Journal of Research in Nursing

2016, Vol. 21(8) 623–635

An overview of mixed methods ! The Author(s) 2016
Reprints and permissions:

research – revisited sagepub.co.uk/journalsPermissions.nav

DOI: 10.1177/1744987116674257
jrn.sagepub.com

Louise Doyle
Assistant Professor, School of Nursing and Midwifery, Trinity College Dublin, Ireland

Anne-Marie Brady
Associate Professor, School of Nursing and Midwifery, Trinity College Dublin, Ireland.

Gobnait Byrne
Assistant Professor, School of Nursing and Midwifery, Trinity College Dublin, Ireland.

Abstract
Mixed methods have become increasingly popular in health-related research, allowing a broader
and deeper understanding of complex human phenomena. This growth in popularity has increased
the range of usage and complexity in design approaches, producing greater need for understanding
of logistical and practical application in this field. This paper aims to provide an overview of mixed
methods research and orientate readers to the critical issues that arise for researchers. It provides
an introduction to mixed methods design and the rationale and philosophical underpinnings for this
methodological approach. It navigates the reader through some real world or ‘hot’ topics within
mixed methods, including data analysis, integration and quality appraisal criteria.

Keywords
healthcare research, integration, mixed methods, qualitative research, quantitative research

Introduction
The landscape of mixed methods research has advanced significantly in the years since the
publication of our original ‘overview of mixed methods research’ (Doyle et al., 2009). Its
popularity as a research design has continued to grow, particularly in the fields of nursing
and health sciences (Creswell, 2015a) where it is commonly used as a design for graduate
theses and evaluation research. Concomitant with the increased use of mixed methods is an
exponential increase in the amount of instructional texts guiding novice and experienced
researchers alike (Creswell, 2015a; Sandelowski, 2014). This paper aims to provide an
overview of the current state of mixed methods research with a focus on a number

Corresponding author:
Louise Doyle, School of Nursing and Midwifery, Trinity College Dublin, 24 D’Olier St, Dublin 2, Ireland.
Email: [email protected]
624 Journal of Research in Nursing 21(8)

of critical issues. There continues to be significant debate about what constitutes mixed
methods research (Hesse-Biber, 2015); we are guided by the definition of Creswell (2015a),
who suggests that mixed methods research is an approach in which the researcher collects,
analyses and interprets both quantitative and qualitative data, integrates the two approaches
in various ways and frames the study within a specific design.

Rationale for mixed methods research

It is critically important that the research question is one that lends itself to a mixed methods
design; ideally one in which using quantitative or qualitative methods alone would be
insufficient. The assumption that mixed methods research is inherently better than a
mono-method approach is being challenged (Sandelowski, 2014), thus strengthening the
need for mixed methods researchers to be explicit about the add-on value of their design
(Creswell, 2015b). However, despite this being an essential step and an important quality
criterion (O’Cathain, 2010), mixed methods research is often insufficiently justified (Bryman,
2008). The justification of mixed methods designs has been considered in depth by Greene
et al. (1989), who identified five main purposes for mixing methods (triangulation,
complementarity, development, initiation, expansion), and by Bryman (2006), who further
expanded upon this scheme. The most commonly identified rationales for mixed methods
studies are presented in Table 1.
In addition to methodological reasons, there has been an increase in funding agencies
explicitly requiring mixed methods designs. This is often the case in health research in which
interdisciplinary research is increasingly popular, requiring collaboration across a range of
disciplines grounded in a variety of research approaches. Mixed methods research therefore
heralds an opportunity for greater interdisciplinary collaboration (Hesse-Biber, 2015).

Table 1. Rationales for mixed methods research.

Triangulation  Using quantitative and qualitative methods so that findings may be mutually
(convergence): corroborated.
 This may also be an unanticipated outcome of the study where a mixed
methods study was undertaken for another reason, but convergence was
evident.
Expansion:  The first phase has findings that require explanation qualitatively.
 Unexpected findings that need to be explained.
Exploration:  An initial phase is required to develop an instrument or intervention,
identify variables to study or develop a hypothesis that requires testing.
Completeness:  Provides a more comprehensive account of phenomena under study.
Offset weaknesses:  Ensures that weaknesses of each method are minimised (Creswell, 2015a).
 Caution is required when identifying this as a primary rationale as each
method should be sufficiently rigorous in its own right (O’Cathain, 2010).
Different research  Both quantitative and qualitative questions may be posed at the beginning of
questions: the study in addition to mixed methods questions (Creswell, 2015b).
Illustration:  Qualitative data are used to illuminate quantitative findings.
 Putting ‘meat on the bones’ of dry quantitative data (Bryman, (2006).
Doyle et al. 625

