This guide provides an overview for using MAXQDA with focus groups, and how Matthew H.

Loxton
it can support the collection, analyxsis, and reporting of focus group data:

• Preparing and importing focus group data in MAXQDA

• Creating and importing speaker variables
• Exploring the focus group data
Analyzing Focus Groups
with MAXQDA
• Focus group coding
• What comes after coding?
• Tips and tricks incl. checkli
sts
Real-world examples from healthcare illustrate the explanations. A checklist for
analysis of focus groups with MAXQDA as well as a checklist for conducting
focus groups support the realization of your projects.

Matthew Loxton is a Principal Analyst at Whitney, Bradley, and Brown Inc.

focused on healthcare improvement and strategic foresight, serves on the board
of directors of the Blue Faery Liver Cancer Association, and holds a master’s
degree in Knowledge Management from the University of Canberra.

MAXQDA Guide
MAXQDA
PRESS PRESS
[Link]
Citation: Loxton, Matthew H. (2021). Analyzing focus groups with MAXQDA. Berlin: MAXQDA Press.

ISBN 978-3-948768-06-5
[Link]

© MAXQDA Press, Berlin 2021, 1st edition

MAXQDA Press is a division of VERBI Software. Consult. Sozialforschung. GmbH

Invalidenstr. 74, 10557 Berlin, Germany, info@[Link], [Link]

All rights reserved, in particular the right of reproduction and distribution as well as translation. No part of this work may be
reproduced in any form (by photocopy, microfilm or any other method) or processed, duplicated or distributed using electronic
systems without written permission by the publisher.

MAXQDA is a registered trademark of VERBI Software. Consult. Sozialforschung. GmbH. All other trademarks or registered tra-
demarks are the property of their respective owners and may be registered in the United States and/or other jurisdictions.

Publisher and authors have compiled the information in this book to the best of their knowledge. They give no warranty for the
correctness and assume no liability for the use of the information.
Contents

Introduction ........................................................................................................................ 4
The traditional focus group.............................................................................................. 6
The nominal focus group ................................................................................................. 6
Data and analysis considerations .................................................................................... 7

2 Preparing and importing focus group data in MAXQDA ............................................. 7

3 Creating and importing speaker variables................................................................. 10

4 Exploring the focus group data ................................................................................... 12

5 Focus group coding ..................................................................................................... 16

Memos ............................................................................................................................. 17
Paraphrases ..................................................................................................................... 18
Pre-designed code systems ............................................................................................ 19
Bulk coding ...................................................................................................................... 22
Thematic coding ............................................................................................................. 23
Grounded coding ............................................................................................................ 24

6 What comes after coding? ........................................................................................... 25

Using summaries ............................................................................................................ 25
Codeline ........................................................................................................................... 26
Crosstab ........................................................................................................................... 27
Code Matrix Browser ...................................................................................................... 28
Interactive Quote Matrix ................................................................................................ 28
Developing a report ........................................................................................................ 29

7 Tips and tricks ............................................................................................................. 31

MAXQDA checklist ............................................................................................................ 32

Checklist for conducting focus groups ............................................................................ 33

Bibliography...................................................................................................................... 35
4 Introduction

Introduction
This guide provides an overview for using MAXQDA with focus groups, and how it can support the collection,
analysis, and reporting of focus group data.1 The guide will cover some of the theoretical underpinnings, be-
nefits, and risks of focus groups. The guide was written in the context of healthcare, but the theory and me-
thods described apply generally to other environments. The guide will draw on examples from past
healthcare focus groups that mainly dealt with process improvement related to policy implementation, tech-
nology deployment, and workflow optimization.
Popularized by Merton at the US Bureau of Applied Social Research at Columbia University in 1946 (Lee,
2010), and possibly first identified as a method by Bogardus in the 1920’s (Jackson, 1998), focus groups have
been used extensively in varied applications, such as marketing, public relations, political campaigns, pro-
duct design, quality management, and computer user experience and interface design. Focus group studies
have enjoyed broad adoption as a relatively low-cost and moderately effective means to explore open-ended
questions with small groups of people selected to represent some target population. Focus groups are often
used to test ideas, products, concepts, or scenarios, elicit reactions and sentiments, and spark innovation.
Due to the high level of interaction between members, it is commonplace to refer to them as participants
rather than “subjects” or “interviewees”. The focus group environment can allow participants to build on
each other’s responses in a generative manner that may allow discovery beyond the ability of individual in-
terviews or surveys. While focus group participants are likely to be selected to match a specific profile relevant
to the topic, their input often yields unexpected facts, and leads to unanticipated discoveries (Schultz,
Schultz, & Schultz, 1994). As Bogardus stated in the 1926 paper that defined focus groups, the group environ-
ment “brings out points that otherwise would remain obscure”. Likewise, Bogardus notes that the group dy-
namic may result in new points that “probably would not have been secured in a personal interview”
(Bogardus, 1926, p. 372).
Although there is a high degree of interaction between participants, the questions used in focus groups
are often tightly scripted, and focus groups are often therefore somewhat towards the more structured end of
the qualitative spectrum. In terms of Creswell’s five qualitative inquiry traditions (Creswell & Poth, 2018),
focus groups may be said to be less (but not zero) grounded theory, and more narrative and ethnographic
content. A focus group traditionally sets up to 12 topics or questions (Krueger, 2002), but it is not unknown

