“The statement that scientific proof is based on falsification. The method is based on the premise that science cannot prove with absolute certainty anything to be absolutely true.”
“The statement that scientific proof is based on falsification. The method is based on the premise that science cannot prove with absolute certainty anything to be absolutely true.”
– Inferential statistics: dispassionate and objective interpretation of the data
– Probability: likelihood degree of an event
– Confidence interval– range of value, true value with high probability
Two way: a. sample size and confidence limit variant with interval. b. Mean and SD.
– Hypothesis Testing:
Why the null hypothesis:?
Problem in interpretation?
Hypothesis is about a population, whereas the study is done on sample, no study can prove weather hypothesis is true, classical way consists of How low is the probability that the hypothesis is false, (5% conventionally)
Eg: The tests examine the null hypothesis, and gives the probability (p value) of the null hypothesis being true. If p value is less than 0.05, it means that there is less than 5% probability of null hypothesis being true. In other words,, there is less than 5% probability that the treatment isnot effective. Stated another way, the probability of the treatment being effective is more than 95%.
Hypothesis alternate– not possible by absolute value– error determination (lowering)– So go for null hypothesis– p value- comparison, true/false– finally conclude for alternate (our aim hypothesis)
Null hypothesis u1=u2 or u1-u2=0. Null says same, so that goes signifies research. But to test we need go form falsification i.e difference values are not same (i.e. values are same).
Eg: The treatment isnot effective” is a null hypothesis. This looks little confusing. The problem is : looking by two dimension. Simply to one dimension, The our Treatment modality is better is alternate and for null hypothesis is our treatment modality is not effective (ie same). So, for better understanding go by one dimension point.
Alternate hypothesis: CL,
Null hypothesis: p value,
– Type I and Type II error: The avoid of one type of error increases the probability of the other types of errors. (So, 95% CL). The type II error can be avoided by not fixing alpha at two low at level:
– The whole propaganda is due to sample size: If sample size is large, the propaganda is less so, The one factor which can reduce both types of error is large sample size and result obtained on the sample truely reflect the results that would obtained of the whole population were to be studied.
– Power: attribute of the study. Correcting validating the hypothesis, if the hypothesis is true. Why equals to 1- beta not 1- alpha. Why power was needed: due to sample size,