Philosophical underpinnings
Paradigms can be viewed as a system of beliefs and practices that influence how researchers
select both the questions they study and the methods they use to study them (Morgan, 2007).
The ‘paradigm wars’, which centred on the perceived differences in philosophical assumptions
between the positivist/post-positivist paradigms and the constructivist/interpretivist
paradigms, raged from the 1970s to the 1990s (Teddlie and Tashakkori, 2003). This war
fuelled the ‘incompatibility thesis’, the belief that the two paradigms could never be mixed
due to the inherent differences underlying them. It is beyond the scope of this paper to provide
an in-depth discussion of the paradigm wars; readers are directed elsewhere (Howe, 1988;
Morgan, 2007; Sale et al., 2002; Teddlie and Tashakkori, 2003). However, the incompatibility
thesis does pose a challenge – how can researchers mix methods when the paradigms on which
they are based have vastly difference ontological, epistemological and methodological
assumptions? For some, the adoption of a dialectical approach which advocates using two
or more paradigms together (Shannon-Baker, 2015) provides a solution to this issue. For
most, however, the answer is to adopt an alternative paradigm, that embraces a plurality of
assumptions and methods (Greene, 2007). Critical realism is one such paradigm which
supports the belief that quantitative and qualitative research can work together to address
the other’s limitations (Shannon-Baker, 2015) and offers strategies for mixed methods
researchers to better understand the context of what they study (Maxwell and Mittapalli,
2010). Other approaches include the transformative paradigm, which is a framework of belief
systems that places priority on social justice and human rights (Mertens, 2010) and the
incorporation of value-based goals within a mixed methods study (Shannon-Baker, 2015).
Pragmatism, however, is the most frequently identified alternative paradigm on which
mixed methods researchers base their work (Feilzer, 2010). The utility of pragmatism is that
it aims to find middle ground between philosophical dogmatisms (Johnson and Onwuegbuzie,
2004). On a practical level, pragmatism offers health researchers the freedom to choose the best
methods to answer the research question to hand (Bishop, 2015a), advocating for a balance
between subjectivity and objectivity throughout the research (Shannon-Baker, 2015). On a
philosophical level, pragmatism supports the view that while qualitative and quantitative
methods are distinct, they are also commensurate as both advance knowledge production
(Bishop, 2015a) and shared meaning making (Shannon-Baker, 2015).

Mixed methods designs

A plethora of mixed methods designs and typologies have emerged, which can be confusing
for both novice and experienced mixed methods researchers. Some of these designs are
overly complex and there is a call for mixed methods researchers to return to more simple
mixed methods designs with the option of seeking out variants of them if required (Creswell,
2015a). In this paper three basic mixed methods designs (convergent, explanatory sequential
and exploratory sequential) and one advanced design (embedded intervention) are presented
with examples of their use in health research. The Figures (1–3) presented below are broadly
based on those in Creswell and Plano Clark (2011).

The convergent design

The convergent design (Figure 1), sometimes called the convergent parallel or concurrent
triangulation design (Creswell and Plano Clark, 2011; Creswell et al., 2003), addresses one
626 Journal of Research in Nursing 21(8)

Figure 1. Convergent design.

Source: Adapted from Figure 3.2(a): Creswell and Plano Clark (2011: 69).

