Levels of Measurement

different levels of measurement mean…

different levels of data that vary in precision

what data can be divided into different levels of measurement?

quantitative data

3 main levels of measurement

• nominal

• ordinal

• interval

nominal data

• separated in categories

• also referred to as categorical data

• relies on a frequency count where you count how many items fall into each category

• discrete data - one item can only appear in one of the categories (e.g. a person’s vote can only appear in one category)

what level of measurement is nominal data -

the most basic level because it provides very little detailed information/insight as it is a head-count

only tells researcher how many people are in a group/how many times a specific behaviour occurred

ordinal data

• data which is ordered in some way

• e.g. a rating

what usually produces ordinal data?

questions and surveys as they use arbitrary scales and measure psychological construct in a more subjective way which lacks precision

why does ordinal date provide NO CERTAINTY as to the intervals between each value

• the intervals are not of fixed, equal, definable size

• the rating could be subjective e.g. on a scale someone’s rating of 10 may be another person’s version of an 8

interval data

• fixed equal intervals between each unit of measurement

• e.g. centimetres, timings

• they are publicly accepted units of measurement

how precise is interval data?

interval data is the most precise and sophisticated level and is necessary criterion for the use of more powerful statistical tests

how can we produce interval data from rating scales and questionnaires?

though they tend to produce ordinal, when we use well-established standardised psychological tests/questionnaires, that data is considered interval

converting data hierarchy

You can convert data from higher to lower levels of measurement

Interval → Ordinal → Nominal

what happens if you convert data from higher to lower?

you lose precision in your data and statistical power

conversion - what to remember before you convert interval to ordinal

• to convert interval to ordinal data, you must rank the scores from highest to lowest

what to remember when you convert ordinal to nominal

• you must sort the ranks into distinct named categories