Translate

Jumat, 26 Agustus 2016

QUANTITATIVE dan QUALITATIF RESEARCH



Judul buku : Conducting Educational
Research
A Primer for Teachers and
Administrators
Patricia D. Morrell and James B. Carroll

Ringkasan


QUANTITATIVE RESEARCH 
     Let’s  take  a  more  detailed  look  at  the  two  schools  of  research.  We’ll  start  with quantitative  because  that  is  most  like  the  typical  five  step  process  of  problem solving and the design with which you are probably more familiar. In quantitative designs, you generally start with a question or a premise. Will using inquiry-based activities  affect  student  achievement?  If  my  students  use  this  tutorial,  will  their mathematics grades improve?  Quantitative studies are linked to statistics. Most statistical studies phrase their question in the form  of a hypothesis. A hypothesis is just a way of ormalizing your question in the form of a statement. The hypothesis may be stated in an if/then format;  for  example,  if  I  use  more  inquiry-based  activities  in  my  classes,  then student grades will increase; if I use this tutorial with my students, their scores on the state mathematics test will increase. 
Some  of  you  may  recall  the  “null  hypothesis”  from  a  statistics  class.  If  you remember  that  hrase,  you  can  state  your  hypothesis  as  a  null  hypothesis.  This means  you  are  saying  the  independent  variable  (the  factor  being  changed)  has no  effect  on  the  dependent  variable  (the  variable  being  measured);  for  example, teaching  students  using  an  inquiry  approach  will  not  result  in  any  change  in achievement  gains.  (In  this  example,  the  independent  variable  is  the  type  of instruction while the dependent variable is achievement gains.) Using this tutorial with  my  students  (independent  variable)  will  not  affect  their  scores  (dependent variable) on the state mathematics test. Quantitative researchers do this because it would be very hard to say that we have proved something will always be true when we are studying people. There are usually exceptions in any situation. So, we ask questions about how likely it is that the lack of what we thought we would see is a good explanation for what we are studying. If the lack isn’t a good explanation then  that  tells  us  that  what  we  thought  we  were  going  to  see  in  the  first  place  is  a better explanation.
You  will  become  more  comfortable  with  thinking  like  this  as  you  apply  these ideas  to  your  specific  research.  We  will  go  over  writing  a  quantitative  problem statement in detail later. And, for most cases, the quantitative problem statement for educational research does not need to look or ound like something out of a statistics primer! The important thing is to be able to clearly state what it is you are planning to investigate.
    To answer quantitative questions, you typically follow some kind of experimental design. You usually have some baseline data, try an intervention, collect more data and see if the intervention made a difference. Or you have two sets of data that you are trying to compare. For instance, to see if teaching with inquiry methods affect.
student learning more positively than teaching using direct instruction, you might test  the  students  before  and  after  using  one  method  and  then  do  the  before  and after  tests  using  the  other  instructional  method.  After  making  some  assumptions (like  the  two  units  are  comparable  in  difficulty  and  student  interest),  you  can compare the scores the students have on the tests of the two units. You would use statistical  methods  to  determine  if  one  teaching  methodology  did,  indeed,  yield greater improvements in scores than the other. As an additional example, perhaps you want to see if the age at which children started walking is related to when they started  reading.  Here  you  are  not  doing  any  intervention,  but  you  are  comparing two sets of information to see if a relationship exists between them.  Quantitative  research,  then,  is  a  structured  process  in  which  you  gather  and analyze  quantitative  data. Ideally,  it requires large, random  samples.  That  is, you would  use  your  whole  class  rather  than  a  just  a  handful  of  selected  students.
Quantitative research follows a deductive model—you go from the general to the specific.  You  have  an  idea  or  theory,  you  apply  it  to  your  sample  (class,  school, etc.), and see if it holds true. The findings tend to be generalizable to other people who are similar to those who were in your study. If teaching ith a more inquiry-like approach works for your students, it will probably work for all similar groups of students. If direct instruction worked better for your students, it would probably work  better  for  other  students  in  your  school.  Finally,  quantitative  research  is concerned with outcomes it looks at a product.