1 Some of the text and ideas in this guide are based on the research example in Loxton (2021).
Introduction 5

for topics to be dropped or radically shifted based on participant inputs. For example, a nursing focus group
set to discuss bed linen preferences may radically diverge from the set topic if a participant observes that bed
cleaning is not adequately controlling the spread of a highly contagious disease. It would be a very poor mo-
derator who insists that the question of bed linen thread count remain the focus in such an eventuality.
In terms of typical focus group size, analysis of 220 papers published in 117 journals suggested that the
median participant count is approximately 5, and ranged from 1 to 96 (Carlsen & Glenton, 2011). Opinions
on optimal size vary within a low range, for example 7–15 (Van Bennekom, 2002) and 8–12 (Robson, 2011),
while another position is that the size should vary according to the background population and purpose; 10–
12 for marketing purposes, and 5–8 for non-commercial groups (Krueger & Casey, 2009).
Whatever size of focus group you use, you are likely to be analyzing data that represents not just a set of
single interviewees, but the interaction of several participants. This level of interaction and participation has
data and functionality implications for your project, methods, and qualitative data analysis (QDA) tool needs.
MAXQDA offers a wide array of integrated features that help to make analysis of focus group data seamless
and efficient. The full array of MAXQDA’s rich functionality for analysis and reporting data can be used on
focus group data, including metadata variables, analysis and mixed methods tools, visual tools, reporting,
and statistical analysis and comparison tools. There are, however, some MAXQDA features that are specifi-
cally designed for focus group data, or which greatly lend themselves to the needs of focus group analysis.
There are some aspects of focus group data that are not typical of other kinds of research. As described
above, members of a focus group (perhaps especially so in healthcare) behave more like participants than
interviewees, and this element is sufficiently different to other kinds of research interaction to warrant spe-
cialized features. Focus groups have been described as collaborative co-discovery sessions, and are perhaps
most similar to participatory design, generative brainstorming, and facilitated card sorting. Focus group mo-
derators may use specialized techniques such as storyboarding (Cross & Warwick-Booth, 2016) when
conducting participatory research to increase participation and enhance data collection, and moderators
may themselves be active participants at times. Focus group sessions may include the use of discussion
prompts such as product examples, images, audio tracks, or video clips. Some may involve walking through
a location, such as a ward floor, surgery, or other medical environment. These artifacts themselves require
capture and analysis within the QDA.
We may also use the MAXQDA focus group functionality in ways that stratify degrees of focus group struc-
ture. Going back to the initial description by Bogardus of a “group interview” (Bogardus, 1926), we see two
levels of group interview: traditional, and nominal or pseudo focus groups.
6 Introduction

The traditional focus group

The traditional focus group presumes a group setting in which members have direct interaction with each
other and participate in a generative process in which one participant’s input might be built upon another.
This assumption of strong interaction and role of the agents as participants, rather than subjects, is a key
feature of the traditional focus group. As a result, each participants’ text needs to be coded both as an indivi-
dual and in context of their interaction with others.
A risk with traditional focus groups is that some participants may unduly influence others, and the
presence of observers, bosses, or outsiders as participants may inhibit the free flow or expressions of o-
pinions, experiences, or ideas of others (Van Bennekom, 2002). The risk of influence by one participant on
another thus suggests the need for an active moderator, and hence a need to capture not just participant
interaction, but also to record, identify, and analyze the interaction of the moderator qua participant in our
data. For the traditional focus group, we may therefore wish to code interaction types that existed between
the participants and between participants and the moderator.
Focus groups may also prove to be logistically difficult—the moderator must find suitable locations for
group discussions, and must situate the sessions in places and at times that are accessible to enough partici-
pants to have a functional focus group. These logistical burdens often prove to be practically insurmountable.

The nominal focus group

Borrowing from the concept of the nominal group technique (Gallagher, Hares, Spencer, & Bradshaw, 2017),
the nominal focus group reduces the risk of undue influence from other participants or observers, and un-
tangles the logistics, by having no formal group at all.
In this approach, we can select individuals who have a shared weltanschauung, traditions, or occupational
environment that may result in a latent sharing or exchange of beliefs, opinions, and assumptions. Members
of the nominal focus group are not ever in the same session and do not participate with each other in a group
interview setting, but there is a reasonable supposition that they are in direct or indirect communication as
part of their shared characteristics. Putting all the transcripts together in a single document allows us to use
MAXQDA’s Focus Groups Transcripts import functions and treat them as if they had been in a common ses-
sion.
For example, I might select several nurses from different wards in the same hospital and form a nominal
group. They will not have been interviewed in the same session, but there is a strong likelihood that they share
enough occupational experiences and commonality of work that they can be treated as nominally belonging
Preparing and importing focus group data in MAXQDA 7

to a group. The various responses can be grouped by question, as if they had been in the same session, and
had responded to each question in turn.
The nominal focus group gives the researcher considerable freedom and scope, and in principle allows
the assembly of a group composed of people who may be from different times, or even entirely fictional cha-
racters. From a data perspective, the nominal focus group implies a need to capture the selection process and
argument as part of the project data.

Data and analysis considerations

There are some general considerations about types of data and analysis options that are important for work
with focus groups. From the descriptions of the traditional and nominal focus groups above, several data
types may be noted.

1. Transcripts of sessions
2. Associated audio and video recordings of the sessions
3. General session notes captured by the moderator(s)
4. Participant interaction notes captured by a moderator
5. Audio, image, or video artifacts used as prompts
6. Group activities such as walkthroughs, tours, or hands-on exploration used during the sessions as
prompts
7. Planning and logistical notes
8. Data related to characteristics of the participants, such as age, gender, role, position, salary, race, etc.
9. Data related to the interview time or setting

2 Preparing and importing focus group data in MAXQDA

The advice of many experienced researchers is often that one should design the data collection plan with the
final product in mind. This is wise advice, but researchers often receive data over which they had no control,
or the collection environment demands other priorities than what would most suit analysis. Another valuable
piece of advice is to tailor the data acquisition to the methodological approach (Creswell & Poth, 2018).
MAXQDA provides wide flexibility in data import and capture. Whether the data are transcripts with or wit-
hout timestamps, or audio or video clips, MAXQDA has import and transcription tools to suit the situation.
8 Preparing and importing focus group data in MAXQDA

For focus group data, specifically, MAXQDA allows the researcher to import transcripts with or without
timestamps, but more crucially, allows speaker changes to be easily specified in the text. If the import func-
tion for focus groups transcripts encounters a “:” symbol in the first 63 characters of a paragraph, it assumes
the preceding text is the name of a speaker, and the subsequent text is what they said. MAXQDA then creates
a code with the name of the speaker, and auto-codes the text with that code. Figure 1 shows an example of an
edited transcript in a Word document. The participant names and questions have been formatted in boldface
for readability, but this has no implications for the import process.

