlier than I do. I have trouble with it.
That was the case with the electric wheelchair. I didn’t want him to have one
because in my eyes he seemed so far gone back. While he experienced it as a
liberation, because he could move outside the house on his own. I learnt from
that event that I should not prevent him from using aids because they can solve
In step 4 of the analysis process we want to compare in pairs at the level of the
couple. For this comparison it need to be two partners who are compared. One of
the relationship issues that appeared to be of importance was commitment to the
relationship and shared future plans of partners. Again P4 and S4 talk about this
issue and reveal the way they deal with this:
Fragment on Relationship Issue Commitment in Interview S4
You are married and you promised to share the good times and the bad times
and I am behind that idea. If your husband is afﬂicted with MS I think that
two things can happen: either you grow apart or you grow towards each other.
And that happened to us, we speak the same language. Of course we do things
together that no husband and wife do together. For example, I take him to the
toilet. In a normal family the man goes out to work all day and comes home
during the evening. We are together all day, we do everything together. But
this doesn’t mean that I don’t have a hard time now and then. I ﬁnd it very
burdensome, because I always have to be there for him.
Fragment of Relationship Issue Commitment and Future Plans in Interview P4
In the daytime my wife and I have become a real twofoldness. We have grown
towards each other. And that went so far, that my wife could not deal with it
and became over strained. Then I went to an activity centre and my wife began
406 HENNIE BOEIJE
voluntary work. And because she is doing something else then, she can’t be
concerned about me at the same time. We both went to different places and
now we got something to tell to each other when we come home.
Four themes challenged all couples, namely: communication and problem solving,
reducing the inﬂuence of the illness, commitment and future plans and ﬁnally re-
spect and appreciation. In step 5 these themes functioned as criteria to compare
the twenty couples. First we show one fragment from an interview with spouse
8 (S8), a male care-provider. From this fragment the difference with couple 4
becomes clear. Finally, we show a table that displays dimensions of the couples
on the different criteria.
Fragment on Relationship Issue Commitment and Future Plans in Interview S8
We made a deliberate choice that my wife is going to live on her own in
a special residence with professional help. More and more, we have grown
apart and live a life of our own. At a certain moment I think it has no excess
value to live together. She can’t do a lot of things anymore, so you don’t do
many things together. She sleeps a lot. In our family the fun disappeared and
I ﬁnd that hard to accept. I have tried to prevent splitting up, but not against
6. Summary and Discussion
The present ﬁndings demonstrate that during qualitative analysis there are many
moments of comparison. The term ’constant’ might be a slight exaggeration, but
comparison is at the heart of the analysis process just as Glaser and Strauss pointed
out in their discovery of and later elaborations on the grounded theory approach.
The step by step approach to the constant comparative method (CCM) described
here systematizes this method for analysing interview data. Going about CCM in a
purposeful way and reporting the researchers’ own experiences when implement-
ing the step by step approach, increases both the traceability and credibility of
researchers’ analysis in their qualitative studies. The example reinforces the point
that it is not necessary to compare everything with everything else, but that the
comparisons must be conducted according to a sound plan. A plan implies that the
researcher knows beforehand which comparative steps are needed in the analysis
regarding the elements that are compared, the aims, the questions asked and the
expected results of each step.
To answer the ﬁrst research question, how varied the comparative moments are,
it is found that the elements that are compared vary along different steps. When a
study involves dyads, like our example about married couples, there are different
data sources. This is reﬂected in the data that are used in the different comparative
steps, namely: one interview, more interviews within the same group, interviews
A PURPOSEFUL APPROACH TO CCM 407
Table III. Dimensions in dealing with relationship challenges by couples
Relationship challenges Dimensions in dealing with challenges by couples
Communication and problem solving Settling things together by Attempts at communication end
talking about it up in discord
Respect and appreciation Feelings of being competent in Feelings of incompetence in
relationship roles, intimacy relationship roles, criticism
Reducing inﬂuence of the illness MS does not dominate the Problems associated with the
relationship or their lives illness spill over into other areas
Commitment and future plans Partners develop a shared Partners have different views on
perspective on their situation their future and see themselves
and future go separate ways
408 HENNIE BOEIJE
from different groups, interviews that constitute a couple and interviews with dif-
ferent couples. These data sources can also be used in studies into other dyads, such
as parent and child, siblings or employer and employee. Accordingly, there are less
data resources available when the study involves only one group of people who
undergo the same experience from the same perspective. Accordingly, there are
even more data sources and more comparative moments when more perspectives
are involved, for example triads. The available data dictate the number of steps.
