Mixed-Methods Research

Mixed-Methods Research

by Michael K. Gardner,

Department of Educational Psychology,

University of Utah

“For my doctoral

dissertation, I thought I might do

a mixed-method study.

Why Do Mixed-Methods Research?

   Mixed-methods  research  has  several  strengths.  First,  mixed-method  research  can  help  to  clarify  and  explain relationships  found  to  exist  between  variables.  For  example, correlational data may dicate a slight negative relationship  between  the  time  students  spend  at  home using  a  computer  and  their  grades that  is,  as  student computer time increased, their grades suffered. The question is raised as to why such a relationship exists. Interviews with students might show that the students fell into two distinct groups: (a) a relatively large group who use the computer primarily for social nteraction (e.g., e-mail and instant messaging) and whose grades are suffering, and (b) a smaller group who use the computer for gath-ering  school-related  information  (e.g.,  through  the  use of search engines) and whose grades are comparatively high. When the two groups were initially combined, the larger number of students in the fi rst group produced the negative  relationship  found  to  exist  between  computer usage  and  student  grades.  The  subsequent  interviews, however,  showed  that  the  relationship  was  somewhat spurious, due more to  the reasons why  students used their computers, not to the use of computers per se.

Second,  mixed-methods  research  allows  us  to  explore  relationships  between  variables  in  depth.  In  this situation,  qualitative  methods  may  be  used  to  identify the important variables in an area of interest. These variables may then be quantifi ed in an instrument (such as a questionnaire) that is then administered to large numbers of individuals. The variables can then be correlated with other variables. For example, interviews with students might  reveal  that  study  problems  can  be  categorized into  three  areas:  (a)  too  little  time  spent  studying;  (b) distractions in the study environment, such as television and radio; and (c) insuffi cient help given by parents or siblings. These problems could be further investigated by constructing a 12-item questionnaire, with four questions  for  each  of  the  three  study  problem  areas. After administering  this  questionnaire  to  300  students,  researchers could correlate the study problem scores with other variables, such as student grades, standardized test performance, socioeconomic level, and involvement in extracurricular activities, to see if and how any of these other variables are related to particular study problems.

Third,  mixed-methods  studies  can  help  to  confi rmor cross-validate relationships discovered between variables, as when quantitative and qualitative methods are compared to see if they converge on a single interpretation  of  a  phenomenon.  If  they  do  not  converge,  the reasons for the lack of convergence can be investigated. For example, a professor specializing in mixed-methods research might be asked to investigate the satisfaction of middle school students with their teachers’ grading practices. He or she could prepare a questionnaire designed  to  determine  the  attitudes  of  students  and  then conduct  focus  group  with  various  samples  of  the  students. If the survey responses generally reveal satisfaction with the teachers’ grading practices, yet the focus group  participants  indicate  a  considerable  dissatisfaction with them, a possible explanation might be that the students felt that their teachers would see the responses to the surveys (and thus they were reluctant to be critical). However, in the focus groups, with no teachers or other adults present, they could feel free to express their true feelings. Thus, the apparent lack of convergence in this case might be explained by a third variable: whether teachers would have access to the results.

Drawbacks of Mixed-Methods Studies

   At this point you might wonder why all research problems  are  not  addressed  using  mixed-methods  designs. Several drawbacks exist. First, mixed-methods studies are  often  extremely  time-consuming  and  expensive  to carry out. Second, many researchers are experienced in only one type of research. To conduct a mixed- methods study  properly,  one  needs  expertise  in  both  types  of research.  Such  expertise  takes  considerable  time  to develop.   Indeed, the resources, time, and energy required to do a mixed-methods study may be prohibitive for a single researcher to undertake. This drawback can be avoided if multiple researchers, with differing areas of expertise, work  as  a  team.  However,  if  a  single  researcher  does not have suffi cient time, resources, and skills, he or she would probably be better off doing a purely quantitative or qualitative study and doing it well.

 

Are Some Methods Incompatible with Others?

   Some researchers in education (as well as other disciplines) argue  that  quantitative  methods  are  incompatible  with qualitative  methods.  They  state  that  the  basic  assumptions of  each  method  actually  prevent  the  use  of  the  other  in  the same study. Many qualitative researchers argue that qualitative  methods  are  based  on  a  point  of  view  about  the  nature of the world—that reality is constructed, not revealed. Since every individual sees the world in his or her own way, there is no single reality “out there” to be discovered; in fact, multiple realities exist. Quantitative researchers, on the other hand, reject this point of view. Still other researchers would argue that this notion of incompatibility has been overblown. Krathwohl, for  example,  has  stated  that  “quantitative  fi ndings  compress into  summary  numbers  the  trends  and  tendencies  expressed in words in qualitative reports. In many instances, counts of coded qualitative data might have produced data similar to the quantitative summaries . . . Many problems, in fact, actually require more than any one method can deliver; the answer, of course, is a multiple-method approach.”

Types of Mixed-Methods Designs

   While  quantitative  and  qualitative  methods  may  be combined in any way suitable to address a particular research  question,  certain  mixed-methods  designs  occur with enough frequency for us to look at them in detail. Three major types of mixed-methods design exist: the  exploratory  design ,  the   explanatory  design ,  and  the  triangulation design .  17   Each involves a combination of qualitative and quantitative data.

THE EXPLORATORY DESIGN

 In this design, researchers fi rst use a qualitative method to  discover  the  important  variables  underlying  a  phenomenon  of  interest  and  to  inform  a  second,  quantitative,  method.  (See   Figure  23.1 .)  Next,  they  seek  to discover the relationships among these variables. This type of design is often used in the construction of questionnaires or rating scales designed to measure various topics of interest.   In  the  exploratory  design,  results  of  the  qualitative phase  give  direction  to  the  quantitative  method,  and quantitative  results  are  used  to  validate  or  extend  the qualitative  fi ndings.  Data  analysis  in  the  exploratory design is separate, corresponding to the fi rst, qualitative, phase of the study and the second, quantitative, phase of  the  study. The  rationale  underlying  the  exploratory design is to explore a phenomenon or to identify important themes. In addition, it is especially useful when one needs to develop and test a particular type of instrument.  The illustration at the beginning of this chapter gives an example of an exploratory design. The student wants to use a qualitative method (ethnography), presumably involving  content  analysis  of  in-depth  interviews  and perhaps  other  narratives  (such  as  essays),  to  identify students’ reasons for joining a high school gang and to see how gang membership affected them. Subsequently, she would use a causal-comparative design to compare subgroups  of  students  who  had  different  reasons  for joining when they were freshmen. To do this, she would have to sort out the subgroups, using her ethnographic data. She would then collect data from them as seniors to see how these groups differ in ways suggested by the ethnography. This will require additional data collection where the preference would be for quantitative information that may require instrument development.