All the time when we undertake research or audit we collect data. The data may be simple such as age of patients or gender of patients. The data we collect may be complex: the Glasgow Coma score at admission. All types of data can be divided into two main groups.
- Main data types:
Each of these groups could be further divided. The categorical group could be divided into nominal or ordinal. The numeric group can be divided into interval or ratio.
Types of numerical data:
Types of categorical data:
Numeric (ratio and interval data) could be tested to assess whether they have normal distribution. However, normal distribution does not apply to catagorical data (nominal, ordinal).
°°Numerical data are sequence of numbers. The difference between one number and a number above is same as the difference between the number and the number below it.
However, as mentioned above there are two types of numerical data. They are interval and ratio. Temperature in centigrade is an interval data. Going from 30°C is to 31°C is the same increment in temperature as going from 31°C to 32°C . However, going from 10°C to 20°C does not make the temperature twice as hot (as the centigrade scale starts from -273.15°C).
However, height is a ratio data. This is because the measurement of height not only has equal interval property i.e if going from 3 feet to 4 feet is same increase in height as going from 4 feet to 5 feet, going from 2 feet to 4 feet is twice as heigh – this exemplifies ratio data.
As mentioned earlier, categorical data has two types. The classical categorical data is nominal e.g. gender (male vs females) or ordinal.
Glasgow coma score is ordinal. Although, while descending the score severity of the of head injury worsens, the interval between sequential scores are not the same i.e. Going from GCS of 7 to 6 is not the same worsening of severity as going from a GCS of 6 to 5.