L5 - Sources of variation Flashcards
What is the difference between observed epidemiological quantities (incidence, prevalence, IRR etc…) and their ‘true’ or ‘underlying’ values?
The observed value for, e.g. incidence of meningitis in Leicestershire, varies from month to month. The true value is the the underlying tendency, which is not known and can only be estimated using confidence intervals surrounding our observed value, which have a specified probability (normally 95%) of containing the true value.
What is the effect of random variation?
Almost all observed quantities in medical science are subject to variation by chance. Random variation is the effect of chance on an observed value. e.g. let’s say the true value of meningitis incidence is 4 cases per month, due to random variation sometimes you may see 10 and sometimes 2.
How do observed values help us towards a knowledge of the ‘true’ values?
- Allow us to test hypotheses about the ‘true’ values
2. Allows us to calculate confidence intervals which include the ‘true’ value within a specified probability
What is a testing hypothesis?
A statement that an underlying tendency of scientific interest takes a particular quantitative value (i.e. the probability of tossing heads is 0.5)
What does formal hypothesis testing ensue?
Calculating the probability of getting an observation as, or more extreme than, the one observed - ASSUMING THAT THE STATED HYPOTHESIS IS TRUE.
If the probability (p-value) is very small, it is reasonable to conclude that
(i) the stated hypothesis is wrong OR
(ii) something very unlikely has occurred and hypothesis is true
What can you do to the hypothesis if p
The data is inconsistent with they hypothesis; therefore there is strong evidence (but no absolute proof) against the hypothesis and you can reject the hypothesis.
What can you do to the hypothesis if p>0.05
You CANNOT REJECT the hypothesis. The observed findings are consistent with the hypothesis and therefore it is quite possible that any difference observed between e.g. incidence rates occurred solely by chance.
Why is the p value cut-off for statistical significance usually 0.05?
An arbitrarily chosen value as it is deemed that a 1 in 20 chance of the data having occurred as chance is sufficiently unlikely.
Define statistical significance
The likelihood that a result or relationship is caused by something other than mere chance. Statistical hypothesis testing or estimation can be used is in order to determine this.
- When a p-value is less than 0.05 it is considered statistically significant
- When the null hypothesis value (e.g. IRR=1, SMR=100) is outside the 95% confidence intervals of an observed value.
Define clinical significance
clinical significance is the practical importance of a treatment effect - whether it has a real genuine, palpable, noticeable effect on daily life.
What is the null hypothesis?
A hypothesis assuming that two things are equal or that there is no effect or difference.
What is meant by the ‘point estimate’?
Our best guess at a true value
What is the meaning of the 95% confidence interval?
The range within which we can be 95% certain that the ‘true’ value of the underlying tendency really lies.
What is the 95% confidence interval range centred on?
The observed value, as it is always our best guess at the ‘true’ underlying value.
IT IS ALWAYS WITHIN THE RANGE!!!!
What does it mean if the null value is within the confidence intervals?
Values inside the confidence intervals are consistent with the observed data and therefore if the null hypothesis is within the confidence intervals, the null hypothesis is consistent with the observed data - and any observed difference from the null hypothesis may be due to chance.
