Qualitative, Quantitative, and Mixed Research Methods

Introduction

Research is a critical educational tool. Scientific, as well as other disciplinary studies are conducted to expand the existing knowledge base by making new discoveries, expounding on them by highlighting new developments and connections, or correcting them by gathering relevant evidence to counter the already recorded data. Every research is therefore beneficial, even if it just confirms already known concepts. However, the war between qualitative researchers, mixed methods and quantitative researchers on the superiority of their respective approaches is a derailing factor for inquiry in general. Though they look similar, they feature significant differences based on the type of data used (textual vs. numerical), logic (inductive vs. deductive), investigation criteria (explanatory vs. confirmatory), data analysis (interpretative vs. statistical), theoretical and conceptual basis (interpretative/ critical vs. positivist) and/or (naturalistic vs. rationalistic), amongst other parameters (Driscoll, Afua, Salib, & Douglas, 2007, pp. 19-28). The irony of their conflict is that either approach applies procedures of its opponent in compiling the research literature. It seems possible to codify and enumerate qualitative data. However, quantitative data requires interpretation, which translates to the usage of words in explanation. Evidently, there is a clear distinction between the two methods on the type of data. Qualitative research uses words in plenty while quantitative research is statistical. The deficiency of each approach’s independent operation is a glaring reality that begs for redress. Mixed research is coming up as a different approach from these traditional methods. (Driscoll, 2007, pp. 25) incorporating both qualitative and quantitative research methods while by taking the best of each side. The result is an efficient hybrid. However, this method is still at its infancy, and requires further study to refine and equip it with a comprehensive working framework. This paper outlines an in-depth analysis of the definitions, descriptions, comparisons and contrasts of all three methods, including detailed insight into their strengths, weaknesses, reliability and validity as research methods.

Definitions

Qualitative research encompasses all studies that result in findings not derived from any form of statistical procedure or any means of quantification. Data is in the form of words, images, objects and categories (Driscoll et al., 2007, pp. 19-28). It is, “an empirical enquiry that investigates a contemporary phenomenon within is real life context using numerous forms of evidence, when the boundary between the phenomenon and context is vague” (Yin, 2006, pp. 41-47). At the basis of this research method, is the tendency to focus on human experiences and reaction to various phenomena including the study of socio-cultural concepts, which form the context or setting of the study.

Quantitative research uses strictly statistical or numerical data. It quantifies variables and draws conclusions from the findings of the analysis of these numbers collected. Large samples are usually collected which can be analyzed and the conclusions drawn can be generalized. “Generalizability” of findings is a distinctive feature, and a clear goal of this type of research (Teddlie, & Tashakkori, 2003, pp. 36). This approach involves an investigation research questions that are predetermined and take the form of a hypothesis. It has a clear-cut framework that is uniformly applied during research by all its proponents, and it is based on a positivistic school of thought.

Mixed research incorporates both qualitative and quantitative research approaches as well as their paradigm characteristics in its methodology. However, it is selective in this incorporation, seeking to include only “complementary strengths and non-overlapping weaknesses of each” (Teddlie, & Tashakkori, 2003, p. 43). Consequently, this method is a hybrid of qualitative and quantitative research. The mixing can occur at one stage or across multiple stages of the research process.

Descriptions

Qualitative research as defined above involves collection of non-numerical data in the form of words, images, and objects. The main idea is to understand the ‘whole’ or gain a holistic comprehension, using knowledge gained from critical, in-depth exploration of a particular phenomenon done using an “inverted” approach that is inductive in nature. The researcher generates a ‘new’ theory and hypothesis from the data collected in the field. There exist five types of qualitative research, which include ground theory, where the data collected is used to develop a theory as well as phenomenological research where the researcher attempts to understand how one or more individuals experience certain phenomena, without the backing any theories, deductions, or assumptions from other disciplines. There is also the case study, which gives a detailed account of the reaction of one or more specific: individuals, groups, events or institutions to a certain phenomenon (Yin, 2006, pp. 41-47). Ethnography, as another type, focuses on describing the culture of a particular community, where culture refers to “…shared attitudes, values, norms, practices, language, and material property of a group of people” (Yin, 2006, p. 45). This is done through keen observation of these socio-cultural characteristics, Historical- orderly collection and unbiased examination of data on past events to test feasibility of hypothesis on causal trends in a bid to explain present events.

Qualitative data collection

The three main methods used in the qualitative method are the interactive interviewing where people are questioned for verbal responses on their description of various phenomena, written descriptions by participants who provide researches with written accounts of their understanding of various phenomena. Lastly, observation- researchers make detailed observations of verbal and non-verbal behavior among participants.