Interviewer 1: Hi. So just some background. Like I said before, we are working with various primary care stakeholders to
help set up and facilitate planning and workflow development sessions for activating the Ambulatory Care Center (ACC) in
October. XXXXX suggested that you would all be able to help us to get some detailed insights on the patient and staff flow
aspects of current day to day operations.
Interviewer 1: Can we start with introductions? Can you describe your role and involvement in the project?
Participant 1: I am in PCMH, only one of a few now. Am one of the unit managers team leader for unit 1. One unit is offsite,
and I am unsure how their workflow is done. Right now, each unit is under a nurse manager, and I have been working in a
nursing home for 2-3 years, as well as the OR, doing med-surg. Last 4 years I was in primary care.
Unit 1 has three care managers, and we are short staffed. Have 8-9 doctors but they are part time. Two medical residents.
Only clinic operating on Saturday, but that is the NP.
Participant 2: Hi
Interviewer 1: Hi What is your role, and how long have you been here?
Participant 2: Hi, I’ve been at XXXXXXXXX for 12 yrs. I currently coordinate the mental health program and intake using the
biopsychosocial model. I also cover crisis care management. I was recently taken on as acting supervisory psychology role,
but I wasn’t previously directly involved in the care integration team
Participant 3: Sorry. Can you repeat that first bit again? Is this to do with the flow workshop?
Interviewer 1: Sure! Hi, yes, we are working with XXXXXXXXXX to help set up and facilitate the workshop next week for
planning and workflow development for the new facility, especially the integration of primary care and mental health in
the new facility. XXXXXXXX suggested that you would be a good person to provide us with some insights on the patient and
staff flow aspects of current day to day operation.

Figure 1: Edited transcript

To import a transcript, we can click on Import (Figure 2, Item 1) on the ribbon bar, and then click Focus
Group Transcripts (2) to obtain a list of import options. Since this transcript does not contain timestamps,
we click on Transcripts without Timestamps (3).
Preparing and importing focus group data in MAXQDA 9

Figure 2: Transcript Import

Figure 3 shows the results of importing the same Word document transcript. The count of coded segments
(participant contributions) is reflected in a number to the right of the code.

Note: Because of the relationship of the focus group codes created by the import process and the associated
transcription documents, the codes cannot be manually deleted. However, if the transcription document is
deleted, all associated participant codes will also be deleted. This is also important to note because if there
are duplicate entries for a participant, perhaps due to varied spelling in the transcript, the codes cannot be
merged either through the code context menu options, or in the Creative Coding function that would be pos-
sible with other codes. It is therefore important to verify participant code integrity immediately after impor-
ting.
10 Creating and importing speaker variables

Figure 3: Focus group transcript after import in a MAXQDA project

MAXQDA also supports direct import of transcripts and data from a number of industry applications and
services, including several using machine learning to generate automated transcripts. In version 20.3.0,
MAXQDA supported direct import from 15 transcription products, including AmberScript, f4/f5, and
Transcriber Pro. Some transcription tools may automatically insert colons after each speaker name, but it is
still wise to check that no extraneous colons exist in the text prior to importing.
If transcripts contain timestamps, these will be removed from the text by MAXQDA and converted to ti-
mestamp icons left to the text once the associated media file has been assigned to it. If no transcript is
available, the full features of the inbuilt MAXQDA transcription tools are available to you, i.e., you can import
an audio or video recording of a focus group into a project and transcribe it directly in MAXQDA.

3 Creating and importing speaker variables

Similar to how document and code variables are handled in MAXQDA for other data types, the Variables tab
in the ribbon menu has a function to provide a List of Speaker Variables, as well as a Data Editor for Speaker
Variables. As the names imply, the former is a quick way to list all existing user-defined variables or create
new variables. The latter function reveals the values of the variables associated with each speaker. It is pos-
Creating and importing speaker variables 11

sible to import and export speaker variables using an Excel Sheet. Typically, desired speaker variables such
as gender, employment level, job title, etc. may come from a human resources application, or other external
source. The Import Speaker Variable function can save time compared to manually entering them in
MAXQDA. To import, first create a single user-defined variable, export as an Excel sheet, merge any external
data into the Excel sheet, then import into MAXQDA. Figure 4 shows the List of Variables pane superimposed
on the Data Editor after variables were imported for Facility and Gender.

Figure 4: Focus group speaker variables

The ability to compare participants according to a range of variables is very powerful, but must be done with
due caution in the context of low sample size typically involved in focus groups. For example, since hospitals
are often very hierarchical places of work, it is very instructive to view focus group results by comparing dif-
ferent role level, gender, and specialty, but one cannot draw any statistically significant conclusions based on
the small number of people selected to participate in the focus group.
In addition to being able to import speaker variables, MAXQDA allows the researcher to Export Speaker
Variables for use in external software or as part of a report. The Speaker Variable Statistics function
(available in the Variables ribbon tab but also accessible from within the Data Editor) shows tabular or gra-
phic frequencies or percentages for all speakers in the project or for the activated ones.
12 Exploring the focus group data