overarching research question and is used primarily, although not exclusively, when
researchers are looking for convergence affording a more complete understanding of
phenomena (Doyle, 2015). Within this design, quantitative and qualitative data are
collected concurrently but remain separate – the findings of one phase are not dependent
on the results of another (Creswell and Plano Clark, 2011). In most convergent designs,
equal priority is assigned to quantitative and qualitative data and results are usually merged
in the interpretation phase of the research where meta-inferences are developed. It is an
efficient design as all data are gathered around the same time so the population remains
accessible. This is particularly beneficial in nursing and healthcare research where the sample
(frequently patients) may be discharged, thereby making a second phase of recruitment
difficult. One of the main challenges of a convergent design is deciding what to do if,
instead of converging, the findings actually diverge. Novice researchers in particular often
express a desire for convergence in their study in the hope of tying all the results up neatly
and strengthening the validity of their findings. However, as discussed later in this paper, it is
important to note that divergence is not necessarily a sign that something is wrong with the
study (O’Cathain et al., 2010). Stoddart et al. (2014) utilised a convergent (concurrent)
design to explore the new clinical leadership role of senior charge nurses. Quantitative
data were collected from 50 participants through the use of an online survey, and
nine participants took part in qualitative interviews to explore the views and
experiences of senior nurses regarding the implementation of a national clinical leadership
policy. This design provided a depth of responses that would not have been achievable with
the sole use of quantitative methods, with the additional benefit of a good degree of
convergence between quantitative and qualitative responses, thereby increasing the
validity of the study.

The explanatory sequential design

The explanatory sequential design (Figure 2) usually consists of a larger quantitative phase
followed by a smaller qualitative phase, the aim of which is to follow up and explain the
quantitative results. Data collection and analysis usually occur sequentially, and as
the results of the quantitative phase guide the development of the qualitative phase, the
quantitative phase is normally dominant. The advantage of this design is that it is relatively
Doyle et al. 627

Figure 2. Explanatory sequential design.

Source: Adapted from Figure 3.2(b): Creswell and Plano Clark (2011: 69).

straightforward with distinct sequential phases of data collection making it possible for a
lone researcher to complete (Creswell and Plano Clark, 2011). Challenges include the fact
that because it is generally an emergent design where the second phase cannot be fully
developed until the first phase has been completed, the study may have to go before an
ethics committee or institutional review board (IRB) for a second time to gain approval for
the qualitative phase (Creswell and Plano Clark, 2011). Furthermore, the longer timeframe
required to undertake the study may make recruitment for the qualitative phase problematic
as the sample may no longer be available to the researcher. An explanatory sequential design
was used by Jellesmark et al. (2012) to explore fear of falling (FOF) and functioning ability
among older people following a hip fracture. In the first phase, surveys were used to identify
FOF and related factors. This was followed up by qualitative interviews with a sub-group of
participants who were identified from survey data as having a high degree of FOF. The
qualitative data allowed for an in-depth explanation of what influenced participants’ fear of
falling and how this impacted on their life.

The exploratory sequential design

The exploratory sequential design (Figure 3) is characterised by a primary qualitative
phase which builds into a quantitative phase. The qualitative phase is useful in the
development of instruments where none exist, in the identification of variables that are
unknown and in the development of theory or hypotheses (Creswell and Plano Clark,
2011). The quantitative phase can serve to test the instrument and/or generalise the
qualitative results to a wider population. Priority in an exploratory design depends on
the purpose of its utilisation; if used to develop a theory the qualitative phase is normally
dominant, but if used to develop and test an instrument the quantitative phase usually
takes priority. Similar advantages and challenges arise with this design as the explanatory
design; however, an additional challenge concerns the added complexity that is inherent in
developing and testing a new instrument (Creswell, 2015a). Stoller et al. (2009) conducted
a sequential exploratory study to explore factors that affect the decision to curtail alcohol
consumption in those with hepatitis C, a previously under-researched topic. Interviews
were conducted with 42 participants with hepatitis C who had been advised to curtail
alcohol. From these interviews, 17 decision factors were identified which then fed into the
development of a survey measuring these 17 new factors. This survey was then
administered to 577 people with hepatitis C, thereby testing these new factors in a
larger sample and providing prevalence estimates.
628 Journal of Research in Nursing 21(8)

Figure 3. Exploratory sequential design.

Source: Adapted from Figure 3.2(c): Creswell and Plano Clark (2011: 69).

