Statistical tests - Level of data

Statistical tests - Level of data

1. There are 3 decisions to be made when choosing a statistical test

   • Are you looking for a difference or correlation? Whether it is an independent or repeated measures design? What level data are you getting?

2. Nominal data - categorical data

   • 500 P’s. Males and females. Tell us about categories. Called nominal data. Most basic type. Alternatively, we could say that men = group 1 and women = group 2. Just categories.

3. Ordinal data - put an order to it

   • Also called ranked data. Rank data from highest to the lowest. We can rank the data from lowest to highest (1st, 2nd etc, level of satisfaction (1-7)) Ranks might be uneven. Data is seen as more sophisticated than nominal data because we get more information from it but we don’t know the distances between the scores.

4. Interval Data - has fixed intervals

   • So interval data gives us more information. For example temperature. We can clearly see which is hotter of 2 temperatures and there is a meaningful difference between the two. There is the same amount throughout the range. This is, therefore, more sophisticated than ordinal data. There is no fixed 0 in interval data.

5. Ratio Data

   • The final level of data is very much like interval data except there IS an absolute 0 This level is called ratio data. It is so similar to interval data (and sometimes hard to distinguish between the two) that both are often combined as interval/ratio data. You can have 0 times so timings are ratio data.

6. Levels of data

   • We get different types. Some data is more sophisticated - tell us more about the population we are studying.

Images and files in the map