QUALITATIVE RESEARCH
Qualitative  research  designs  are  quite  different  from  quantitative  methodologies. Instead of starting with a hypothesis, you end up with one. You start with something you want to learn more about, but have no hypothesis to test. For example, rather than trying to measure student attitudes toward school, you want to find out what is it they like or dislike about school. What affects their attitudes? Rather than using an  experimental  set  up,  you  study  the  students  in  an  actual  setting.  You  don’t manipulate  how  the  teacher  teaches  or  what  students  are  in  the  class,  but  rather look at the situation in context. You will collect descriptive rather than numerical data. You might do classroom observations, you might interview students, you might have  students  journal  about  their  thoughts.  You  would  NOT  give  them  a  list  of items and ask them which statements were true about their attitudes toward school.
You would have the students generate the list. In qualitative designs, we typically select small, non-random samples to study. Rather than study all students enrolled in  a  history  class,  we  might  study  a  handful  that  we  think  possess  (or  have  been suggested  to  us  as  possessing)  positive  attitudes  and  some  that  have  negative attitudes. Because the numbers are small and the sample is non-representative, the findings  will  not  be  generalizable.  They  hold  true  only  for  that  specific  group  of students. 
In qualitative research, you don’t come up with an outcome or a product. Rather than a “what,” you usually come up with “hows and/or whys.” The data are speculative, based on the perceptions of those involved. You actually “conclude” with a  hypothesis.  This  may  provide  an  excellent  statement  to  be  tested  in  another setting  or  with  other  research  designs.  After  talking  with  and  observing  a  elect group  of  students,  you  may  come  up  with  some  notion  of  what  it  is  that  affects students’ attitudes toward school. Students might like it when they feel they have more of a say in what happens and more freedom of choice in their schools. From that  you  might  hypothesize  that  students  in  schools  with  an  active  and  strong
student  government  would  have  more  positive  attitudes  toward  school;  however, you could not conclude that—you are just guessing based on what you discovered with  this  select  group—that  assertion  is  something  that  would  need  testing (probably using a quantitative design). Qualitative research tends to be inductive.
You look at specific instances and try to come up with a generalization.
    So, quantitative and qualitative studies differ in important ways (See Figure 1-3).
1. Purpose of the study—are you looking to test something specifically or determine an  outcome  or  are  you  interested  in  coming  up  with  possible  explanations  or descriptions?
2. Kinds of data collected—are you looking at quantitative outcomes (test scores, scales,  etc.)  or  observational/descriptive  data  (you  are  the  primary  collection tool)?
3. Methods  of  data  collection—are  you  using  an  experimental  set  up  or  a  more naturalistic approach; is the process more objective or subjective?
4. Analysis of the data—are you applying statistical procedures or using inductive reasoning? 


Quantitatif
Qualitatif
numerical data    
Qualitative
deductive model     
inductive model
large, random sample        
small, purposeful sample
Generalizable
not generalizable
outcome oriented     
process oriented