Data analysis

After the first collection of data, analysis follows to provide an insight on the possible future collections. Data is examined to identify the patterns, themes, and holistic features (Maxwell, 2005, pp. 52-60). Various Qualitative Data Analysis (QDA) software programs can be applied during analysis. While generating the collected data, the researcher has an obligation of availing the process of the study for purposes of vetting the authenticity or accuracy of the research by other parties. He/she must also write descriptively including nuanced descriptions as well as well thought interpretations. These findings are presented in a “narrative report which focuses on the local, the personal, and the subjective” as Maxwell (2005, pp. 52-60) points out. However, the method has its strengths as well as flaws.

Strengths/Weakness

Findings remain characterized by rich, in-depth, and nuanced descriptions capable of pointing out to the reader the idiosyncrasies of the situation. It also results in a “fine hermeneutic study” characterized by “penetrating, and context sensitive interpretations.” (Maxwell, 2005, p. 63). This means that qualitative research’s reports or findings are expressed in detailed explanations that enable a reader to understand both hypotheses and findings. As such, one just needs to read them to understand the idea, unlike in quantitative research where expertise in interpreting values and symbols is required for comprehension of methods, findings and analyses. However, the data collected is subjective. It is a result of a researcher’s personal interpretation of the situation based on his values and attitudes hence may be inaccurate. Secondly, the conducting of the research takes place based on small samples to ensure detailed study of certain phenomena. Consequently, it proves impossible to generalize the findings in case one wishes to infer about the larger population.

Validity and Reliability

Credibility, transferability, dependability and confirm-ability are some of the measures used to measure the validity of qualitative research. Others include the “descriptive validity, which refers to the factual accuracy of the account as documented by the researcher and the interpretive validity: the extent to which an interpretation of the account represents an understanding of the perspective of the underlying group and the meanings attached to the members’ words and actions. Theoretical validity too refers to the degree to which “a theoretical explanation developed from research findings is consistent with the data” (Onwuegbuzie, 2006, p. 52) amongst others. This method is not so reliable due if one intends to use the findings for generalization purposes; the usage of small samples for detailed research makes it only useful for conceptualized studies.

Quantitative Research

Quantitative research is deductive in nature, meaning that the researcher sets out to the field with a predetermined theory and hypothesis that he/ she seeks to prove or disprove. This translates to the presence, and analysis of variables (something with varying possible values or categories). The different variables can be classified using two main criteria: Measurement Level: Categorical variables, which denote different kinds of a phenomenon, e.g. gender; male &female.); Quantitative variables (variations in degree/ amount of phenomenon, e.g. temperature; zero- high temperatures). Variable’s role: Independent variables (IV), which cause changes in other variables; Dependent Variables (DV) – affected and caused to change by another variable; Mediating/ Intervening Variables- intermediary variable that delineates the change process between variables; Moderator Variables- delineates how a relationship of interest changes under different conditions or on the introduction of new features. Data is collection as samples, which may be either random or organized. However, large samples are collected o facilitate the goal of generalization of findings (Driscoll et al, 2007, p. 26).

There are several methods applied in quantitative research. Among them are experimental method, which involve the manipulation of an independent variable, and the use of a randomly assigned group of participants to establish a causal (cause-effect) relationship. It is characterized by the presence of a ‘control experiment’, which is a similar group of participants upon whom the ‘treatment’ or manipulation is not applied. Consequently, the findings will provide insight on efficiency of the manipulation technique. Using random assignment ensures that the groups are “equated” hence the possibility of having alternative explanations for findings due to the existence of extraneous variables is eradicated (Driscoll et al., 2007, pp. 27). Non-experimental- neither manipulation, nor random assignment of participants is used. Therefore, one cannot use this method to establish causal relationships. It can either be used for Causal- Comparative Research: a categorical (IV) is compared to a quantitative (DV), or Correlation Research- comparison between a quantitative (IV) and a quantitative (DV), where the relationship is drawn by calculating the correlation coefficient (ranges between -1 and +1, 0 ‘indicates no relationship’). If the coefficient is negative, the relationship is of inverse proportionality, if positive, it is of direct proportionality (Teddlie, & Tashakkori, 2003, pp. 43).

Data collection and analysis

The data is collected using precise measurement instruments such as fixed/ close-ended questionnaires, rating scales and behavioral responses. It is then subjected to statistical analysis aimed at inferring statistical relationships.

