Analysing medical research data
In a nutshell
Analysing and evaluating data is very important in medical research. There are a few important features you need to identify and questions you need to answer when evaluating data.
Looking for patterns
Scientists try to identify useful links between variables, sometimes there is no link. When you interpret graphs, you should look for key features.
1.  Identify patterns or trends. 
2.  Use axis labels AND units when describing a graph. 
3.  Identify maximum and minimum values; the range; outliers; and patterns that do not fit the trends. 
4.  Quote numbers to provide evidence of your point 
5.  Use key words such as 'increased', 'decreased', 'constant', 'plateau', 'maximum' and 'minimum'. 
Correlation
Definition
A correlation is an association between two sets of random data. If one variable increases as the other increases, this is a positive correlation. If one variable increases as the other decreases, this is a negative correlation. This correlation does not necessarily mean causation and further investigation needs to be carried out to prove a causal mechanism.
Examples
Correlation  Graph  Description 
Positive
  As A increases, B increases. Therefore, A and B are positively correlated. This does not prove that A affects B, or B affects A. A and B could be caused by another variable.

Negative   As A increases B decreases. Therefore, A and B are negatively correlated.

None   There is no correlation between A and B.

In medical research, scientists need to determine the links between treatments and cures, or risk factors and diseases. This can be difficult if the proposed cause of a disease only sometimes results in the disease. For instance, if one disease has many possible cures or there is a long delay between the proposed cause and effect.
Evaluating data
When evaluating data, there are some questions you should always consider.
1.  Are the data reliable? Was there an appropriate control group? 
2.  Are the data valid? 
3.  Was the sample size sufficient? 
4.  Could the data be better collected? How long was the study and how was the sample chosen? 
5.  How could more accurate data be collected? Do the data answer the questions? 
6.  Are there any anomalies in our data? Can these be explained? 
7.  How confident are you that the evidence supports the conclusions made? Has the causal relationship been proved? 
8.  Could anyone use the same data to support a different conclusion? 
9.  Are the data biased? Who completed the research? Do they have a personal interest in the data that could influence the results? 