SUBCAMPS OF QUANTITATIVE AND QUALITATIVE RESEARCH

While specific types of quantitative and qualitative research methodologies will be discussed in  more detail later, it is useful to provide  an  overview of  the types  of studies  that  fit  into  these  paradigms  now.  We  think  this  will  help  solidify  the similarities and differences between the two major research models.
Quantitative  categories.  Common  quantitative  designs  include experimental, causal-comparative,  correlational,  and  survey  or  descriptive  designs  (see Figure  1-4).  An  experimental  design  is  used  when  you  want  to  test  a  particular variable.  Using  our  previous  examples,  if  you  want  to  see  whether  inquiry-type learning produces different achievement results than direct instruction, you would use  an  experimental design. The  variable  being tested is the  type of instructional methodology (independent variable) and the outcome being  measured (dependent variable) is achievement. Experimental designs allow for manipulation of independent variables and make cause/effect conclusions possible by examining data that comes from dependent variables. In actuality, in most educational research studies that we will  do,  we  will  be  using  a  quasi-experimental  design  rather  than  a  true experimental design. This is because we cannot randomly select our students. We generally  work  with  the  sample  we  have  the  easiest  access  to,  and  this  is  not usually randomly assigned.
    Experimental design – testing a particular variable
 Causal-Comparative (Ex post facto) – testing a variable but  researcher cannot control the independent variable
   Correlational study – determining relationships between variables
   Descriptive quantitative study – gathering information to clarify characteristics of a group
   Survey research – relying on answers to questions
When  you  cannot  control  or  manipulate  variables  but  want  to  see  the  effect  a  variable  may  have,  you  typically  are  doing  causal-comparative  research,  also known as ex post facto research. For example, you want to determine if reading daily  to  preschoolers  (independent  variable)  affects  their  reading  readiness  in kindergarten  (dependent  variable).  You  cannot  (and  should  not)  determine  which preschoolers  get  read  to,  so  you  compare  two  intact  kindergarten  groups—those that had been read to with those that had not. It is similar to an experimental design in that groups are being compared to determine an outcome (the differences in the dependent  variable);  but  dissimilar  in  that  the  experimenter  does  not  control  the independent variable (the factor being changed).
13
    Quantitative  categories.  Common  quantitative  designs  include experimental, causal-comparative,  correlational,  and  survey  or  descriptive  designs  (see Figure  1-4).  An  experimental  design  is  used  when  you  want  to  test  a  particular variable.  Using  our  previous  examples,  if  you  want  to  see  whether  inquiry-type learning produces different achievement results than direct instruction, you would use  an  experimental design. The  variable  being tested is the  type of instructional methodology (independent variable) and the outcome being  measured (dependent variable) is achievement. Experimental designs allow for manipulation of independent variables and make cause/effect conclusions possible by examining data that comes from dependent variables. In actuality, in most educational research studies that we will  do,  we  will  be  using  a  quasi-experimental  design  rather  than  a  true experimental design. This is because we cannot randomly select our students. We generally  work  with  the  sample  we  have  the  easiest  access  to,  and  this  is  not usually randomly assigned. 

   Experimental design – testing a particular variable
   Causal-Comparative (Ex post facto) – testing a variable but 
researcher cannot control the independent variable
   Correlational study – determining relationships between variables
   Descriptive quantitative study – gathering information to clarify
characteristics of a group
   Survey research – relying on answers to questions

When  you  cannot  control  or  manipulate  variables  but  want  to  see  the  effect  a  variable  may  have,  you  typically  are  doing  causal-comparative  research,  also known as ex post facto research. For example, you want to determine if reading daily  to  preschoolers  (independent  variable)  affects  their  reading  readiness  in kindergarten  (dependent  variable).  You  cannot  (and  should  not)  determine  which preschoolers  get  read  to,  so  you  compare  two  intact  kindergarten  groups those that had been read to with those that had not. It is similar to an experimental design in that groups are being compared to determine an outcome (the differences in the dependent  variable);  but  dissimilar  in  that  the  experimenter  does  not  control  the independent variable (the factor being changed). 
         A correlational study is used when you want to determine if a relationship exists between  two  or  more  variables.  Is  there  a  relationship  between  socio-economic status  and  performance  on  open-ended  mathematics  tasks? Is  there  a  relationship between  homework  completion  rate  and  chapter  test  scores?  These  are  not cause/effect studies, but help to determine the degree of a relationship that exists between/among variables. It may be enough for your study’s purposes to understand that  two  variables  behave  in  the  same  (as  one  increases,  the  other  increases)  or opposite  ways  (one  increases  while  one  decreases)  but  often  the  results  of correlational studies help in the design of further experimental studies.  
       Descriptive  quantitative  studies  gather  information  about  a  group  so  that  it  is easier  to  understand  the  nature  of  the  group.  The  purpose  of  the  study  is  not  to compare groups or to understand the impact of an intervention, but simply to have a clearer picture of the characteristics of a group. As an example, you might want to track changes in your school district’s yearly expenditures over time to understand changing funding sources. In this case you are asking a descriptive question about the group being studied and not comparing groups.
        A  form  of  descriptive  research  is  survey  research.  As  the  name  suggests, survey research relies on answers to questions. Unlike qualitative descriptions, the data collected in this type of design are quantitative. Those in the sample respond to  provided  statements  such  as  “choose  from  the  following”  or  “rate  on  a  scale from one to five.” Typical data collection instruments include tests and questionnaires; pencil  and  paper  forms  that  the  respondents  can  complete  on  their  own  (or  via
telemarketers).
Qualitative categories. Qualitative research can be categorized in a myriad of ways,  but  the  main  types  typically  are  grounded  theory,  ethnography,  case study, phenomenology, and historical research. (See Figure 1-5) In a grounded theory  approach,  the  purpose  is  to  collect  and  analyze  data  with  the  intent  of coming  up  with  a  theory  that  explains  the  research  situation.  It  is  considered “grounded”  because  it  develops  out  of—or  is  grounded  in—collected  data.  You would use  this approach when existing theories do  not apply  to your situation or context and when studying new phenomena. You are basically trying to understand a research situation; that is, what is happening and how those involved feel about it.  As  an  example,  you  may  be  interested  in  understanding  why  conflict  exists between parents  of children in  special  education  programs and  school  officials or  how  to  best  integrate  the  use  of  interactive  white  boards  in  a  classroom.  For grounded  theory,  the  data  collection  and  analysis  process  is  an  iterative  cycle  of collecting information, identifying commonalities in the data, and verifying those commonalities,  narrowing  the  key  components  as  you  repeat  the  cycle.  The outcome is a potential theory that explains the relationship. 