The embedded intervention design

The embedded intervention design (Figure 4), also called the experimental intervention
design (Creswell, 2015b), is characterised by the inclusion of a qualitative phase embedded
within an experiment or intervention trial which can help to minimise some of the problems
associated with intervention studies. The qualitative phase can serve a number of different
functions and can be used prior to, during or after the intervention, or at all three time points
(Creswell, 2015a; Sandelowski, 1996). Prior to the study a qualitative phase can be
introduced to develop an instrument or inform recruitment for the trial. During the study,
a qualitative phase can enhance understanding about how participants are experiencing the
intervention leading to modification of the intervention if there is a clearly identifiable,
fixable problem (Creswell, 2015a). However, this needs to be done cautiously to avoid
threatening the validity of the trial (Creswell, 2015a; Fetters et al., 2013). After the trial, a
qualitative component can contribute to an understanding of why the intervention did or did
not work, can identify how it might be improved and can aid exploration of statistically non-
significant cases in which participants responded very well or not at all (Drabble and
O’Cathain, 2015; Sandelowski, 1996). The use of qualitative research within intervention
trials has increased in response to the realisation that within healthcare research, such trials
may have little or no clinical significance or real life importance to patients, clinicians and
policy makers (Drabble and O’Cathain, 2015). By increasing the ‘human element’ of
rigorous controlled trials, an embedded intervention design can make the findings more
meaningful for all concerned (Creswell, 2015a). An embedded design was utilised by
McCabe et al. (2013) wherein a randomised control trial (RCT) sought to determine the
effect of new media art using a virtual window on health-related quality of life in patients
experiencing stem cell transplantation. Findings from the RCT demonstrated that patients
exposed to the intervention had better overall experiences of treatment, including more
positive depression and anxiety scores. Findings of embedded qualitative interviews
provided participants with the opportunity to explain the beneficial effects of the
intervention, thereby providing an example of how embedded intervention designs can not
only identify if interventions work, but importantly, how they might work.

Critical issues in mixed methods research

Attention is now focused on a number of key issues in the application of mixed methods
designs, including data analysis, integration and quality appraisal.
Doyle et al. 629

Figure 4. Embedded intervention design.

Data analysis
Most mixed methods studies separately analyse quantitative and qualitative data with
methods suitable to their own tradition, merging the findings in the interpretation phase
(Creswell and Plano Clark, 2011). However, there is potential to further integrate data
analysis in mixed methods research using a number of techniques.

Typology development. Analysis of one method provides a framework for analysis of the
second; in most cases, quantitative data provide a framework to ‘hang’ qualitative data
on. A coding matrix is developed which identifies the main quantitative findings with a
text box underneath for the explanatory qualitative data. The qualitative data are then
630 Journal of Research in Nursing 21(8)

usually subjected to another level of thematic analysis. This method is particularly useful in
sequential explanatory designs as it helps to ensure that questions raised in the quantitative
phase are answered. However, it is important not to ignore other qualitative data, which
may not be directly linked to the quantitative data but may nonetheless provide valuable
insights.

Data transformation. Here one form of data is transformed into another. It can take the form
of qualitising data, where quantitative data are transformed into qualitative data. This is not
a frequently used method and usually involves constructing narrative profiles from
quantitative data. The most common form of data transformation occurs when
qualitative data are transformed into quantitative data (quantitising). This can occur
when qualitative data are assigned a binary value and are then analysed statistically
(Collingridge, 2013). Quantitising can also take the form of ‘counting’, where responses of
participants to various questions are counted. This generates descriptive statistics, which
allows for the identification of patterns in the data that may have been missed through
qualitative analysis (Collingridge, 2013). However, there is a note of caution with the use
of quantitising: qualitative data by their nature are not randomly selected or normally
distributed so caution is required when interpreting statistical tests and the generalisability
of the results (Maxwell, 2010). Furthermore, there is a real danger that the very essence of
qualitative research (production of rich, depth data) will be lost through this method, so
quantitisation should not replace qualitative analysis.

Software programs. Programs such as NVivo and MAXQDA can incorporate quantitative
data into a qualitative analysis. They can also facilitate the linking of demographic detail or
responses on a survey to qualitative data from individual cases (Bazeley, 2012), thereby
facilitating further integration of data.

Integration
Data integration has become a ‘hot topic’ in mixed methods research and is one of the major
issues facing those using these designs (Creswell, 2015b). Although much of the literature
refers to integration as the mixing of qualitative and quantitative data, it is broader than that
and can be defined as the bringing together of quantitative and qualitative components of a
study (O’Cathain et al., 2010). It is essentially the point(s) where quantitative and qualitative
phases intersect – it is the ‘mixing’ in a mixed methods study. Bryman (2007) suggests that
those reading mixed methods research deserve more from the findings than being presented
with parallel accounts that barely connect. Good integration in a mixed methods study
should provide ‘a whole greater than the sum of its parts’.
In a move away from the notion that integration in mixed methods research refers only to
data integration, Fetters et al. (2013) have identified how it can occur at the design, methods
and reporting/interpretation levels of research.

