It is not uncommon for medical students to claim they chose a career in medicine because of their allergy for mathematics. While it is historically true that in the initial years statistics itself was considered a sub-specialty of Mathematics, it has outgrown its parent discipline and has established its own academic space. Today it will not be uncommon for departments of statistics to housed in buildings separate from those of mathematics. It is important that in promoting statistical thinking that we stress the use the minimum number of formulas and manual computation. Such a therapeutic attitude towards mathematics however requires maximum clarity of minimum ideas and conventions of mathematics.
Most clinicians have had sufficient exposure to statistical ideas during their school days. Thus students in the initial years of medical and nursing have greater aptitude for statistics than in their later years . It could also be advantageous to introduce statistical thinking in the initial years of clinical education.
One of the modus operandi in this blog is the use of small data sets so that the reader can visualize the impact of the manipulations.
In understanding the cognitive developent of Statitical Thinking among clinicians we can learn much from frameworks in understanding the evolution of statistical reasoning in children. As much of this blog promotes a data based approach to statistics we will follow the frame work of jones et al :
a. Reading Data .
b. Reading between data.
c. Reading beyond Data.
In this post we will address some preliminary etiquette to get by :sy
1. Scientific Notation.
2. Statistical terms and their Greek symbols.
3. Conventions in grouping data.
4. Arithmetic manipulation of Percentages.
5. Standardization Maneuvers.
In building frequency distributions there is some confusion about class intervals . Class interval such as 60-62 is a symbol for that particular class. 60 is called lower class limit and 62 upper class limit . How ever what happens if we have value of 62.5. thus in practice we use something called class boundaries or true class limits ie 60.5 -62.5. The mean or adjoining limits . 62+63/2=62.5. the real issue is class boundaries should not coincide with actual observatons .Thus a observation was 62.5 then it would not be possible to decide whether it belongs to the class 60.5-62.5 or 62.5-64.5. the class midpoint or class mark is computed by adding upper and lower limit and dividing by 2. for mathematical purposes all members of a class are assumed to have the value of the midpoint. thus all members of the class 60-62 have value of 61.this often leads to grouping error.as there is loss of information due to grouping. this is often minimised by choosing classintervals such that observations coincide with the midpoint.
observations should not coincide with class boundaries but should coincide with class midpoints.