4 Exploring the focus group data

After importing focus group data, it is prudent to begin by reviewing the “Document System” and “Code Sys-
tem” to ensure all files were imported, all participants and moderators were represented, and no unexpected
participants were listed. Unexpected items in the “Document System” or “Code System” are typically the
result of errant colons in the first 63 characters of a line in the transcription text, or variation in the spelling
of participant names. The easiest corrective action is often to make a note of all the unexpected items, and
then before doing any other work in MAXQDA, use the Undo (Ctrl+Z/cmd+Z) function to back out of the
import, and correct the issues in the source transcription text. Failure to correct at this stage may result in the
need to take more complex remedial actions later on.
Once satisfied that all and only the expected participants have a code in the “Document System” and the
“Code System”, it is useful to review each of the transcripts to establish that they are intact and complete. By
double-clicking a document’s name (in the example in Figure 3, “Group 1 Session 1”), the full transcript is
listed in the “Document Browser” pane. Double-clicking a participant code displays the “Coded Segments”
pane listing, and a preview of all the coded segments representing what that participant said (Figure 5). Cli-
cking on any of the rows in the “Coded Segments” pane results in a preview of the segment text, as well as
highlighting the associated text in the “Document Browser”.

Figure 5: Participant’s coded segments

As with many other results panes, the “Coded Segments” pane also enables the use of autocoding. The au-
tocode feature is covered in more detail in the next chapter’s “Bulk coding” section.
Exploring the focus group data 13

Right-clicking the “Document Group” results in the familiar context menu, which now has an additional
entry, Focus Group Speakers. Clicking on “Focus Group Speakers” results in an overview pane (Figure 6). The
content of the pane can be filtered by clicking on the Only activated focus group speakers icon, or by using
the Set Filter option available in the context menu of a column’s heading. The table provides an overview of
the number of coded segments and coverage associated with each participant, and is an early indicator of
whether one or more participants were outliers, and either dominated the discourse or were mostly silent.
Similar to other MAXQDA table overviews, the results can be viewed in numerical values or as graphs, and
can be exported in various formats. The table also offers the ability to view existing or adding new user vari-
ables. In this example, the variable “Role” has been added to distinguish between the major work divisions at
the hospital. Variables can also be accessed by going to the Variables section in the ribbon bar, and clicking
the List of Speaker Variables, or the Data Editor for Speaker Variables icons.

Figure 6: Overview of Focus Group Speakers

If participant codes are assigned different colors, a similar perspective on contributions can be obtained by
using the Document Portrait function in the context menu of a focus group transcript in the “Document
System” (Figure 7). In that case, grouping the participant codes by color can quickly reveal if participant
groups or individuals tended to be outliers in their amount or distribution of contributions. This is particu-
larly useful in detecting if the moderators tended to over-talk, but also helps to see if some participants speak
more or less frequently than others, or at different times than others. Hovering the mouse over any specific
tile will yield a short preview of the text and some basic descriptive statistics.
The portrait view in this example shows some expected results, such as that the moderators (light blue)
started the dialogue, are scattered throughout, and are not overly represented. However, it also alerts us to
the fact that the sessions aren’t terminated with the moderator—no light blue at the end. This is an example
of not seeing an expected element in the visualization, which prompted us to query if text was cut off. Clicking
14 Exploring the focus group data

on the Ordered by color frequency icon yields a further revelation—Participant 5 (purple) contributed almost
nothing (Figure 7). Clicking on the purple tiles takes the “Document Browser” view to the associated coded
segment, and in this example, reveals that Participant 5 quit the sessions shortly after it started, and was lost
to contact. The Document Portrait gives a quick, but highly effective overview of the data, and highlights po-
tential issues upfront.

Figure 7: Document Portrait for one focus group; ordered by contribution sequence (left) and ordered by participants, i.e.,
by color (right)

A further data orientation feature in the context menu of a focus group transcript is the Word Cloud (Visual
Tools > Word Cloud), which in conjunction with a stop list gives a tabular, and a graphic view of the most
frequent words in the transcript. To view a Word Cloud specific to only participants or only some participants
selective activation and the “Retrieved Segments” pane can be used in conjunction. To have a Word Cloud
for only specified speakers in a focus group, activate the associated focus group documents in the “Document
System” and their focus group speaker names in the “Code System”. This will result in the “Retrieved Seg-
ments” window containing only their contributions. Then click the Word Cloud icon in the head of “Ret-
rieved Segments” window to see a Word Cloud with only the words from the selected speakers.
MAXQDA users with access to the MAXDictio suite of functions are able to request frequencies of multi-
word phrases (MAXDictio > Word Combinations). The frequency of words and phrases is a good initial indi-
cation of what was most salient to the participants, and perhaps an early hint at elements of the coding re-
quired. It is also useful to compare word or phrase clouds of different strata of participants. For example, did
Exploring the focus group data 15

the participants from one facility, or one role, or one gender use different terms to another grouping, or use
terms more or less frequently than another?
The Keyword in Context (KWIC) function, also available on the MAXDictio ribbon tab, retrieves text pre-
ceding or trailing a specified keyword. This function allows you to look for keywords in the context of
surrounding text and therefore notice patterns and nuances that may otherwise have escaped attention.
Figure 8 shows instances of occurrence of either “handoff” or “hand-off” across all participants. For this
focus group, the KWIC raised three points for further exploration—firstly, it was unexpected that participants
would refer to both “warm” and “hot” handoff of patients. “Hot-handoff” was an expected result of recent
changes and a defined terminology, but there was no such thing as a “warm” handoff in the project definition.
Noticing the use of “warm-handoff” as a term helped identify that clinicians were not always able to achieve
the desired handoff, but counted as “warm” those that “almost satisfied” the requirements. Secondly, since
the handoff was a critical success factor, eight occurrences was unexpectedly low. Thirdly, the clinician who
used the qualifier “when it happens” raised further off-line questions as to what was causing handoff failure.