Integration at design level. An overlooked point of integration in mixed methods studies is the
setting of qualitative and quantitative questions within one study and/or the setting of
research questions that are inherently mixed methods from the outset (Creswell, 2015b).
Research questions that require both quantitative and qualitative methodologies to
answer them necessitate a mixed methods design – the development of which is a second
Doyle et al. 631

point of integration in mixed methods studies. The research design of a mixed methods study
should have at least one connecting point of integration and may have more. In the
explanatory and exploratory designs, one phase feeds directly into another. In convergent
designs, the two forms of data are directly compared.

Integration at methods level. At the methods level there may be several points of integration,
particularly in studies using a sequential design. Findings from the first phase may inform
development of a sampling frame for the second. This may occur in an explanatory design
where the qualitative sample is purposively selected to followup unusual results. Integration
can also occur when findings from the first phase are used to inform the development of the
second phase, termed integration – though building by Fetters et al. (2013). In an
exploratory study, findings from interviews may be used to develop an instrument that is
then tested quantitatively in phase two. In an explanatory study, findings from the
quantitative phase can inform the development of the interview guide and other data
collection tools such as vignettes to help inform the qualitative phase. The final point of
integration at the methods phase is at data analysis level as previously discussed.

Integration at reporting/interpretation level. In many mixed methods studies, findings are

presented separately; however, there is a move to increase integration of data in the
findings section. One method is to present quantitative and qualitative findings together
on a theme by theme basis. This type of integration is particularly useful for convergent
designs where meta-themes are presented which cut across findings from both methods
(Farmer et al., 2006; O’Cathain et al., 2010). In studies where findings are presented
separately, there is still potential to increase integration through ‘weaving’ (Fetters et al.,
2013). This can work well in explanatory designs where quantitative and qualitative findings
are mostly presented separately, but where the main quantitative findings are also briefly
woven into the qualitative findings to act as a hook for the explanatory findings. Another
strategy to increase integration is the use of a joint display. Creswell (2015b) provides an
example of how this might be done for a sequential explanatory study where within a table
one column presents the quantitative findings, a second column presents the qualitative
findings and a third presents information about how the quantitative findings were
explained by the qualitative ones.

Integration at discussion level. At the very least, a mixed methods study must be integrated at
discussion level (Creswell and Plano-Clark, 2011). Within the discussion section, inferences
may be drawn from each phase of the study separately. A crucial step forward is to then
develop meta-inferences, which are integrated understandings derived from both the
quantitative and qualitative data (Tashakkori and Teddlie, 2008). Meta-inferences go
beyond what is learned from the quantitative and qualitative components in isolation;
they provide a sense of the overall findings. In essence, the meta-inference(s) should
answer the mixed methods questions.

Quality in mixed methods studies

Current debate about quality in mixed methods research focuses on whether there should be
separate quality appraisal criteria applied to the quantitative and qualitative components or
whether there should be bespoke mixed methods criteria (O’Cathain, 2010). In recognition
632 Journal of Research in Nursing 21(8)

that high quality quantitative and qualitative components are necessary but not sufficient for a
quality mixed methods study (Tashakkori and Teddlie, 2008), a number of mixed methods
quality frameworks have been advanced; however, there are no agreed criteria for evaluating
mixed methods studies (Ivankova, 2014). In a bid to advance the ‘quality’ argument, O’Cathain
(2010) developed an eight-domain quality framework focusing on important components
including planning quality, design quality and reporting quality. This quality framework can
be a useful tool to determine quality. However, there are two caveats: first O’Cathain herself
recognises that in practice, this framework has too many criteria and that there is a need to
prioritise the most important elements; and second, quality frameworks should not be viewed
as rigid templates, but rather as general guidelines for use (Creswell, 2015b).

