4_6033130600662045188.pdf
Transcript
What is hypothesis? Research problems Steps of hypothesis testing Hypothesis A hypothesis is a statement or supposition of relationship between two or more variables that permits empirical testing which is liable for acceptance or rejection. Research Problems...
What is hypothesis? Research problems Steps of hypothesis testing Hypothesis A hypothesis is a statement or supposition of relationship between two or more variables that permits empirical testing which is liable for acceptance or rejection. Research Problems Example: Patients with diabetes have a higher blood pressure than non-diabetic. Oral contraceptive pills can cause cancer breast The idea about research problem should be formulated in the form of a research question as follow Can diabetes cause increase in blood pressure among diabetics? Can oral contraceptive pills cause breast cancer? It should have the following criteria: - It translates the research question into a prediction of expected outcome. - It has 3 components: Population, variables and relationship between the variables. - It is impossible to prove any hypotheses, but it is easy to disprove it. Hypothesis testing steps State the hypothesis Check the assumption Define required statistics Define level of significance Detect the significance test Draw & state the conclusion We have 2 statistical hypotheses Null hypothesis “refutable hypothesis”(H0): hypothesis of no difference between groups, and both groups are equal. Alternative hypothesis (HA): hypothesis of there is difference between groups, and both groups are not equal. - Type of variables - If the data is normally or not normally distributed - Equality of variance - No. of groups & No. of observations in each group - Whether the groups are matched or not required statistics Quantitative Qualitative variables variables Mean SD NO. Percentage Confidence interval Confidence level Level of significance (P- value) Confidence interval: It is the interval which with a certain degree of confidence (confidence level) it contain the parameter being estimated. Confidence level: level of assurance that if you repeated the survey you would get the same results. Level of significance (P- value): It is the probability that the observed difference between groups is due to chance and errors. So it is the level at which we judge the null hypothesis. The investigator is the one who choose the level of significance. In medicine we usually use 0.05. It means that the researcher allow himself 5% to commit alpha error. Interpretations of p-value: P – value ≤ 0.05 , we reject null hypothesis and we conclude that the difference between groups is significant. P – value > 0.05 , we accept null hypothesis and we conclude that the difference between groups is non-significant. Type 1 error (alpha error-false positive): It is the error which Types of arise from rejecting null hypothesis when it is actually true. error p = 0.05 means you allow yourself 5% chance of committing type I error Type 2 error (Beta error - false negative): It is the error which arise from retaining null hypothesis when it is actually false. الحــقـيـقــة Null hypothesis True False (No difference) (No difference) (difference) Rejected Type 1 error Correct decision (difference) (False positive) البحث بتاعك Not rejected Correct decision Type 2 error (No difference) (False negative) Quantitative (normally Qualitative distributed data) Difference Association Difference between groups Association (Correlation & between 2 or more (odd`s ratio) regression) groups Matched 2 Not matched More than 2 2 groups (macnemar (chi square groups test) test) Unmatched: Matched: Unmatched: Matched: Anova test Multilevel independent T paired T test (F test) Anova test P – value ≤ 0.05 , we reject null hypothesis and we conclude that the difference between groups is significant. P – value > 0.05 , we accept null hypothesis and we conclude that the difference between groups is non-significant. State the hypothesis and formulate the research question for the following: 1. The relationship between caffeine consumption and sleep quality 2. The association between daily social media use and anxiety levels in adolescents
