Design a sampling strategy

• Relevance: Define what is a relevant text
  • What will you count?
  • What will you ignore? And why?
• Selection Method
  • How will you select individual texts?
• Random Sample
  • Every item has equal chance of being selected
  • Good if no significant differences in population.
• Systematic Sample
  • Items selected according to regular pattern
  • Selects texts from entire distribution of population.
• Example:
  • 20% of stories in newspaper for one month
  • Random Sample:
    • Write dates on paper scraps: pull six from hat
    • Number days of week 1-5, roll a die
  • Systematic sample
    • Select every fifth day
    • Select every fifth article, each day
    • Five aspects
    • Time frame for the entire population of data
    • Frequency new content generated/distributed (estimated size of universe or population)
    • Anomalies in the medium (may effect the data you collect)
    • Define what is a relevant text
    • Method of text selection (will affect the data you collect)

Content analysis

• Analytical category:
  • Specific content and/or quality to be studied
  • Set of content and/or quantities you count as you examine the medium.
    • Should be clear from your research question
    • Must collect the data you require
• Identifier
  • Name of text
  • Author/producer of text
  • Location/data of production and/or distribution
  • Position of text within medium
    • E.g. page nubers, time into broadcast program
  • Size or duration of text
• Analytical
  • Relevant data which will answer your question
  • E.g. symbols, words, images, colours, topics.
• Define the analytical categories
  • Kubrin (p. 368), do song lyrics legitimize violence? She looks for these elements in lyrics: respect, willingness to fight, wealth, retaliation, sexism, nihilism. These are her analytical categories; 6 altogether.
• Create a Code book
  • Process of converting complex qualitative data into simple coded data to facilitate quantitative data.
    • Generally, codes are symbols or numbers
    • Codes should facilitate counting, measurement
  • Two advantages:
    • Accelerate data entry; only enter single symbol or number rather than a word or phrase.
    • Numbers much easier to analyze with computer
  • Code book
    • Explicit rules to guide interpretation and coding of text, presents your interpretation of your analytical categories.
    • Always helpful, but particularly when:
    • Texts are vague or ambiguous
    • Other researchers want to replicate your study
    • There is more than one person coding
  • Code book: example
    • Analytical category = sexuality
    • How should we code these two figures
      • Queer
      • Straight
      • Ambiguous
• Coding Schedule
  • A template with all categories of analysis listed, and all possible answers also listed. In many projects using content analysis, you need one schedule for each text examined.
• “Piloting” or testing:
  • Test your research plan on a limited sample of texts
  • This allows you to work out bugs:
    • Will medium and sample stragey produce data?
    • Is there an anomaly in the medium
    • Do codes cover all significant possibilities
• “Conducting” the research:
  • Collecting data using your research plan (aka actually doing the research)
  • How?
    • Everytime you find something?