Challenges when undertaking a mixed methods design

One of the main challenges when conducting mixed methods research is what to do about
divergent findings. Most researchers strive for congruency between quantitative and
qualitative findings; however, divergent findings can uncover new theories and insights
(Creswell et al., 2008). Nonetheless, there is a requirement that mixed methods researchers
acknowledge and attempt to address inconsistencies between the two sets of findings as the
quality of a study may be adversely affected by not addressing divergent findings (Ivankova,
2014). One potential reason for divergence is methodological differences between the two
phases of research; for example, the use of anonymous methods in a quantitative phase and
non-anonymous methods in a qualitative phase might lead to different responses,
particularly when exploring sensitive topics, as reported in a study of suicidal behaviour
(Safer, 1997). Another explanation is that quantitative measures may not be sensitive enough
to pick up on complex experiences that have been reported qualitatively, as reported in a
study of maternal mental health following miscarriage (Lee and Rowlands, 2015).
Divergence may also have a theoretical explanation (Fetters et al., 2013) and it may be
appropriate to collect additional data to resolve the discrepancy leading to the
development of a new study (Creswell et al., 2008; Fetters et al., 2013); however, this may
not always be possible where the study period is limited.
A further challenge in mixed methods research is the potential threat of tokenistic use of
qualitative research within a prominently quantitative design, perhaps stemming from the
historical dominance of quantitative methods in health research (Bishop, 2015b). While
Morse (2015) suggests that the quantitative/qualitative imbalance will correct itself with
time, there is still the risk that qualitatively driven mixed methods research will continue
to be dwarfed by quantitative research and will continue to play a secondary role in terms of
importance and function (Giddings, 2006; Hesse-Biber et al., 2015).

Conclusion
The use of mixed methods research continues to grow, and as a design it has much to offer
nursing and healthcare researchers. Its appeal is largely grounded in its ability to offer a
broader and deeper understanding of complex health circumstances and human phenomena
and in its ability to incorporate ‘the patient’s voice’ (Morse, 2015). This is evident in some of
the examples of mixed methods designs presented in this paper in which patients’ experiences
were qualitatively elucidated to explain results in more detail, making them more meaningful
to healthcare practitioners. To date, there have been limited texts providing an overview
Doyle et al. 633

of mixed methods research for those considering its use in healthcare research. The focus
instead has been on specific issues within mixed methods, including the paradigm wars and
the issue of data integration. This paper provides a sound starting point for novice researchers
and more experienced researchers unfamiliar with mixed methods designs by identifying key
issues and potential challenges in the use of this design. We hope that this paper encourages
healthcare researchers to undertake mixed methods studies, successful application of which
will be characterised by wisely selected research designs and careful diligence in planning and
addressing issues such as integration, analysis and quality appraisal.

Key points for policy, practice and/or research

. Mixed methods research remains a popular design in nursing and healthcare research.
. The increasing complexity of some designs can be perplexing, particularly to novice
researchers.
. Nursing and healthcare researchers should focus on identifying a rationale for their
mixed methods study, choosing an appropriate design to meet the objectives and
identifying clearly where integration takes place.
. An integrated mixed methods study can ultimately facilitate a greater understanding
of complex human phenomena that exist in healthcare research in addition to allowing
the patient’s voice be heard.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or
publication of this article.

Funding
The author(s) received no financial support for the research, authorship and/or publication of this
article.

References
Bazeley P (2012) Integrative analysis strategies for mixed In: Bergman MM (ed.) Advances in Mixed Methods
data sources. American Behavioral Scientist 56(6): Research. London: SAGE, pp. 87–100.
814–828. Collingridge DS (2013) A primer on quantitized data analysis
Bishop FL (2015a) Using mixed methods in health research: and permutation. Journal of Mixed Methods Research 7(1):
Benefits and challenges. British Journal of Health Psychology 81–97.
20(1): 1–4. Creswell JW (2015a) A Concise Introduction to Mixed Methods
Bishop FL (2015b) Using mixed methods research designs in Research. Thousand Oaks, CA: SAGE.
health psychology: An illustrated discussion from a Creswell JW (2015b) Revisiting mixed methods and advancing
pragmatist perspective. British Journal of Health Psychology scientific practices. In: Hesse-Biber S and Johnson RB (eds)
20(1): 5–20. The Oxford Handbook of Multimethod and Mixed Methods
Bryman A (2006) Integrating quantitative and qualitative Research Inquiry. New York: Oxford University Press,
research: How is it done? Qualitative Research 6(1): 97–113. pp. 61–71.
Bryman A (2007) Barriers to integrating quantitative and Creswell JW and Plano Clark VL (2011) Designing and
qualitative research. Journal of Mixed Methods Research Conducting Mixed Methods Research, 2nd edn. Thousand
1(1): 8–22. Oaks, CA: SAGE.
Bryman A (2008) Why do researchers integrate/combine/mesh/ Creswell JW, Plano Clark VL and Garrett GL (2008)
blend/mix/merge/fuse quantitative and qualitative research? Methodological issues in conducting mixed methods research
634 Journal of Research in Nursing 21(8)