Figure 8: Keyword-in-context (KWIC) function

16 Focus group coding

5 Focus group coding

Once an orientation had been achieved by looking through the transcript, viewing the Document Portrait,
and examining word, phrase, and KWIC results, the more fine-grained task of analysis can start. Focus group
transcripts can be read like any other interview or document, and the researcher can use memos, summaries,
and paraphrases to annotate, interpret, or demarcate the speaker contributions. What may be different in
addressing focus group data is the need to note interactions between participants.
One practical approach to focus group coding is to view the coding requirements in terms of analytical levels.
Kuckartz and Rädiker (2019) propose the following four analytical levels:

 Topic
 Participant
 Group Interaction
 Moderator

There are three basic approaches in reading focus group transcripts for the purposes of coding, depending
on the research focus. It is possible to use one, a combination, or all the methods in conjunction.

1. Read the full transcripts in order of occurrence. This gives an overall perspective of the sequence, and may
help to identify any maturing or shift in how topics were presented, or how the moderators may have
adapted over time.
2. Read the combined transcripts for each topic. In many projects, the same topics or questions are posed
to several groups over a period of time. Reading by topic, gives a strong narrative perspective of the group
responses to the topic.
3. Read the transcripts by participant. This approach gives a perspective of the contributions of a single par-
ticipant at a time, and allows the researcher to gain closer understanding of themes, habits, or styles of
each participant that may otherwise have been lost in the interaction.

MAXQDA provides an easy way to see all the contributions of an individual participant. Right-clicking on any
specific participant code in the “Document System”, will result in a context menu with the Overview of Con-
tributions option. The resulting Coded Segments pane provides a listing of all contributions for that partici-
pant, and has all the familiar options to select some or all contributions and code them with another existing
code, a new code, or to export the list in several formats. Figure 9 shows an example of this process and the
resulting Coded Segments pane.
Focus group coding 17

Figure 9: Overview of Contributions for a single speaker (“Participant 1”)

Memos
As with other types of research, memos are often added prior to coding, and used as a vehicle to develop
codes. For focus group coding, in-document memos are an excellent place to store specific moderator or
observer notes related to a segment. This may include any notes about reactions that one participant had to
the speech or actions of another, or moderator behavior that may have influenced participants. For example,
in one session, a nurse clapped her hands to her face when another spoke of a near-miss due to confusion
between two similar-sounding but very different medications. This would be very difficult to capture in a
transcript, but if it appears in moderator notes, it can be attached to the relevant segment as an in-document
memo.
Likewise, the in-document memo can be used to record events, such as the details of a prompt, or a
walkthrough. Memos can also record moderator or observer notes to themselves, such as “Participant-2 rol-
led her eyes when the topic of the ePortal was raised, and I am unsure what this meant. We did not get another
chance to ask her to elaborate”. Additionally, you can use the in-document memos as records of possible
code suggestions by different team members. Memos that contain follow-up suggestions or questions can be
given the “?” memo label.
18 Focus group coding

Figure 10 shows examples of recording events, moderator notes, and code or project suggestions with in-
document memos.

Figure 10: Use of memos

Paraphrases
Paraphrases accurately, but concisely state the meaning of a particular text. This may be important in focus
group settings when one or more participants interrupt or interject when another is speaking. In these cases,
it may be necessary to piece together a participant’s full contribution in order to present it in its full and
uninterrupted form. It can be a challenge because a contribution may spread across several paragraphs and
interleave with other speakers. An effective approach is to construct the precis and attach it to the partici-
pant’s initial text where they first started a train of thought.
This re-ordering cannot be done in the transcript itself without destroying the sequence and interaction,
so it is best done in a paraphrase (Figure 11).
Focus group coding 19

The Paraphrase Document function set can be invoked either by going to the Analysis tab in the ribbon
bar and clicking on Paraphrase, or by using Ctrl/cmd+Shift+P. The paraphrase tools include options to ca-
tegorize existing paraphrases, view a paraphrase matrix, or print the current document including paraphra-
ses. The Categorize Paraphrases function (also available at Analysis > Paraphrase) is especially useful for
developing new codes because it provides a side-by-side view of original vs paraphrase text.

Figure 11: Capturing paraphrases in the “Document Browser”

Pre-designed code systems

We often use existing code systems specifically designed for focus group analysis, or which are institutional
or traditional to the healthcare industry. For example, the group interaction coding system developed by
Morgan and Hoffman (2018) is specifically designed for focus groups to address the interaction between par-
ticipants. In Figure 12, there is an example of the Morgan and Hoffman coding system, that has been slightly
adapted for use in MAXQDA. Existing code systems can be imported using the Import Codes and Memos
from Excel Spreadsheet function in the Codes menu.
20 Focus group coding

Figure 12: Morgan and Hoffman (2018) group interaction code system

What can perhaps improve the Morgan and Hoffman coding structure, is a way to indicate the directionality
of interaction, and researchers may wish to apply codes to indicate directionality of group interaction that is
specific to their study topic. For example, it may be important to note whether it is a male interrupting or
supporting a female, or whether doctors support nurses.
There are many possible approaches to coding directionality, but here are three that you may consider:

1. Create a family of “directionality” codes specific to the context. For example, if I want to depict male vs
female directionality, I might have the following codes “M->F”, “F->M”, “F->F”, and “M->M”, and then
code any segment reflecting a directional action or speech act. Used in conjunction with the Morgan &
Hoffman Group Interaction system, the directionality codes applied to the same segment could denote,
for example, that a male disagreed with a female.
2. Another approach using codes and variables, is to create a code family for “Target of Action” and duplicate
all the participants and moderators as sub-codes. In this case in a speech act by “Participant-1”, we might
code the segment as being “Support”, and code it with the “Target of Action” as “Participant-2” to show
who did the supporting, and who received the support. The participant variables could contain biogra-
phical data such as gender, and therefore allow us to show by gender, who was supporting whom by using
Visual Tools > Code Relations Browser.
Focus group coding 21

3. A third option is to use the Edit comment function to add a text to the interaction code saying who
addressed whom, e.g., by right-clicking on a coding stripe in the “Document Browser”.

