Senin, 26 April 2010
Senin, 19 April 2010
FOR SMP/MTS STUDENT
para siswa SMP/MTs yang kami sayangi dimana saja berada, mari kita berlatih ngomong/talk/discuss bahasa Inggris dengan mengunjungi (mendownload) or buka situs :www.syahrani.blogspot.com
ENGLISH SIDE
CHAPTER 4
PRINCIPLES OF RESEARCH DESIGN
After a researcher has identified a research problem and has completed at least some review of the literature, it is time to develop a research design - a plan or strategy for conducting the research. This chapter focuses on general principles of design, but because of process differences, qualitative and quantitative research are discussed separately.
The Purposes of Research Design
Kerlinger (1986, p.280) identified two basic purposes of research design are
(1) To provide answers to research questions and (2) to control variance.
The first purpose is general and straightforward – to provide answers to the specific research questions. Research should be valid, which includes being able to interpret results and, through those results, answer the research questions or problems being posed. Good research design assists in understanding and interpreting the results of the study and ensures that a researcher obtains usable results.
All research is conducted for the purpose of explaining variance – the fact that not all individuals are the same or have the same score or measurement. In a broad definitional sense, this may be true, but variance can be evident in a number of ways. For example, variance in elementary school students’ achievement, motivation, attitude, age, and family background can be considered. Also, when the variance of any one variable is considered, it may influenced by any number of factors. Variance in achievement, for example, may be due to aptitude and motivation, to mention two possible factors.
Variance is considered somewhat differently in qualitative and quantitative research. In qualitative research, the researcher attempts to explain the phenomenon understudy, including the variance, through description of a logical interpretation of what has been observed. The interpretation is based on a holistic concept of the situation. As we will see later in the chapter, in quantitative research more direct design procedures are taken for controlling variance than in qualitative research. Variance takes on quantitative meaning and, at least to some extent, variance is partitioned as attributable to variables under study.
Design in Qualitative Research
Qualitative researchers, for the most part, do research in natural settings; they do not manipulate or intervenes (except possibly by their presence) in the situation. Therefore, research design requires flexibility and a tolerance for adjustment as the research progresses. Smith and Glass (1987, p. 259) refer to this as a working design, similar to what McMillan and Schumacher (1989, p. 179) call and emergent design. From the identification of the research problem, decisions must be made about beginning the study.
1. Working Design
The Working design is the preliminary plan by which the research gets underway. Decisions are made about the subjects or sites to be studied, the length of time for data collection, and possible variables to be considered.
2. Working Hypotheses
Qualitative research uses the induction model, which conceptually means that data collection commences without any preconceived theories or hypotheses. However all researchers are influenced by their own backgrounds, and some information is likely to be available about the research problem. Earlier the concept of foreshadowed problems was introduced. Although technically these are not hypotheses statements, foreshadowed problems come in at this point.
3. Data collection
Methods of data collection include observation, interview, and the collection and review of related documents. In the dropout example, the researcher might engage in the following activities, although data collection certainly would not be limited to these.
1. Interview students and faculty, including guidance counselors.
2. Observe the interaction taking place between students and between students and faculty.
3. Review school records relative to factors such as grading patterns.
4. If in any way available, interview recent dropouts.
The data record of a qualitative research study can become quite massive with all the interview and observation protocols, document information, and so on. As recommended by some authors (Bogdan and Biklen, 1982), researchers should keep written accounts of their own thinking about what is being collected.
4. Data Analysis and Interpretation
Data analysis in qualitative research begins soon after data collection begins, because the researcher checks on working hypotheses, unanticipated results, and the like. In fact, data collection and data analysis usually run together; less data are collected and more analysis is produced as the research progresses. Qualitative data analysis requires organization of information and data reduction. The data may suggest categories for characterizing information. Comparisons can be made with initial theories or working hypotheses. Early data collection might suggest a hypothesis or theory, and then more data might be collected that would support, disconfirm, or extend the hypothesis or theory.
Data analysis in qualitative research is a process of categorization, description, and synthesis. Data reduction is necessary for the description and interpretation of the phenomenon under study.
Components of Research Design in Qualitative Research
Design in Quantitative Research : The concept of Controlling Variance
One of the purposes of research design in quantitative research is to control or explain variance. This is done by putting certain restrictions on the research conditions, such as including control variables.
Procedures for Controlling Variance
This section continues the chemistry example and illustrates procedures by which control of variance can be enhanced. There are basically four ways by which this can be done:
1. Randomization.
2. Building conditions or factors into the design as independent variables.
3. Holding conditions or factors constant.
4. Statistical adjustments.
1. Randomization spreads an effect of a variable evenly across the groups of the study. Often, organismic variables are so controlled.
2. Building in factors as independent variables
Building factors into the design as independent variables enables the researcher to determine the effects of those factors. Too many independent variables, however, can unnecessarily complicate the research design
3. Holding factors constant
When a factor is held constant, a potential variable is reduced to a constant. This eliminates, or at least substantially reduces, any effect the factor may have on the dependent variable.
4. Statistical Control
Statistical control, in essence, consists of adjusting the dependent variable scores to remove the effect of the control variable.
5. Using procedures for control in combination
The four procedures for enhancing control can be used singly or in combination.