designs. In: Bergman MM (ed.) Advances in Mixed Methods McCabe C, Roche D, Hegarty F, et al. (2013) ‘Open Window’:
Research. London: SAGE, pp. 66–83. A randomized trial of the effect of new media art using a
Creswell JW, Plano Clark VL, Gutmann ML, et al. (2003) virtual window on quality of life in patients experiencing
Advanced mixed methods research designs. In: Tashakkori A stem cell transplantation. Psycho-Oncology 22(2): 330–337.
and Teddlie C (eds) Handbook of Mixed Methods in Social & Maxwell JA (2010) Using numbers in qualitative research.
Behavioral Research. Thousand Oaks, CA: SAGE, Qualitative Inquiry 16(6): 475–482.
pp. 209–240. Maxwell JA and Mittapalli K (2010) Realism as a stance for
Doyle L, Brady AM and Byrne G (2009) An overview of mixed mixed methods research. In: Tashakkori A and Teddlie C
methods research. Journal of Research in Nursing 14(2): 175– (eds) Handbook of Mixed Methods in Social and Behavioural
185. Research, 2nd edn. Thousand Oaks, CA: SAGE,
Doyle L (2015) Mixed methods. In: Henly SJ (ed.) Routledge pp. 145–167.
International Handbook of Advanced Quantitative Methods in Mertens DM (2010) Transformative mixed methods research.
Nursing Research. Abingdon: Routledge, pp. 411–422. Qualitative Inquiry 16: 469–474.
Drabble SJ and O’Cathain A (2015) Moving from randomized Morgan D (2007) Paradigms lost and pragmatism regained:
controlled trials to mixed methods intervention evaluations. Methodological implications of combining qualitative and
In: Hesse-Biber S and Johnson RB (eds) The Oxford quantitative methods. Journal of Mixed Methods Research
Handbook of Multimethod and Mixed Methods Research 1(1): 48–76.
Inquiry. New York: Oxford University Press, pp. 406–425. Morse JM (2015) Issues in qualitatively-driven mixed methods
Farmer T, Robinson K, Elliott SJ, et al. (2006) Developing and designs: Walking through a mixed-methods project.
implementing a triangulation protocol for qualitative health In: Hesse-Biber S and Johnson RB (eds) The Oxford
research. Qualitative Health Research 16(3): 377–394. Handbook of Multimethod and Mixed Methods Research
Inquiry. New York: Oxford University Press, pp. 206–222.
Feilzer MY (2010) Doing mixed methods research
pragmatically: Implications for the rediscovery of O’Cathain A (2010) Assessing the quality of mixed methods
pragmatism as a research paradigm. Journal of Mixed research: Toward a comprehensive framework.
Methods Research 4(1): 6–16. In: Tashakkori A and Teddlie C (eds) Handbook of Mixed
Methods in Social and Behavioural Research, 2nd edn.
Fetters MD, Curry LA and Creswell JW (2013) Achieving Thousand Oaks, CA: SAGE, pp. 531–555.
integration in mixed methods designs – principles and
O’Cathain A, Murphy E and Nicholl J (2010) Three techniques
practices. Health Services Research 48(6): 2134–2156.
for integrating data in mixed methods studies. BMJ 341:
Giddings LS (2006) Mixed-methods research: Positivism c4587.
dressed in drag. Journal of Research in Nursing 11(3):
Safer D (1997) Self-reported suicide attempts by adolescents.
195–203.
Annals of Clinical Psychiatry 9(4): 263–269.
Greene JC (2007) Mixed Methods in Social Inquiry. San
Sale JEM, Lohfeld LH and Brazil K (2002) Revisiting
Francisco, CA: Jossey-Boss.
the quantitative–qualitative debate: Implications for
Greene JC, Caracelli VJ and Graham WF (1989) Toward a mixed-methods research. Quality and Quantity 36(1):
conceptual framework for mixed-method evaluation designs. 43–53.
Educational Evaluation and Policy Analysis 11(3): 255–274.
Sandelowski M (1996) Using qualitative methods in
Hesse-Biber S (2015) Introduction. Navigating a turbulent intervention studies. Research in Nursing & Health 19(4):
research landscape: Working the boundaries, tensions, 359–364.
diversity, and contradictions of multimethod and mixed Sandelowski M (2014) Unmixing mixed-methods research.
methods inquiry. In: Hesse-Biber S and Johnson RB (eds) Research in Nursing & Health 37(1): 3–8.
The Oxford Handbook of Multimethod and Mixed Methods
Research Inquiry. New York: Oxford University Press, Shannon-Baker P (2015) Making paradigms meaningful in
pp. xxxii–liii. mixed methods research. Journal of Mixed Methods
Research, Epub ahead of print 20 March 2015. DOI:
Howe KR (1988) Against the quantitative–qualitative 10.1177/1558689815575861.
incompatibility thesis or dogmas die hard. Educational
Stoddart K, Bugge C, Shepherd A, et al. (2014) The new
Researcher 17(8): 10–16.
clinical leadership role of senior charge nurses: A mixed
Ivankova NV (2014) Implementing quality criteria in designing methods study of their views and experiences. Journal of
and conducting a sequential QUAN-QUAL mixed methods Nursing Management 22(1): 49–59.
study of student engagement with learning applied research
Stoller EP, Webster NJ, Blixen CE, et al. (2009) Alcohol
methods online. Journal of Mixed Methods Research 8(1):
consumption decision among nonabusing drinkers
25–51.
diagnosed with hepatitis C: An exploratory sequential mixed
Jellesmark A, Forysth Herling S, Egerod I, et al. (2012) Fear of methods study. Journal of Mixed Methods Research 3(1):
falling and changed functional ability following hip fracture 65–86.
among community-dwelling elderly people: An explanatory Tashakkori A and Teddlie C (2008) Quality of inferences in
sequential mixed methods study. Disability and mixed methods research: Calling for an integrative
Rehabilitation 34(25): 2124–2131. framework. In: Bergman MM (ed.) Advances in Mixed
Johnson RB and Onwuegbuzie AJ (2004) Mixed methods Methods Research: Theories and Applications. London:
research: A research paradigm whose time has come. SAGE, pp. 101–119.
Educational Researcher 33(7): 14–26. Teddlie C and Tashakkori A (2003) Major issues and
Lee C and Rowlands IJ (2015) When mixed methods produce controversies in the use of mixed methods in the social and
mixed results: Integrating disparate findings about behavioral sciences. In: Tashakkori A and Teddlie C (eds)
miscarriage and women’s wellbeing. British Journal of Health Handbook of Mixed Methods in Social & Behavioral
Psychology 20(1): 36–44. Research. Thousand Oaks, CA: SAGE, pp. 3–50.
Doyle et al. 635

