PLS 210-A01

    Levels of Measurement

  • ratio
  • interval: specific number, the number 0 is crucial
  • (above 2 can be combined to “Interval or better”

  • ordinal: order or class
  • nominal: attributes not quantities
    Forms (formats) of data

  • Individual (raw score) data
  • grouped frequency distribution
  • ungrouped frequency distribution (only appllies to integer/ratio level data)

Nominal dichotomy (2 categories only):
GENDER f=
Male 11
Female 7
n= 18

Ordinal grouping of data: must be two categories, everything must be in one OR the other

n-category nominal
Religion (nominal attribute, not quantitative data)

Catholic
Islam
Presbyterian
Baptist
Methodist
Other Protestant
Eastern Orthodox
Hinduism
Taoism
Judaism
Buddhism
Agnostic
Atheist

ordinal dichotomy
Income f=
High
Low
n=

n-category ordinal
Income f=
High
medium
Low
very low
n=

if categories are out of logical sequence, they then become nominal rather than ordinal

(Likert scales are always ordinal)

“individual ordinal data” is currently atypical

Going below 0 indicates interval level
32°F = 0°C
0°K = -459.6°F = -273.1°C

Ratio: 0 is absolute — nothing lower than 0 being measured (ex: Kelvin temp scale)

(grouped frequency distribution — Class intervals: lists a low & high for a range of scores
ex: age 80-89, 70-79, 60-69, 50-59, 40-49, 30-39, 20-29, 10-19, 0-9

class intervals must be closed-ended (both low & high value); equal in magnitude (each is same size)

above age 89 = open-ended class interval
below age 10 = closed-ended class interval

group interval level data, when grouped, must follow the two bolded rules above, without exception

Advertisements
Both comments and trackbacks are currently closed.