Strengths/Weaknesses

Mathematical precision is guaranteed due to use of measurement, which enhances accuracy of findings, and informs precise conceptualization. Large samples used make ‘time and context-free’ generalization possible. There is use of random assignment, which guarantees integrity of findings of causal relationships. Its objective quality keeps bias in check by numerical data. However, due to the mono-dimensional direction of research, there occurs neglect of phenomena that may be crucial for the research question. Standardized research instruments that are not designed to capture explanations may leave out critical data, especially where researchers lack any culture-specific bases of knowledge.

Validity

Internal validity, which denotes the fundamental requirements that guarantee interpretability of an experiment) of quantitative research is hindered by numerous factors, among them being history: events occurring between the 1st and 2nd measurements. Maturation as another barrier denotes the changes within participants due to passage of time, e.g. ageing, and exhaustion; testing- the effect of participation in the experiment on the results of the second experiment, and instrumentation- any changes in calibration or scorers reflect on the next findings. External Validity, which determines whom the findings can be generalized upon, can be adversely affected by factors such as reactive effects of experimental arrangements, interaction, and multiple treatment interference- where the effects from the previous measurement cannot be erased, hence have a bearing on a second measurement. This method is reliable due to the accuracy of the statistical findings, which is guaranteed by the use of precise measurement methods. However, the lack of subjectivity associated with this method limits its usage to scientific research.

Mixed Research

The mixed research principle directs that, when planning mixed research, the researcher mixes methods or procedures in a way that the resulting mixture or combination has complementary strengths and non-overlapping weaknesses” (Teddlie, & Tashakkori, 2003, p.42). Mixed research involves the synthesis of qualitative and quantitative methods (during one phase of research or across several phases) and findings. Findings can be synthesized through assimilation where similar findings (confirmatory) are incorporated into each other or repetitive findings pooled together to result in one trend of information with a multiplicity of sources to pose as evidence. This process borrows heavily from aggregative analysis (quantitative approach directed at identifying recurring findings from report studies). Configuration: findings that complement each other are “meshed” in such a form so that they build on each other in n orderly manner. This can be reflected in either a conceptual model, or a meta-narrative. This process bases on interpretative or rather qualitative) approaches characterized by discovery of new phenomena, or modification of already-existent phenomena.

Mixed research designs include the triangulation, embedded, exploratory and explanatory designs (Yin, 2006, pp. 45). They involve a sensible/eclectic collection of both forms of data that is guaranteed to research problems at hand in this context of study. It is a balanced approach to research, which wholly incorporates both inductive and deductive methods to corroborate and compliment findings. It is characterized by two types of methods. Mixed Method Research for instance involves the application of a qualitative research paradigm for one phase, e.g. intensive interviews for data collection and then a quantitative research paradigm for the next phase, e.g. quantification of qualitative data through codification (Maxwell, 2005, p. 55). In Mixed Models Research, both paradigms are mixed either within one stage of the study, or across different stages of study.

Data collection

Concurrently, the mixed method data collection is used for the validation of one form of data by another as well as converting data for comparison, or addressing different types of research questions because it is intuitive for participants as it links structured and unstructured responses. However, it may sometimes preclude follow-ups on interesting or confounding responses. In the sequential mixed method data collection, the data collected in one phase is used to direct data collection for the next phase.

Data analysis

This involves the conversion, and integration of different forms of data to generate balanced findings. Mostly, the most difficult conversion involves quantification of qualitative data, which remains detailed and based on varying themes and patterns (Driscoll et al., 2007, p. 25). Several software programs have been formulated to take up this task. These include the QCA- Qualitative Comparative Analysis software, Atlas, NVivo, and winMAX among others (Onwuegbuzie, & Johnson, 2006, pp. 48-63). They analyze data by counting the occurrence of particular qualitative codes; enumeration of themes; calculating: the percentage of themes associated with a given category of respondents; the percentage of people selecting specific themes, and quantifying the occurrence or absence of each code per participant (Onwuegbuzie, & Johnson, 2006, pp. 48-63).