Some code systems are typical to an industry. For example, in healthcare, it is common to have a number of
codes specifically related to safety that will always be applied to any interview or focus group transcript in
addition to any codes developed for a specific study purpose or derived through a grounded theory approach.
Such “institutional code systems” may be applied more easily through the use of standardized search terms
as described in the next section.
The research sponsor may also have posed specific questions, or desire the coding to be done to suit spe-
cific interests. The questions or interests can be operationalized as a code system. In some cases, the resulting
code system may have only one level, and in others there may be parent codes, with as many levels as are
required to exhaust the constructs. For example, in Figure 13, the combined code system contains a parent
code for the topics that were set by a project sponsor, a parent code for group interaction, and a parent code
for codes developed using an inductive approach inspired by grounded theory. Each interview topic is cap-
tured in the code system as subcodes to the parent code of “Sponsor Topics”, and the sponsor subcode for
Shift Scheduling has a fourth level of coding for the responsibility and any issues related to scheduling.

Figure 13 Combined Code System

22 Focus group coding

Bulk coding
The Lexical Search tool (available in the Analysis ribbon tab) can be used to search within the transcripts,
and to bulk-code segments in focus group transcripts where there are pre-determined codes. This feature
saves time when applying codes that relate to searchable constructs. To help save time and reduce inter-
coder variance, complex search strings can be saved and reused. The researcher can apply any institutional
coding relatively quickly, and thus not detract too much from the core focus of a specific focus group project.
For example, Figure 14 shows the Lexical Search dialog box loaded with a stored search string that will look
through transcripts for “patient safety” OR “hazard” OR “hospital acquired”, etc.. Figure 15 shows the results
of the stored search, and this pane allows autocoding of the hits by means of a single click on the icon Au-
tocode search results with a new code .

Figure 14: Stored Lexical Search

Focus group coding 23

Figure 15: Lexical Search results and starting point for autocoding

Note: Care must be taken not to let bulk coding obscure the need for careful analysis and coding.

Thematic coding
A key part of focus group research is identifying and coding the participant text that matches the themes set
by sponsor research questions. In addition to speaker codes, these thematic codes will reflect text segments
that inform the sponsor questions, and whose summaries will later serve as the basis for the project report. It
is often the case that the research questions will be operationalized into a code system containing codes that
are individually necessary to answer the research questions, and collectively sufficient. Thematic coding in
focus group research offers a detailed description of the data rather than developing a theory, but does so in
a constrained universe of questions that are of interest to a sponsor. Thematic coding greatly facilitates the
development of summaries described in the next chapter.
The thematic codes can be applied to a discrete piece of contribution by a single speaker, or can also en-
capsulate several contributions. As a best practice, when several successive contributions follow each other
in the transcript, they are coded as a group rather than applying the same code multiple times.
24 Focus group coding

Figure 16 shows a code system in which a number of codes are used to address sponsor research questions
for a healthcare focus group. It is often useful to specify in the code memo, how each code aligns to the
sponsor questions, which may often be too broad or vague to specify directly as a code.

Figure 16: Sponsor research question (thematic) codes

Grounded coding
Although focus groups typically have a set number of questions or topics that have been derived from the
sponsor questions, or in some cases directly dictated by a sponsor, there are likely to be times when a parti-
cipant response is unexpected or takes a direction that was not anticipated by the sponsor or project designer.
When these variations add value, they can be captured by adopting procedures for developing codes inspired
by a grounded theory approach, from which novel understanding or insights may emerge. As such, you
should be prepared to develop a parallel grounded code system that addresses and explores emergent themes
or issues. Figure 17 shows codes that have been developed through grounded theory to describe emergent
themes that were not captured by the predetermined thematic codes.
What comes after coding? 25

Figure 17: Grounded codes

6 What comes after coding?

Many authors and training courses assert that coding is not really analysis, and merely a precursor to analysis,
but this is perhaps a superficial view. In many ways, analysis has already started by virtue of the cognitive
process of coding, and even more so in grounded theory studies when codes are created from textual analysis.
Analysis is arguably involved in the process of merging, splitting, or changing codes, and part of the interpre-
tation performed to create paraphrases.
For many researchers, the next step in the focus group analysis process, is to begin articulating what has
been discovered by means of summaries, and using them to populate a report.

Using summaries
Paraphrases religiously reflect the voice of the participant, but the summary reflects the voice of the narrator.
Summaries provide the significant events, exchanges, and meanings of a transcription segment in the words
of the researcher. As such, there is overlap between how in-document memos and summaries can be used to
code for a focus group transcript.
In-document memos are typically created before coding and as a means to develop codes, whereas the
summary is written after initial coding is completed. A summary reflects all the segments for a specific exis-
ting code within a document, i.e., one focus group. Summaries can provide an analysis for reporting or can
be used to develop additional codes or make code refinements. As such, MAXQDA provides a toolbox for
developing summaries using the intersections of codes and the focus group transcripts. All the intersections
26 What comes after coding?

of a transcript document and the codes used on it are available in the Summary Grid view (available in the
Analysis ribbon tab).
Figure 18 shows a Summary Grid in which the intersection of a code used to denote questions between
participants, and the transcript for “Group 1”. The center column contains the original transcript text of the
coded segments, with a highlighted hyperlink to the document browser location. The right column contains
the researcher’s description of the text.

Figure 18: Summary Grid—the content of the coded segments for each relevant code is summarized for each focus group
document

MAXQDA provides two other tools related to summaries: The Summary Table offers a compilation of sum-
maries, and is useful for presentations and reports, while the Summary Explorer enables the researcher to
compare the summaries of different cases or groups. Both tools are also available in the Analysis ribbon tab.

