hypothesis & variable

36
HYPOTHESIS & VARIABLE wien – fmub 1

Upload: eden-faulkner

Post on 31-Dec-2015

34 views

Category:

Documents


0 download

DESCRIPTION

HYPOTHESIS & VARIABLE. wien – f m ub. HOW TO GET SCIENCE?. Logi c. Dedu ctive. Library. Rationally world. Predi ction. Science. H ypothe sis. Empirical world. Indu ctive. Hypothesis Testing. Statisti cs analysis. Methods. Fa cts. PROBLEM SOLVING CYCLE IN MEDICAL PROCESS. - PowerPoint PPT Presentation

TRANSCRIPT

HYPOTHESIS & VARIABLE

HYPOTHESIS & VARIABLEwien fmub 1LogicDeductive Rationally worldEmpirical worldPredictionHypothesisHypothesis TestingFactsInductiveScienceLibraryStatisticsanalysisMethodsHOW TO GET SCIENCE?22PROBLEM SOLVING CYCLE IN MEDICAL PROCESSPatients complaintGeneral examinationTemporary diagnosisPemeriksaanlanjutanConclusionTherapy3 HYPOTHESISHypo = under; thesis = truth

Temporary conclusion towards problems which is needed to proveIn scientific methods:after describing the Problem

define Hypothesis

HYPOTHESIS4Is hypothesis always needed ?Survey, explorative and case study (descriptive)usually hypothesis is not needed Mostly, the purpose of medical research is to know the relation between variables

hypothesis is important5Theory : the data which is composed systematically as a conclusion that apply generally is needed as a general guide

Concept: a thought or an abstraction of a certain comprehension towards phenomenon which is formulated generally not measurable

broken down into variable measurable Hypothesis: - is derived from theory - is needed to explain towards problem6HYPOTHESISA prediction or reference which is described and accepted for a moment, that can explain the facts observed, and can be used as a guide to do research (Good and Scates, 1954)Some temporary explanation about the facts which can be observed (Trelease, 1960)The statement that it predicts the relation between two or more variables (Kerlinger, 1973)7Hypothesis function1. To verify the truth of theory2. Gives idea to extend theory3. Widen knowledge about the phenomenon which is observed8Hypothesis benefitsGives the limitation towards scope and research workingReminds about condition of the fact and the relation between the facts which is observedAs a simple tool to focus the facts into unityAs a guidance to do verification process9Good HypothesisIs described in declarative sentence that is clear and the simple oneHas strong theoritical base or conceptDeclare the relation between one independent variable and one or more dependent variablesPossible to be tested empiricallyShould be specific and show the variables that will be measuredIs described before start the research10Smart observation of the researcherImagination and creative thinking of the researcherAnalysis framework that is usedMethods and design that is choosenGood hypothesis are affected by: 11HYPOTHESIS FORMULATIONDo your questions need to know the cause of the problems?No, just to know the problemsYes, I want to solve it need to formulate hypothesis

12Problem - Hypothesis A tentative statement about the relation of the probably cause with the problemProblem : Why the heartbeat of the patient 100 per minute?Hypothesis : Hyperthyroidism is the probable etiology of tachicardia

a temporary answer to the problem / question13Hypothesis (theoretical meaning)1. Incomplete / hasnt established statement (hypo); in term of the truth so, must be tested2. The most probably answer / explanation of research problem3. Theoretical conclusion derived from literature study.4. Its an intellectual prediction is needed for experimental & observational analytic research1414Hypothesis (technical meaning)Statements about parameters / variables will be tested its validity empirically using collected sample

may be:correlational studycomparative study15Hypothesis should be developed through strong theoritical base, experience, other strong scientific resources

One hypothesis is only allowed for one independent variable

For complicated hypothesis must be divided16Example : Education level of mother has contribution role towards growth and development of child simple but that it can not be tested by only one hypotehesis

Divided into 2 sub-hypothesis:1. The education level of mother associated with the child growth2. The education level of mother associated with child development

1717 The child who is never comunicated with adult people will be experienced with severe mental retardation can not be tested empiricallyBad hypothesisFormula is too general and more than one meaning must beavoided Giving drug X can cure gastrointestinal function disorder

which one of gastrointestenal function (digestion, absorbtion, atau excretion?) 1818Hypothesis statement which is to risk: Studying is not affect the students performance from the data collected there was an evidence that students who did not study can be passed the hypothesis was proved the conclusion by students: studying is not useful it will have destruction effect (this is caused by an error in describe hypothesis)

