Content analysis

• Content analysis


  An indirect observational method where we observe human behaviour through the things humans make (artifacts)
• Quantitative data


  • Can make tables, charts, work out averages and do statistics on it
• Qualitative data


  Data in the form of words
• Content analysis


  1. Decide research question

  2. Select sample

  3. Define coding units/behavioral categories

  4. Operationalise categories

  5. Record in frequency table
• Artifacts for content analysis

  • 1950's advertisements
  • Art of the middle ages
  • Films from the late victorian age
  • Greek pots
• Content analysis allows us to give a quantitative description of qualitative human communication
• Coding units/behavioral categories


  • Linked to research question, need to be operationalised (defined so they can be precisely measured)
• Waynforth and Dunbar (1995) conducted a content analysis on 881 lonely heart adverts from 4 American newspapers
• Evolutionary theory of mate choice

  Men look for younger, more attractive mates as indicators of fertility, women look for older mates with more resources to provide for a family
• The content analysis backed up the evolutionary theory - men looked for significantly younger mates and women significantly older mates, men showed off resources more than women, and women mentioned their own attractiveness more than men
• Conducting a content analysis

  1. Decide measurable categories to record

  2. Tally each time a category appears
• Test-retest reliability


  Running the content analysis again on the same data on another occasion and comparing the results
• Inter-rater reliability


  Getting a second researcher to complete the content analysis separately using the same operationalised categories, then comparing the two sets of data
• Researchers generally accept a correlation coefficient of 0.8 between the ratings as showing the data is reliable
• Strengths of content analysis

  • High external validity as the material was not created for research
  • Easy to get a sample as the data already exists
  • Replication is possible
• Weaknesses of content analysis

  • Possibility of observer bias
  • Lack of validity as the data is not a record of actual behaviour
• Thematic analysis


  A variation on content analysis where the researcher starts by attempting to discover deeper meanings in the text/interviews, spotting patterns that can be coded, and then identifying themes (rather than using predetermined categories)
• Thematic analysis is intended to stop the researcher imposing their own ideas on the texts