   Grounded theory – generating a theory to understand a problem/situation
   Ethnography – examines the “culture” of a group
   Case study – detailed account of an individual or group
   Phenomenology – looking at something through the eyes of those being observed 
   Historical Research – understand events that have already occurred

Another  common  qualitative  approach  is  an  ethnography.  In  an  ethnography, you are interested in studying the “culture” of a group. You might examine how a group is formed, how the members interact, and what type of interactions occur. You might be interested in looking at “skateboarding culture” or behaviors of teachers in a teachers’ lounge. One of our colleagues studied a group of graffiti artists to get a deeper understanding and insight into that subculture (Christen, 2003). 
    A case study is considered by some to be a subset of ethnography. Others consider it a separate type of qualitative design. As its name implies, a case study is an indepth examination of one or more special “cases.” A case can be a specific person or  a  small  group.  It  is  similar  to  an  ethnography,  but  has  a  more  defined/limited focus.  One  of  our  students  was  doing  her  practicum  in  a  grade  school  that  was gradually  implementing  a  new  mathematics  curriculum.  Teachers  were  given  the option  of  when  and  how  they  wanted  to  start  using  the  newly  adopted,  student-centered  program  as  long  as  everyone  had  it  fully  adopted  by  the  end  of  three years. Our student chose a small group of teachers to study, which included early adopters,  resisters,  and  some  fence  straddlers.  She  examined  how  they  felt  about the change, about mathematics in general, and about student learning. She hoped to understand why there was such a difference in attitude and behavior concerning the
implementation of this new type of curriculum. 
    Phenomenology  is  looking  at  something  (the  phenomenon)  through  the  eyes of a subject group. That something might be an event or an interaction; the focus is on understanding the perspectives of those you are studying. You attempt to find the  “commonalities  of  their  perceptions”  (Slavin,  2006,  p.  147).  Phenomenology applies  a  social  constructivist  viewpoint  to  the  research  process.  You  might  be interested in studying student teacher/mentor teacher interactions or how students deal with the loss of a classroom pet.
    Historical research is used to help us understand events that have already occurred. We worked with a teacher who had started the school year in a new district. The physical design of her school building was very different from any she had worked in or seen. Her research question was why the school was designed as it was. She undertook a historical research project, examining school board minutes, looking
at  architectural  plans,  speaking  with  members  on  the  school  building/planning committee, to gain an understanding of the reasons for the physical configuration of the building. 
    As  with  anything,  these  subcategories  of  quantitative  and  qualitative  research are not exhaustive. There are other distinctions and subcategories within subcate-gories. These are just the more common approaches you’re likely to encounter or pursue yourself.

