Oct 28, 2012
Method of Educational Research
Documented by Marsigit
Quality Improvement and lessen time of study. Problems are comming from the students. The different among the supervisors about the qualitative, quantitative and mix approach as well descriptive approach. Most of the graduate students carry out Research and Development. There are also Research by Project, Classroom Action Research. Four and five years ago most of us implement Factor Analysis. Research and Development uses Structure Model. How to have the same perception of the research in Post Graduate program. The different of paradigm, abstract: 4 or 5 paragraph. Topic and problem.
Topic has reflective the relationship among variable. Identification of problem and limitation of problems has not relevant. Problems and aim of the research has not synchronized yet. Conclusion and formulation of problem was not relevant. Problems of research or hypothesis. The conclusion was still very long (4 pages). Recommendation did not base on the related research. Citation problems: name, the length, the space line, etc. In S2 most of them tend to Research and Development.
How to improve Research and Development.R&D aim to produce on education to solve practical problem in teaching learning. Developmental Research is different. R & D is not to examine theory of hypothesis but to examine and improve the products. Are the product relevant to solve the problems e.g motivation. The facilities, time, supporting financial.
The factual problem there is a gap between fact and ideas. E.g there is no innovation because of there is no references (Book) so the R & D need to develop books. R&D there preliminary study: method, model to tool, theory and ; the problems: innovative, motivation, Focus: Design T&L and implementation. Specification of product.
The result of product: computer software, module, evaluation instrument, etc. Product specification: the scope, ...Criteria of choosing the product: the need t & l, rational plan. Aim of R & D: improve quality of TL, efektiveness, Three Component R & D: Development Model, Procedure of Development, Try out Product: design, subject,..Model of Development: .., Prosedural Model,...Structure Model, memaparkan the component, Goodnes and Feet (with the Statistics). 10 Step for R & D (Bob and Gol. R&D + CAR = Research by Project. Descriptive is not similar to qualitative research.
For S2 is okey research of Descriptive and Qalitative. Need to be improve on its description and qualitative method. Methodology more enxtends than method. Problems: popular issues-with theory-philosophy. S3 research should be published. The problems should be interesting. How to understand methodology. Research problem should be elaborated from research problem.
Qualitative: looking for meaning; looking for pattern. "Si UPIK mencari Cinta" looking for.. . R&D about soft skill, how to evaluate soft skill. Gender isues: lead to guide line for Local Government Official. Hirarchical linear model? BOOK: Organizing and Managing your Research (2007), 2. Doing Research (2005). 3. Research Design : Qualitative and Quantitative. 4. Action Research:.. Gelingher.Research Formulation: Gelingher: Plan (Liniaer + Cyclic), Controlable (by design , by Statistics), Empiric, Critical, Irrational Finding (Conceptual + Calculation error), Guided by Theory. Quantitative: theory is the ground.
Qualitative: theory to help the understand the meaning. The students sometime were still confused. In the theory there is a pre-assumption. Theory is a collection ..phenomena-menjelaskan -fenomena itu. Theory central peripheral (gaung). About variable. Variable is not the concept. Variable is the operationalization of the concepts. Dynamic variable should be more. Static variable is only limited. Static variable: 1, Dynamic Variable: 4 (e.g). Dependent/independent; intervining variable. Intervining variable can not be observe? Variable as moderator.
Variable concomiltat/iringan --> covariance. Intervining just lay in our minds. Development using corellation model. Causal : the relationship of reciprocallity, model sank. Research paradigm (sometimes is just be defined as the relation of variabel. Paradigma penelitian kualitatif dan paradigma penelitian quantitative. Not only just a relation. Model: is a structure to make easier to understand. Factor learning : regression. Measurement model in a symbol. Path Analyses: use diagram. How to develop model.
Design of Research: Plan & Structure. Each research has its design. Ethnography. The students should have a clear picture about their design. Survey Design, Experiment Design.
Quantitative-variant-statistics-(parametric). Sample can not be chosen purpossivelly.
Quantitative: clear with its aim: explanation, prediction, estimation, examine the model.
Quantitative: feasibility of getting data, accuracy data, supporting references, Match between the theory and method.
Procedure and the sample should be correct. Just leave the following: Rule of thumb. Sepersepuluhan : yahudi. Assumption: number of population is known or not. For the population is not known: effect size, .. if not ...Morgan. Harry Cane: only for population of 2000. Significant? We can determine the number of sample. Correctness of procedure. Clarity of finding. Title: explicit or implicit variable. Hypothesis based on conceptual framework. Map of concept. Research Area. Accurateness of data. Fundamental Research for S3: developing science, ...
Analyses quantitative data. Definition of statistics-analyses data-just to help. Empirical data. Measurement-the instrument should be good-questionnaire-valid or not. Test-teachers' certification-the impact/effect. Descriptive analyses-some province-pass certification -higher or lower achievement after certification. Data-information-normal-mean-median. Median-parametric. Mean- Standard Deviation. Stem and leaves. Infferential: generalization.