Louise Doyle (PhD, MSc, BNS, RPN, RNT) is an Assistant Professor in mental health
nursing at the School of Nursing and Midwifery, Trinity College Dublin. She has
experience using a range of mixed methods, designs primarily within the area of mental
health focusing on topics including self-harm in young people, adolescent help-seeking and
the mental health of lesbian, gay, bisexual, trans and/or intersex (LGBTI) people. She has
led mixed methods projects which evaluated innovative mental health interventions
including wellness interventions for the general population. She delivers workshops for
PhD students and early career researchers on mixed methods research and has published
mixed methods studies.

Anne-Marie Brady (PhD, MSc, DSN, Pg Dip Stats, Pg Dip CHSE, RGN, RNT) is Director
of the Centre for Practice & Healthcare Innovation in Trinity College Dublin. She is an
experienced mixed methods researcher in the field of healthcare systems, evaluation and
quality improvement. She has employed a range of mixed method studies in collaboration
with health service providers to examine real world issues around health service delivery,
integration of care, system- and patient-related outcomes, safety, workload and competency
among healthcare workers.

Gobnait Byrne (MPH, Pg Dip Stats, BNS, RGN, RNT) is an Assistant Professor in general
nursing in the School of Nursing and Midwifery, Trinity College Dublin. She has recently
completed a mixed methods study exploring the health, lifestyles and use of health services
by Polish migrants living in Ireland.