MIXED METHODS
As mentioned earlier, sometimes a research study uses a combination of qualitative and quantitative methods. You may not be interested in just what teachers’ attitudes are  toward  state-mandated  curricula  (which  can  be  measured  using  a  scaled instrument)  but  also  why  they  feel  that  way  (answered  by  having  them  answer some open-ended or interview questions). You might want to know what methods neighboring districts use to identify talented and gifted students and how teachers feel about those methods. You may want to examine if and how student achievement. scores  are  affected  by  switching  to  PowerPoint  presentations  and  what  students think  about  your  use  of  that  type  of  medium.  Using  a  mixture  of  qualitative  and quantitative  methods  in  one  study  is  perfectly  acceptable,  and  in  some  cases, necessary.  Your  question  dictates  the  methodology  or  methodologies  you  should use! (You’ve heard that before.)
In  a  study  that  uses  mixed  methods,  the  methods  can  be  used  sequentially  or simultaneously (see Figure 1-6). For example, you might initially administer a quantitative  questionnaire  to  gain  information  on  students’  attitudes  toward  a  specific content area. You would then use those results to help identify a sample of students with a range of attitudes from positive to negative to interview for a deeper insight into  the  whys  of  these  attitudes.  This  is  referred  to  as  an  explanatory  design:  the quantitative portion precedes the qualitative data collection. Conversely, you  might follow an exploratory design, carrying out qualitative methods prior to quantitative methods. You might interview a sample of students first to determine their attitudes toward a subject and the factors that affect that attitude. After analyzing the data to determine pertinent themes and patterns, you would use those findings to design a quantitative questionnaire that could be administered to a larger group of students.
    At  times  your  question  might  require  that  you  collect  both  types  of  data simultaneously  or  concurrently.  You  might  be  integrating  technology  into  your instruction  and  you  want  to  determine  if  (a)  this  change  in  methodology  will impact  quantitative  student  achievement  and  (b)  whether  students’  behaviors  in class are affected by the implementation of technology. You would use a quantitative design for the former and a qualitative design for the latter and be conducting both
at the same time. This is referred to as a triangulated design.
    In  these  types  of  mixed-methods  studies,  you  actually  have  a  twopronged  or compound  research  question.  One  part  of  the  question  would  be  answered  using quantitative methods and the other using qualitative means. You could design the study  as  if  you  are  doing  two  studies,  using  both  types  of  methods  and  data analysis, but base your conclusions on the two sets of findings. 

Explanatory
Quantitative precedes Qualitative
Exploratory
Qualitative precedes Quantitative
Triangulated
Quantitative occurs along with Qualitative

