Type of Data
Imagine you have now collected your data from the collection method in step 3. Now you have to determine what type of data you have collected. Before you can perform any data analysis, ie interpret your data and present your data, you need to identify what type of data you have. There are 4 types of data: RATIO, INTERVAL, ORDINAL OR NOMINAL data.
A good example of ratio data is collecting height in metres, weight in Kg, quantity in litres. Ratio data is simply true numbers. It is data that deals with numbers that have a zero, go up in order and can be multiplied and divided. For example, weight measurements (Kg) increase and decrease in order, 20kg is twice as heavy as 10kg, and there is such a thing as 0kg. Another example is number of goals scored, 10 goals is twice as many as 5 goals, and you can score zero goals. A further example is minutes to complete a run, 30 minutes is twice as fast as 60 minutes. Ratio data is the easiest to use because you know the exact value between each unit. As a result, if you have or plan to collect weight, height, times, number of shots, goals etc you are most likely dealing with ratio data.
Good examples of interval data are collecting shoe size or recoding temperature in Celsius. Interval data is dealing with number, the numbers go up in scale, ie 5 is higher than 1. Interval data also increases and decreases with the same intervals between each unit, for example the difference between 1 and 2 is the same as the difference between 2 and 3. However, there is no true zero. Therefore, you cannot multiple and divide the numbers. For example, there is no such thing as 0 shoe size (that would mean no feet!), size 12 feet are not twice the size as size 6 feet, however size 12 is bigger than 6 and the interval difference between each shoe size is the same. The same is true with temperature, 20 degrees is not half the temperature of 40 degree, but it is a colder temperature and the intervals between each degree are the same.
A good example of ordinal data is collecting survey results where the participants have been asked whether they were very satisfied, satisfied, ok, dissatisfied, very dissatisfied. Or how much do you agree, select from 1-5, 1 being completely disagree, 5 being completely agree.
Ordinal data can deal with numbers and words, there is normally a scale, for example you know that ‘very satisfied’ is a higher score than ‘satisfied’. However, you are not sure what the interval is between the scale, for example is ‘very satisfied’ twice as good as ‘satisfied’? A little bit better? A lot better? The same is true with how much do you agree, you know that a 4 score is better than a 2 score but you are not sure whether is twice as good, marginally better, a lot better.
Good examples of nominal data are collecting data like gender (male or female), preferred sport (rugby, football, tennis etc), what team do you support? (Chelsea, Arsenal, Fulham etc), what colour hair do you have? (brown, blonde, grey etc).
Nominal data is fairly easy to determine because it is data that does not have a quantity value and there is no scale to go by. For example, there is not a number value between answering what type of football boot you wear, adidas or nike have no numerical difference.