Stem and leaves: the information does not lost. Normal, t, Q square. Sampling Distribution. Statistical Estimation of parameter. Sampling Distribution: Distribution Normal, Dist t, Dist F, Q-Square. Problem: Variance should hommogen. Dist t: correlation, test beda. Fizher. The length life of a lamp. Variant Philip-varian Osram-Fisher. Modelling: Q square-goodness of fit. The supporting data itu bukan satu-satunya. Concept= definition-from theory-theoretical-operational-indicators-item-check with theory.
S2: Exploratory Factor Analysis (EFA); S3: Conformity Factor Analyses. Significant-degree of kebebasan- n(number of data). Goodness of fit-ajustment. Blue print of the instrument. "Mampu memilih dan mengunakan" (two concepts). reliabilitas=indeks keandalan- belah dua-sperman brown-asumsi tak bisa dipenuhi turun lagi Crown-Cronbach-Congeneric. Reliabilitas =1 kesalahan = nol(impossible). Try out. Theory response item: function information of the instrument. Test Beda: two grou-t Distribution; more than two group: analyses variance. Analyses Variance. Factorial: by design, Analyses Covarian: by statistics. Tetrachoric: continue-discontinue-nominal.Analises Varian Factorial: interaction. Assumption: Analises Variance: random sample-..,Analises Covarians: A1, A2, A3,... Assumption Covarian Analyses. Linear and Non-linear Prediction. Causal Analyses: Path Analyses- High Reliability > 0.8, SEM: Structural Equation Modeling.
Quasi-Experiment: non random-hypothesis? Experiment ResearcH, three control: phisical control, statistical control, procedural control? Quasi: t-test? Clarity of concepts.
Instrument free from kesalahan; R&D: Problems-relevant of statistics; Issues: qualitative-quantitative-mix; please indepth study the philosophy: naturalistik combines with positivist; Standard Book: Handbook of qualitativ-naturalistic; standard research - standard research book is important.\
Experiment-laboratory; Quasi: field laboratory-not ideal condition. Control and experiment group should be comparable. Meet the assumption or not? Variant is the same or not.
Full variant model; separated variant model; n the sama, variant not the same or vise versa. Understanding Edicational Statistic Book. Phisical control: lamp, laboratory; society sometimes impposible. Statistic: other cases contaminate need to be control.
Nurturant variable: Anova; anacova. Prosedural: SOB (Standard Operation Baku). The influence of breakfast on students competence, in fact, they should eat rice. Three references:..,..,..
R&D: 10 step, can use experiment in try-out the model. In R&D does not always use SEM. ANOVA that model is effective. Varian analyses? Assumption. SEM should has a power assumption. Except that the product is the concept itself. Qualitative: ethnography-antropology from USA. Quantitative "Kemiskinan" asupan calorie. Supplement-complement. Merger them not on paradigm but on their procedure.
The newest resources is good. Mr. Bruri: Hypothesis, experimen, random, intake design, no hypothesis in quasi. Answer: Control and experiment group should be comparable. Variance is the same? In each experiment there should be hypothesis. Time series experiment- improvement- Contextual Analyses. Effect on individual and group. Treatment affect to individual or to group?
If there is no generalization there should be no hypotheses. The parameter is their themselves. Rrsearch question- conceptual framework; Problem from the field. Research question can not contradict with problem.
Random for individual or group? Hyphoteses? Different test? T-test. Answer: unit analyses is the class and not individual. Analyses is Contextual Analyses. Affect to individual, group and interaction. Random on group, t-test.
There are references to develop guideline. How to random? Random error, systematic error (person). Handbook design and analyses. Unit analyses: school or class, we examine the variance. Individual make the power of statistics higher. The variation of analyses can be very great. Survey=margin error; Experiment=power of statistics. Homogenity can be lower or higher.
Face Validity; Rooster- Not contextual, Eestimation, quantitative? estimation? Instrument-content-construct. Content: Penguasaan isi, Validation: expert judgment-practitioner-content reflect competencies. Test validity-Vandalen. Validity -internal consistency. If it based on concept: psycology, motivation- model-tree-and branch-construct-sub-construct-indicator-descriptor-item: analyses factor.
Orthogonal-oblique: factor. Misunderstood: correlation between item and total? this is just validitas item and not validity of instrument. Now uses factor analyses. The sensitivity of instrument: measuring-index; more than 27 % can be used. CAR should has the concret problems in a certain school. There should be cooperation with the teachers. Alternatives actions.
CAR by practitioner to improve the quality. So that we need the simmilar perception. Collaboration is important. To evaluate their self program can not be carried out. First procedure: 1. the students and supervisor, 2. to be reviewed, 3. supervised by the director.
Results-process. Evaluation research cannot be conducted at his school. Analyses contextual should be done. Prof. Djemari: item analyses-daya beda. Norm-Criterion. KTSP: the competence of students and teachers. ANOVA and ANACOVA parametric? For Internal validity there should be control.