Characteristics of good Research Design in quantitative Research
1. Freedom from bias
2. Freedom from confounding
Two or more variables are confounded if their effects cannot be separated.
3. Control of extraneous variables
4. Statistical Precision for Testing Hypotheses
PRINCIPLES OF RESEARCH DESIGN
After a researcher has identified a research problem and has completed at least some review of the literature, it is time to develop a research design - a plan or strategy for conducting the research. This chapter focuses on general principles of design, but because of process differences, qualitative and quantitative research are discussed separately.
The Purposes of Research Design
Kerlinger (1986, p.280) identified two basic purposes of research design are
(1) To provide answers to research questions and (2) to control variance.
The first purpose is general and straightforward – to provide answers to the specific research questions. Research should be valid, which includes being able to interpret results and, through those results, answer the research questions or problems being posed. Good research design assists in understanding and interpreting the results of the study and ensures that a researcher obtains usable results.
All research is conducted for the purpose of explaining variance – the fact that not all individuals are the same or have the same score or measurement. In a broad definitional sense, this may be true, but variance can be evident in a number of ways. For example, variance in elementary school students’ achievement, motivation, attitude, age, and family background can be considered. Also, when the variance of any one variable is considered, it may influenced by any number of factors. Variance in achievement, for example, may be due to aptitude and motivation, to mention two possible factors.
Variance is considered somewhat differently in qualitative and quantitative research. In qualitative research, the researcher attempts to explain the phenomenon understudy, including the variance, through description of a logical interpretation of what has been observed. The interpretation is based on a holistic concept of the situation. As we will see later in the chapter, in quantitative research more direct design procedures are taken for controlling variance than in qualitative research. Variance takes on quantitative meaning and, at least to some extent, variance is partitioned as attributable to variables under study.
Design in Qualitative Research
Qualitative researchers, for the most part, do research in natural settings; they do not manipulate or intervenes (except possibly by their presence) in the situation. Therefore, research design requires flexibility and a tolerance for adjustment as the research progresses. Smith and Glass (1987, p. 259) refer to this as a working design, similar to what McMillan and Schumacher (1989, p. 179) call and emergent design. From the identification of the research problem, decisions must be made about beginning the study.
1. Working Design
The Working design is the preliminary plan by which the research gets underway. Decisions are made about the subjects or sites to be studied, the length of time for data collection, and possible variables to be considered.
2. Working Hypotheses
Qualitative research uses the induction model, which conceptually means that data collection commences without any preconceived theories or hypotheses. However all researchers are influenced by their own backgrounds, and some information is likely to be available about the research problem. Earlier the concept of foreshadowed problems was introduced. Although technically these are not hypotheses statements, foreshadowed problems come in at this point.
3. Data collection
Methods of data collection include observation, interview, and the collection and review of related documents. In the dropout example, the researcher might engage in the following activities, although data collection certainly would not be limited to these.
1. Interview students and faculty, including guidance counselors.
2. Observe the interaction taking place between students and between students and faculty.
3. Review school records relative to factors such as grading patterns.
4. If in any way available, interview recent dropouts.
The data record of a qualitative research study can become quite massive with all the interview and observation protocols, document information, and so on. As recommended by some authors (Bogdan and Biklen, 1982), researchers should keep written accounts of their own thinking about what is being collected.
4. Data Analysis and Interpretation
Data analysis in qualitative research begins soon after data collection begins, because the researcher checks on working hypotheses, unanticipated results, and the like. In fact, data collection and data analysis usually run together; less data are collected and more analysis is produced as the research progresses. Qualitative data analysis requires organization of information and data reduction. The data may suggest categories for characterizing information. Comparisons can be made with initial theories or working hypotheses. Early data collection might suggest a hypothesis or theory, and then more data might be collected that would support, disconfirm, or extend the hypothesis or theory.
Data analysis in qualitative research is a process of categorization, description, and synthesis. Data reduction is necessary for the description and interpretation of the phenomenon under study.
Components of Research Design in Qualitative Research
Design in Quantitative Research : The concept of Controlling Variance
One of the purposes of research design in quantitative research is to control or explain variance. This is done by putting certain restrictions on the research conditions, such as including control variables.
Procedures for Controlling Variance
This section continues the chemistry example and illustrates procedures by which control of variance can be enhanced. There are basically four ways by which this can be done:
1. Randomization.
2. Building conditions or factors into the design as independent variables.
3. Holding conditions or factors constant.
4. Statistical adjustments.
1. Randomization spreads an effect of a variable evenly across the groups of the study. Often, organismic variables are so controlled.
2. Building in factors as independent variables
Building factors into the design as independent variables enables the researcher to determine the effects of those factors. Too many independent variables, however, can unnecessarily complicate the research design
3. Holding factors constant
When a factor is held constant, a potential variable is reduced to a constant. This eliminates, or at least substantially reduces, any effect the factor may have on the dependent variable.
4. Statistical Control
Statistical control, in essence, consists of adjusting the dependent variable scores to remove the effect of the control variable.
5. Using procedures for control in combination
The four procedures for enhancing control can be used singly or in combination.
Characteristics of good Research Design in quantitative Research
1. Freedom from bias
2. Freedom from confounding
Two or more variables are confounded if their effects cannot be separated.
3. Control of extraneous variables
4. Statistical Precision for Testing Hypotheses
Langganan:
Postingan (Atom)