ACTION RESEARCH
There is one other main “body of research.” It is action research. Action research can be qualitative or quantitative or use mixed methods. It follows the same planning and guidelines as other types of research. What sets it apart, is that action research is generally undertaken by individuals for their own, personal purpose as opposed to  educational  researchers  in  college,  universities  or  think  tanks.  Teachers  and school administrators look at a problem in their own classroom/school/district to be able to gather information and make an informed “action plan.” The purpose is generally to improve one’s teaching or address a specific, local concern. Is the new tardy  policy  working?  Is  implementing  judicious  discipline  in  my  history  class helping  students  learn  more  deeply  about  democratic  principles?  Is  using  the student response system a more effective way of formative assessment than using non-technological  means?  Action  research  is  not  undertaken  to  have  the  findings apply to any situation except the one that is studied.
    Mills (2007) categorizes action research into two main types: practical action research and participatory action research. Practical action research is research that addresses specific questions in a classroom, school, or district. The purpose is  to  improve  teaching  and  learning  or  provide  necessary  information  to  help  in decision making. Participatory action research differs from practical action research in that it involves a group of people focusing on the same problem (e.g., teachers, administrators, board members). This group of stakeholders collectively formulates the research problem, which leads to this type of action research also being dubbed collaborative action research. Additionally, participation action research attempts to  “improve  social  practice  by  changing  it”  (McTaggart,  1989,  Tenet  1).  It  is  an iterative  process  of  planning,  implementing,  observing  the  effects  of  the  implementation, reflecting on the effects, and revising the action plan. 
While there are entire books written just on action research (e.g., Mills, 2007), the  basic  research  guidelines  needed  in  conducting  such  a  project  parallel  those outlined in this text. Depending on your question, you would follow the guidelines for  conducting  a  qualitative,  quantitative  or  mixed  methods  study.  Once  you  are done with your data analysis, action research requires that you take one additional step. The implications section would be your action plan. Now that you have studied
the issue, what do you propose be done in response to the problem?
A  main  drawback  to  conducting  action  research  is  that  the  researcher  is  an inherent  part  of  the  process,  making  it  harder  to  maintain  objectivity.  The  major advantage is action research focuses on self-reflection; and, because you are studying something  that  is  relevant  to  your  immediate  professional  practice,  it  is  easy  to keep  focused  on  and  enthused  about  the  research  question.  You  will  be  taking  a systematic and critical look at an issue that will ultimately improve your effectiveness as a professional. Those of you who will be examining your own teaching practices or classroom policies may be conducting a study very similar to action research. 

NEXT STEPS
Now that you are armed with a basic understanding of why you should know how research is conducted, why you should learn about designing and conducting your own  research  project,  and  the  basic  schools  of  research  methodologies,  the  next logical step is to begin thinking about your own research project. It’s time to start drawing  on  the  empowerment  that  you  will  get  from  becoming  an  educational researcher. 

CHAPTER SELF-CHECK
Having completed this chapter, you should be comfortable discussing the following:
  the importance of educational research
  what it takes to be an educational researcher
  the typical format of an educational research paper
  a comparison of qualitative and quantitative paradigms
  the description of the following quantitative methodologies: experimental design,
causal-comparative, correlational, survey, descriptive
  the description of the following qualitative methodologies: grounded theory, ethno-
graphy, case study, phenomenology, historical research
  mixed methods research
  what constitutes action research

CHAPTER REVIEW QUESTIONS
1. What is the purpose of educational research?
2. Why  is  it  important  that  you  understand  the  basics  of  conducting  educational research?
3. How is a typical research paper organized? What is included in each chapter?
4. What are the major differences between qualitative and quantitative research?
5. The  text  describes  five  main  types  of  quantitative  research.  What  are  these? Provide a brief description of each.
6. The text also describes five main types of qualitative research. What are these? Provide a brief description of each.
7. Is  it  possible  to  mix  qualitative  and  quantitative  methodologies  in  the  same research study? Why or why not?
8. Compare and contrast action research with qualitative and quantitative schools of research.

REFERENCES
Bodgan, R. C., & Biklen, S. K. (2006). Qualitative research for education: An introduction to theory
and methods (5th ed.). Boston: Allyn and Bacon.
Borko,  H.,  Liston,  D.,  &  Whitcomb,  J.  A.  (2007).  Genres  of  empirical  research  in  teacher  education.
Journal of Teacher Education, 58(1), 3–11.
Christen,  R.  S.  (2003).  Hip  hop  learning:  Graffiti  as  an  educator  of  urban  teenagers.  Educational
Foundations, 17(4), 57–82.
Erickson, F. (1986). Qualitative methods in research on teaching. In M. C. Wittrock (Ed.), Handbook of
research on teaching (3rd ed., pp. 119–161). New York: Macmillan.
Gardner, H. (1983). Frames of mind: The theory of multiple intelligences. New York: Basic.
Gardner, H. (1999). Intelligence reframed. Multiple intelligences for the 21st century. New York: Basic
Books.
McTaggart, R., (1989). 16 tenets of participatory action research. A paper presented at the Third World
Encounter on Participatory Research, Managua, Nicaragua. Retrieved from http://www.caledonia.org.
uk/par.htm

Tidak ada komentar:

Posting Komentar