To answer the second research question, which aims the different comparis-
ons have, the aim and the results of the different steps differ much. Comparative
thinking contributes to the development of criteria to distinguish categories of data,
to the conceptualization of the ﬁeld under study and to the patterning of the data
that ultimately provides an answer to the research questions that are examined in a
particular study. Therefore premature results are needed, such as codes, conceptual
proﬁles, summaries, memos and provisional deﬁnitions which ultimately lead to
the ﬁnal results of the study.
The third and last research question was aimed at the relation between the
different types of comparison. The coherence between the types has to do with
their complexity. First of all, complexity is inﬂuenced by the amount of interviews
involved in the comparison. In the ﬁrst step only one interview is involved, while
other steps encompass as many interviews as are available. Second, complexity
is connected with the diversity of interviews involved. Step two is limited to in-
terviews from one group of people who share the same experience, while from
step three onwards, the interviews used involve people with different perspectives.
Third, complexity is linked with the level of analysis. Steps one to three are on the
individual level, while steps four and ﬁve are on the dyadic level.
It is important to notice that the steps do not form a linear process but can
be found in all the research phases and support the cyclical method in qualitat-
ive research. Every time a new interview is conducted, the comparison within a
single interview takes place and every time an interview with a partner completes
a couple, the comparison in pairs and the comparison of different couples can be
carried out. However, the simpler comparisons (step one within a single interview
and step four in pairs respectively) were carried out more intensively at the begin-
ning of the study and the more complicated steps which involve more and more
diverse interviews (steps two and ﬁve) were extended in later phases.
Finally a word about creativity and theoretical sensitivity. The procedure for
CCM described here focuses on comparisons between interviews conducted to
answer research questions. External data that have not been produced for this spe-
ciﬁc purpose, such as ﬁlms, books or photographs, are highly heterogeneous and
therefore less likely to be inﬂuenced by the researcher. Excluding these sources
from the step by step procedure does not mean that the information and know-
ledge they provide should be ignored. Instead we admit that these comparisons
and occurrences have a very important part to play in research and especially in
qualitative research. There is no procedure or prescription for identifying exactly
A PURPOSEFUL APPROACH TO CCM 409
which comparisons are a fertile source for developing categories and a theoret-
ical model. Making the right choices in this process depends on the creativity,
experience, knowledge, talents, support and sensitivity of the researcher. In this
respect, comparisons go hand-in-hand with interpretation and it is our opinion that
purposeful comparison makes the valuable task of interpreting social phenomena a
much easier proposition.
Boeije, H. R., Bromberger, N., Duijnstee, M. S. H., Grypdonck, M. H. F. & Pool, A. (1999). In relatie
tot MS. Zorgafhankelijke mensen met Multiple Sclerose en hun partners. Utrecht: NIZW.
Chenitz, W. C. & Swanson, J. M. (eds.) (1986). From Practice to Grounded Theory: Qualitative
Research in Nursing. Menlo Park: Addison-Wesley.
Denzin, N. K. & Lincoln, Y. S. (eds.) (1994). Handbook of Qualitative Research. Thousand Oaks:
Dey, I. (1993). Qualitative Data Analysis. A User Friendly Guide for Social Scientists. London &
New York: Routledge.
Glaser, B. G. & Strauss, A. L. (1965). Awareness of Dying. Chicago: Aldine Publishers.
Glaser, B. G. & Strauss, A. L. (1967). The Discovery of Grounded Theory: Strategies for Qualitative
Research. Chicago: Aldine.
Glaser, B. G. (1992). Emergence vs. Forcing. Basics of Grounded Theory Analysis. Mill Valley:
Kimchi, J., Polivka, B. & Stevenson, J. S. (1991). Triangulation. Operational deﬁnitions. Nursing
Research 40: 120–123.
Kuckartz, U. (1998). WinMAX 97. Scientiﬁc Text Analysis for the Social Sciences. Berlin.
Layder, D. (1993). New Strategies in Social Research. An Introduction and Guide. Cambridge: Polity.
Miles, M. B. & Huberman, A. M. (1984). Qualitative Data Analysis: An Expanded Sourcebook.
Beverly Hills: Sage.
Morse, J. M. & Field, P. A. (1998). Nursing Research of Qualitative Approaches. Cheltenham:
Schwandt, T. A. (1997). Qualitative Inquiry. A Dictionary of Terms. Thousand Oaks: Sage.
Sivesind, K. H. (1999). Structured, qualitative comparison. Quality and Quantity 33: 361–380.
Strauss, A. L. (1987). Qualitative Analysis for Social Scientists. Cambridge: Cambridge University
Strauss, A. & Corbin, J. (1998). Basics of Qualitative Research. Techniques and Procedures for
Developing Grounded Theory. London: Sage.
Tesch, R. (1990). Qualitative Research. Analysis Types and Software. London: Falmer press.
Wester, F. (1995). Strategieën voor kwalitatief onderzoek. Muiderberg: Coutinho.