Codeline
The codeline feature is useful for analyzing the sequence of topics and participants. Figure 19 shows an exa-
mple of a codeline for the speakers in Group 1, Session 1 of a healthcare focus group. The codeline is an easy
What comes after coding? 27

way to see participation sequences in the transcript, and thus whether some participants are more active
than others. The codeline visual also shows if some participants reentered conversation multiple times, or
were represented only in a single slot. The codeline feature can be accessed by going to Visual Tools > Code-
line.

Figure 19: Codeline

Crosstab
The Crosstab feature is useful in focus group analysis to compare code frequencies for groups of participants
defined by their variable values. Double-clicking on any of the values, will populate the “Retrieved Segments”
pane. The Crosstab feature for focus group speakers can be accessed by going to Mixed Methods > Crosstab
> Crosstab for Focus Groups.
28 What comes after coding?

Code Matrix Browser

This function allows the researcher to see how code allocation compares between participant groups (co-
lumns) and code frequencies per participant (rows), or as shown in Figure 20, the researcher can see activated
code allocation by groups. The display can be selected to show heatmap nodes as squares or circles, as inte-
gers, or as a combined heatmap with integers. This display provides an easy way to see differences or simila-
rities between the groups or participants. Double-clicking a row populates the “Retrieved Segments” pane.
The Code Matrix Browser can be accessed by going to Visual Tools > Code Matrix Browser.

Figure 20: Code Matrix Browser

Interactive Quote Matrix

This is a useful feature in analyzing focus group data because it provides a means to compare the content of
coded segments, comments on coded segments, and links to associated segment memos per participant, or
group of participants. Based on selection criteria, the results pane can show the coded segments and code
comments in one pane, as well as provide a hyperlinked point of origin document name with start and end
line number. Clicking on this origin link will navigate to the coded segment in the “Document Browser”. The
pane also shows the name of any segment memos. Hovering the mouse over the memo icon in the pane will
What comes after coding? 29

result in a pop-up pane containing the memo text. Figure 21 shows the results pane and the segment content,
link, code comment, and content of the associated in-document segment memo. The Interactive Quote Mat-
rix feature can be accessed by going to Mixed Methods > Interactive Quote Matrix.

Figure 21: Interactive Quote Matrix

Developing a report
An effective way to approach the development of a report is to review the summaries and memos, and to
develop an outline of the themes that are observed to run through the summaries and memos. The summa-
ries, memos, and quoted text can be used in the report body.

 Develop a storyline that draws from the summaries and outlines the themes related to
o questions and outcomes agreed by the sponsor
o emergent themes that may be relevant to the research questions
o unexpected risks, issues, or opportunities that bear description or explanation
30 What comes after coding?

 Use the Code Matrix Browser and Code Relations Browser results (both available in the Visual Tools rib-
bon bar) to support the theme outline
 Use the codes to identify quotable participant text to illustrate the themes
 Use the graphic reports to visualize the themes

It may be useful to develop your own method to keep track of which summaries, memos, and codes have
been “used” in the report, those that have been left out deliberately, and those that remain to be used. For
memos, I use the “Memo Type” icons in conjunction with the Memos > Memo Manager filter to keep track of
which memos I intend to use, and which I have used already. This reduces the chance that I will forget a point,
and gives me a sense of progress in my documentation. To use this feature, click the gear wheel in the Memo
Manager toolbar, and specify a meaning for each memo type icon that you will use. Once the icons have been
defined, you can assign a memo type to each memo, and use the funnel icon in the Memo Manager window
to filter the display by memo type.

Figure 22: Memo Manager

Tips and tricks 31

I use “Code Sets” for the same purpose, and add all the codes I intend to discuss to a “To be Used” set, and
move them to a “Used” set once they have been dealt with. I export the summaries and paraphrases into an
Excel sheet and use that to keep track of what I intend to use and those that have been used.

7 Tips and tricks

MAXQDA has proven to be an effective tool for analysis and reporting of focus group data. The following are
a few lessons learned:

 Remember to document logistics. Use the MAXQDA Home > Logbook feature to keep a diary in the pro-
ject of logistical decisions, such as venues, conferencing tools, etc. It is easy to forget why a particular
location was selected, or who agreed to purchase space, conferencing services, or refreshments. The Log-
book will stay with the project and can be used to answer any questions about the project logistics.
 Determine beforehand how prompts will be captured. In some focus groups, prompts are used to initiate
discussion, and may be the object of the discussion. For example, a set of images might be shown to the
participants, and then several questions related to the image will be posed. However, the prompt might
be a physical object, a video clip to watch, a software app, a tool, or an in-situ walkthrough of a process.
These prompts need to be represented in some fashion in the data, which may take significant forethought
and planning. Usually, it is a good advice to import them into your MAXQDA project alongside the focus
group transcripts, if possible. This allows you to create internal links if a participant mentions parts of the
stimulus.
 Plan how large media files will be handled. Audio and video files for an entire 90-minute focus group
session can become very big. Plan ahead for where you will store these files, and how archiving and ba-
ckups will be managed.
 Consider how you will deal with “narrator’s voice” and significant actions. Before starting your analysis
session, think through how you will denote, code, and use events like people entering or leaving, dropping
things, etc. Will you code them with a particular code like “actions” perhaps, and add a code comment, or
will you add a memo, or do something else?
 Clean before you import. Don’t assume data from a transcription service, or a video clip or audio track,
are clean. They may require significant editing and cleaning prior to import, and may require highly spe-
cialized tools that take time to acquire.
32 MAXQDA checklist

MAXQDA checklist
The following checklist describes steps in MAXQDA that closely resemble the path I typically take with focus
groups. Your experience may differ.