Strengths/Weakness

The method comes in as a ‘hybrid’ of both qualitative and quantitative research therefore building on strengths and limiting weaknesses of both methods. It is therefore a good bargain for most study questions. However, it is not always the best approach, as the research question should determine the method of research. Some of its features or ‘improvements’ include the use of qualitative plot studies to counter the limitation of small samples by guiding selection of participants, using responses to understand complex or contradictory statistics. Further, using statistics has complemented qualitative interviews by identifying structural constraints that interviewers were previously ignorant. The method attains corroboration of findings where similar findings are drawn using different methods thus giving authenticity to the results. The method too discovers what would have been missed if only one approach had been applied. The set of results is expanded with an inclusion of important data on “emergent or unexpected themes” (Onwuegbuzie, & Johnson, 2006, p.63). However, the method lacks the capacity to formulate an integrated presentation technique that incorporates a balanced share of both methods. Quantification of qualitative data results in loss of depth and flexibility of the data; this method is tedious, time consuming, and expensive; the use of generalizations and specific statistical procedures is hindered; most researchers are either qualitative or quantitative and they have no knowledge base on the method they do not conform to. This translates into a lack of skilled mixed method researchers.

Validity and Reliability

Validity, termed as ‘legitimation’ appears in several stages including sample integration, which directs matters of drawing generalized conclusions from inferences obtained from samples as well as the inside-outside stage where both ‘emic’ and ‘etic’ perspectives should be incorporated in data analysis and collection (Onwuegbuzie, & Johnson, 2006, pp. 48-63). Weakness minimization too targets at limiting overlapping weaknesses from qualitative and quantitative methods. Generally, this research method is the most reliable because it incorporates the best practices and procedures of qualitative and quantitative research while limiting the margin of error.

Comparisons

Both qualitative and quantitative methods are similar to the mixed research method. Both also use each other’s paradigm characteristics on their research literature at the aggregate level (Yin, 2006, pp. 41-47). The data acquired is similar with differences only arising in format.

Contrasts

The scientific method or rather the qualitative analysis uses an inductive method that involves formation of a hypothesis and theory based on the research findings, unlike quantitative research, which is deductive, formulating a hypothesis and establishing a theory before embarking on the research. Mixed research integrates both qualitative and quantitative methods (Onwuegbuzie, & Johnson, 2006, pp. 48-63). Based on the perceived objectives, quantitative research sets out to make explanations and predictions whereas qualitative research aims at exploring and discovering phenomenal experiences. Based on the focus, quantitative research is mono-dimensional, fixed, and dichotomous in nature while qualitative research is multi-dimensional, flexible and in-depth (Onwuegbuzie, & Johnson, 2006, pp. 48-63). Considering the setting, qualitative research observes phenomena in their natural setting unlike quantitative research, which opts for a ‘controlled’ setting. Mixed research uses a variety of contexts. In terms of data, quantitative research collects and records data in the form of numbers and statistics. Variables are the most fundamental form of data. It also uses measurement as a data collection method, and presents findings in the form of statistical reports unlike qualitative research, which is big on words, the use of images, objects, and categories; it uses qualitative data collection procedures such as interviewing and observation, and presents findings in the form of detailed narrative reports. Mixed research integrates all these forms of data. In quantitative research, data analysis is through statistical analysis, and findings are generalized while qualitative analysis is characterized by a reflection of themes and patterns while findings are particular in nature, and may be interpreted differently as per reviewers’ perspectives. Mixed method uses both qualitative and quantitative methods of analysis to come up with corroborated finding. It uses a pragmatic approach to present data. The qualitative researcher’s view on reality is objective, unlike that of the quantitative researcher, which is subjective. A mixed method researcher is more open minded, harboring a pragmatic view of reality that accepts whatever serves him best as ‘real’ as Onwuegbuzie and Johnson (2006, p. 64) points out.

Conclusion

In this paper, I have analyzed and made comparisons between three major research types: qualitative, quantitative, and mixed research methods. I have listed the strengths and weaknesses of each method, analyzed validity and reliability, and given a brief description of each method. Consequently, my findings convey that it is high time proponents of qualitative and quantitative research stopped feuding, and instead focused on the acquisition of extensive knowledge of the respective fields to acquire expertise in mixed research.

References

Driscoll, D., Afua, A., Salib, P., & Douglas, J. (2007). Merging Qualitative and Quantitative Data in Mixed Methods Research: How To and Why Not. Journal of Ecological and Environmental Anthropology , pp.19-28.

Maxwell, J. (2005). An interactive approach (2nd. ed.). In Qualitative research design Newbury Park: Sage Publications.

Onwuegbuzie, A., & Johnson, P. (2006). The Validity Issue in Mixed Research. Mid-South Educational Research Association: Research in the Schools , 13 (1), pp. 48-63.

Teddlie, C., & Tashakkori, A. (2003). Major issues and controversies in the use of mixed methods in the social and behavioral sciences. Thousand Oaks: Sage Publications.

Yin, R. (2006). Mixed methods research: Are the methods genuinely integrated or merely parallel? Research in the Schools , 13 (1), pp. 41-47.