19Based on hypothesis aspect, research can be divided into:

1. Hypothesis testing research 2. Hypothesis generating research

20Using medical record data generates hypothesis For hypothesis testing, should be used the other samples set e.g. from the medical record, there are 3000 patients with congenital heart disease.

when a researcher to find out whether an association among the variables generates hypothesis test the hypothesis using the same data (DATA DREDGING) is outlawHYPOTHESIS GENERATING RESEARCH2121HYPOTHESIS TYPES1. RESEARCH HYPOTHESIS - is aimed to make a prediction about the occuring phenomenon e.g. When the environment sanitazion is not good, thus transmitted disease at the area is high There is an influence of smokes towards incidence of lung cancer

22222. STATISTICAL HYPOTHESIS - H null (Ho) & H alternative (Ha)

Ho means there is no difference between ... and....e.g. There is no difference between mortality rate of heart disease in urban and rural There is no relationship between smokes and lung cancer

Ha means there is a difference / relationship usually, it is the research hypothesis

2323HYPOTHESIS TESTINGIs the answer to the questions true ?If you want to verify, you need A SCIENTIFIC HYPOTHESIS TESTING:1. FORMULATE the hypothesis2. IDENTIFY AND FORMULATE the variables included in the hypothesis3. Choose the best and efficient way (DESIGN) to collect the data based on your hypothesis4. Choose the RIGHT METHODS TO ANALYSE AND INTERPRET5. Make the right CONCLUSION: is the hypothesis true or false24HYPOTHESIS TESTINGIt is statistical pocedure to show the validity of the hypothesisIt is needed due to the research is done at the samples (not at population), while researcher want to generalize the conclusion towards populationHypothesis testing is done by Ho statement Is it evidence (by facts) which is enough to refuse Ho, and that so H1 will be achieved ? is interpretated through p (probability) valueThere is the probability to do mistake of hypothesis testing

2525The types of error in hypothesis testingConclusion andDecisionThe factsHypothesis is correctHypothesis is incorrectAccept Hypothesisno errorError type II ()Reject HypothesisError type I ( )no errorpower of study

26Error type I ( = significance level): - the chance to reject Ho in samples, but Ho is correct in population false positive- determined previously by the researcher- value: 0,5 ; 0,1; 0,05; 0,01

Error type II ( ): - the chance to accept Ho, but in fact the Ho should be rejected and value can be reduced by increasing samples 2727VARIABLESIs the characteristic of subject (not that subject) variouslye.g. Type of sex, Body weight, Blood pressureThe types of variable is important due to related with data collecting and analysis technique

Scale of variables 1. Categorical: nominal, and ordinal scale 2. Numerical : interval, and ratio scale282829Type of variables (based on its function)1. Independent & Dependent Variable Independent variable (risk, predictor, causa) - its change will alter other variable Dependent Variabel (effect, outcome, event) - variable which will be changed caused by changes of independent variable

Examples :Treatment by drug A decreasing blood pressure i v d vCholesterol level is different between male student and female student30Satu jenis variabel dapat memiliki fungsi yang berbeda berdasarkan konten risetKeturunanKonsumsi garamMerokokKegemukanHipertensiKematian pada Manula31Is the variable which has association with dependent var. and also with independent variabel but its not an intermediate variableIdentify toward confounding factor is important since it can cause wrong result. e.g. it can be concluded there is association between the variables but actually threre is no association 2. Confounding Variable (Variabel Perancu)32eRelationship between variablesIndepend var.= Independen= Pengaruh= Perlakuan= Penyebab= Regressor= Treatment= Prediktor= Risiko= KausaDepend Variabel= Dependen= Terpengaruh= Terikat= Regressed= Efek= Outcome= Event= ResponsIntermediate varConfounding VariabelOutside var. Outside var. Outside var. Outside var. Outside var. ABCDE33Independ varInterm var iableDependent varconfoundingIndepend varDependent varMIndepend varDependent varVNot confoundingABCRelationship among independent, dependent, confounding variabel34Variabel PerancuMinum kopiPJKMerokokTerdapat asosiasi semu antara kebiasaan minum kopi dengan kejadian PJK

35Variabel perancuMakan permenCarries gigiSikat gigiHubungan antara makan permen dan karies dentis dapat tersembunyi pada anak yang gemar makan permen tetapi rajin gosok gigi.

Thank you36