Week 6

Question 1

What is correlation?

Answer 1

technique for investigating the relationship between two numerical variables.

CORRELATION does NOT imply CAUSALITY
use scatterplot to display correlation

Question 2

Types of correlation

Answer 2

Positive correlation:
eg. as age increases so does wisdom
Negative correlation
eg. as age increases, motor ability increases

Question 3

What to use to calculate the correlation coefficient

Answer 3

these will determine the strength of the correlation between variables

PEARSON correlation coefficient (r)
use when data is NORMALLY DISTRIBUTES and sample size is ABOVE 30

SPEARMAN correlation coefficient (rs)
use when data is SKEWED and sample is BELOW 30

Differences::
- Pearson ranks data first

INTEPRETATION
r = 0 - no correlation
r is close to 1 - positive correlation
r is below 0 - negative correlation

Question 4

Hierarchy of Evidence

Answer 4

It is a ranking system of various study designs. It only applies to quantitative data. There is no one gold standard, depends on the question as to what design should be used.

Why we need it?
We need it to rank study designs based on their bias, chance and cofounders as we need it to be the best study design (have believability, external and internal validity) to answer a clinical question.

Deficiencies?

• hierarchy of evidence does not recognise expert opinion
• if there is absence of published evidence, there is no evidence.