Content Analysis – Univariate Statistics

Sep 17, 2008

Content Analysis – Univariate Statistics

Nominal:
- Numbers assigned to categories to signify qualitative differences
- Basically, nominal figures simply tell us that two things are different
  - Generated by arbitrary assignment of numbers to categories
  - No bearing to any objective or subjective scale.
  - Examples:
    - In sports, jersey numbers
    - In content analysis, used whenever content coded categorically
    - Hibbarb and Keenleyside
Ordinal:
- Numbers assigned to categories to signify preferences
- Basically, ordinal figures tell us two things are different and that one is better
  - Generated by arbitrary assignment of numbers to categories
  - Have bearing only to subjective scale
  - Examples:
    - In pop culture, rankings arbitrarily assigned to favourite songs.
    - The scale of difference between what is popular and what isn’t\
    - Rank –> Points = Difference.
Interval:
- Numbers assigned to categories based on a fixed scale with equal intervals between points.
- Basically, interval figures tell us two things are different and that one is better but also indicates the scale of difference.
  - Generated by measurement against fixed scale
  - Scale may be subjective or objective
    - Examples:
    - Thermometer, interval between 4 and 5 is the same between 5 and 6.
    - Communications & cultural studies, used in Lykert scales.
Ratio:
- Numbers assigned to categories based on a fixed scale with equal intervals between points and a true zero point.
- Basically, the most powerful – they allow for direct comparisons. Ratio figures tell us two things are different and that one is better and also indicates an absolute difference.
  - Generally by counting
  - Or measure against fixed, objective scale with true zero point.
  - Examples:
    - Age, starts at birth (the zero point)
    - Content analysis, numbers generated by counting responses
Descriptive Statistics
- Data set:
  - Data = information gathered by observation
  - Set = all data related to single phenomenon
- Univariate analysis:
  - Description of a single data set, without reference to any other data set.
- Frequency (n): number of observations for each response often expressed symbolically as “n”.
- Range: complete spectrum of observation in a data set. E.g. Grundy = F-A-G. Agazi not.
  - Class marks = 58, 61… 85, 90
  - Range = 58 to 90
  - Range size = 32
- Median: precise centre of range, regardless of weight observation
  - Example: class marks —– 58, 61… 85, 90
  - Middle number = 70 + 72 = 69
- Mean or Average: precise centre of the weighted observations
  - Example: class marks — 58, 61… 85, 90
  - Mean = sum/n
  - 1420/20= 71
- Mean deviation: average difference of each observation from precise centre of all weighted observations.
  - Example: 58, 61… 85, 90
  - Mean = 71
  - Deviation 71-58=13
  - 71-61 = 10, etc…
  - Total sum of deviations = X
  - x/n – 134/20= +/- 6.7
  - Example convoluted
    - Range = 58 to 90
    - Median = 71%
    - Mean deviation = +/- 6.7%