1. Check transcripts for spelling errors or variations in the names of participants

2. Check that speaker changes are indicated by the speaker name and a colon
3. Remove any extraneous colons in the text
4. Import transcripts via Import > Focus Group Transcripts
5. Check for duplicate or unexpected speaker names in the “Document System” and “Code System”
6. Import any speaker variables anticipated for analysis
7. Give each speaker a different code color
8. Use the Document Portrait to see any speakers that are over-represented or under-represented
9. Use the Word Cloud to gain a quick overview of frequent topics
10. Use Keyword-in-Context (KWIC) to explore topics and to spot terminology variances
11. Double-click each participant name in the “Code System” to see a listing of all their contributions
12. Read the full transcripts in order of occurrence
13. Read the combined transcripts for each topic
14. Read the transcripts by participant
15. Add memos
16. Add paraphrases
17. Perform any bulk coding
18. Create a coding scheme containing thematic codes
19. Decide if you need to add codes for interaction
20. Perform thematic and grounded coding
21. Add summaries
22. Start building the report from memos and summaries
23. Celebrate, and then return to #20 until the report is populated
24. Check that all summaries have either been used or deliberately omitted
25. Exhale!
Checklist for conducting focus groups 33

Checklist for conducting focus groups

Researchers will have varied needs and environments, but I have found building a checklist for running focus
group projects to be useful as a reminder. Here are my top twenty checklist items.

1. Do I understand the sponsor’s need, and have they agreed to my operationalization into the focus group
questions or topics
2. Do I have a data retention and destruction plan
3. Do I know what characteristics define a good fit for a participant
4. If participants will be rewarded in some fashion, is that
a. Documented
b. Approved
c. Logistically practical
5. Do I have a clear recruitment plan
6. Do I have commitment from enough participants to make a viable session
7. Do I have signed participant consent releases
8. Do I have dietary requirements captured
9. Is the venue confirmed (either physical or online)
10. Is Security aware of the session (especially when after hours)
11. Have I checked the venue (physical sessions) for
a. Handicap access
b. Findability
c. Lighting
d. Ventilation
e. Security
f. Furniture
g. Parking
h. Sound—echo, external noise
12. Have I checked data (online sessions)
a. Accessibility
b. Bandwidth for video
c. Sound quality
d. Reliability
34 Checklist for conducting focus groups

e. Cost to participants
f. Equipment minima
13. Do I have backup recording devices for audio, video
14. Do I have a fallback plan for cancellation
a. Venue unavailability
b. Insufficient participants
c. Data connection issue
d. Participant or moderator ethics or safety violation
15. Have refreshments been ordered and confirmed
16. Are visual aids and prompt available and secured
17. Are office supplies available and secured
a. Flip charts
b. Pens
c. Notepads
d. Dry board
e. Dry erase pens
f. Extension cords
g. Chargers
h. Projectors
i. Sound systems
18. Have I reiterated participants rights
19. Has every participant agreed to the interview and being recorded
20. Do I need to renew or order MAXQDA licenses
Bibliography 35

Bibliography
Bogardus, E. S. (1926). The group interview. Journal of Applied Sociology, 10(4), 372–382.

Carlsen, B., & Glenton, C. (2011). What about N? A methodological study of sample-size reporting in focus group
studies. BMC Medical Research Methodology, 11, 11–26. [Link]

Creswell, J. W., & Poth, C. N. (2018). Qualitative inquiry & research design: Choosing among five approaches. Sage.

Cross, R., & Warwick-Booth, L. (2016). Using storyboards in participatory research. Nurse Researcher, 23(3), 8–12.
[Link]

Gallagher, M., Hares, T., Spencer, J., & Bradshaw, C. (2017, 1). The nominal group technique: A research tool for
general practice? Family Practice, 10(1), 76–81. [Link]

Jackson, P. (1998). Focus group interviews as a methodology. Nurse Researcher, 6(1), 72–84.

Krueger, R. A. (2002). Designing and conducting focus group interviews. [Link]

[Link]

Krueger, R. A., & Casey, M. A. (2009). Focus groups: A practical guide for applied research. Sage.

Kuckartz, U., & Rädiker, S. (2019). Analyzing focus group data. In Analyzing qualitative data with MAXQDA: Text,
audio, video (pp. 201-217). Springer. [Link]

Lee, R. M. (2010). The secret life of focus groups: Robert Merton and the diffusion of a research method. The Amer-
ican Sociologist, 41(2), 115–141. [Link]

Loxton, M. H. (2021). Using MAXQDA for analyzing focus groups: An example from healthcare research. In M. C
Gizzi & S. Rädiker (Eds.), The art of qualitative data analysis: Research examples using MAXQDA. MAXQDA
Press. [Link]

Morgan, D. L., & Hoffman, K. (2018). A system for coding the interaction in focus groups and dyadic interviews. The
Qualitative Report, 23(3), 519–531. [Link]

Robson, C. (2011). Real world research: A resource for users of social research methods in applied settings (3rd ed.).
Wiley.

Schultz, D. P., Schultz, S. E., & Schultz, D. P. (1994). Psychology and work today: An introduction to industrial and
organizational psychology. Macmillan.

Van Bennekom, F. C. (2002). Customer surveying: A guidebook for service managers. Customer Service Press.
This guide provides an overview for using MAXQDA with focus groups, and how Matthew H. Loxton
it can support the collection, analyxsis, and reporting of focus group data:

• Preparing and importing focus group data in MAXQDA

• Creating and importing speaker variables
• Exploring the focus group data
Analyzing Focus Groups
with MAXQDA
• Focus group coding
• What comes after coding?
• Tips and tricks incl. checkli
sts
Real-world examples from healthcare illustrate the explanations. A checklist for
analysis of focus groups with MAXQDA as well as a checklist for conducting
focus groups support the realization of your projects.

Matthew Loxton is a Principal Analyst at Whitney, Bradley, and Brown Inc.

focused on healthcare improvement and strategic foresight, serves on the board
of directors of the Blue Faery Liver Cancer Association, and holds a master’s
degree in Knowledge Management from the University of Canberra.

MAXQDA Guide
MAXQDA
PRESS PRESS
[Link]