Analysis

1
Analysis

• Making sense of the data

2
Analysis
3
Analysis

• You must sort your codes into some order or into
  groups.
• Hierarchical, group with subgroups
• Arrange into clusters of related codes
• Such arrangements help researchers
  dimensionalize, or recognize the various
  alternatives for or dimensions of similar
  thoughts or behaviors
• E.g., thoughts about how to look masculine
• Short hair
• Plain shoes
• Shirt with collar

4

• Flat codingnonhierarchical list of codes (no
  subcodes)
• Voice is funny
• Sings poorly
• Whiney
• Tree codinghierarchical arrangement of codes
• Things that denote femininity
• Body
• Long hair
• high voice
• Clothes
• Girls shirt
• high heels

5
Analysis

• Relationships between codes become more apparent
  as codes are grouped
• Themes should be explored
• Why do some codes co-occur?
• Why are some dimensions related to other codes
  while others are not?
• Are some codes linked to particular emotions?
• Exploration of themes is analysis. The
  discoveries should be written down.

6
Analysis
As you analyze, new codes may be generated. This
means that you should retrospectively recode the
data that had already been coded.
Source http//onlineqda.hud.ac.uk/Intro_QDA/how_
what_to_code.php
7
Analysis

• As you group codes, you should keep memos to
  yourself. These eventually (with very heavy and
  serious editing) turn into your written text.
• Information to include in the memo about a code
• Why you created the code or category or theme
• Some information defining the code
• Information that says what the code reveals about
  the phenomenon you are studying
• Why you changed a code
• Ideas about the phenomenon that are generated by
  your coding activities in general

8
Analysis

• Data Displays
• Data displays are an organized way of compressing
  information and assembling it in ways that help
  you draw conclusions
• Can be text, diagrams, charts, matrices
• They show systematic patterns and
  interrelationships of the chunks of meaning
  (codes) in the data
• Displaying will often reveal new connections and
  themes in the data beyond those already noticed
• Can display intra-case analysis and/or cross-case
  analysis

9
Analysis

• Select Types of Data Displays
• Partially ordered displayssome but not too much
  internal order aiming to uncover and describe
  what is happening in the local setting no matter
  how messy or surprising
• Context charta network, mapping in graphic form
  the interrelationships among the roles and groups
  that go to make up the context of individual
  behavior
• Checklist Matrixformat for analyzing field data
  on a major variable or general domain of interest
• Transcript as Poemmake a poem

10
Analysis

• Select Types of Data Displays
• Time-ordered Displaysorders data by time and
  sequence, preserving the historical chronological
  flow and permitting a good look at what led to
  what and when
• Event listinga matrix that arranges a series of
  concrete event by chronological time periods,
  sorting them into several categories
• Critical incident chartlimited representation of
  critical elements of a process
• Event state networkcenters on general states
  linked to specific events
• Activity recordsequencing of routine events
• Decision modelingsteps in decision-making
  spelled out
• Time-ordered matrixcolumn arranged by time
  period in sequence so that you can see when
  particular phenomena occurred the rows are what
  else you are studying

11
Analysis

• Select Types of Data Displays
• Role-ordered DisplaysOrders information
  according to peoples roles in a formal or
  informal setting.
• Role-ordered matrixsorts data in its rows and
  columns that have been gathered from or about a
  certain set of role occupants
• Role-by-Time Matrixsorting role information over
  time

12
Analysis

• Select Types of Data Displays
• Conceptually ordered Displaysdisplays the
  concepts or variables.
• Conceptually clustered Matrixrows and columns
  arranged to bring together items that belong
  together. A prior derivation or empirically
  driven. May be ordered by persons or themes or
  both.
• Folk Taxonomydisplaying concepts in network form
• Cognitive mapsdisplays the persons
  representation of concepts about a particular
  domain, showing the relationships among them.
  Descriptive text is associated with it.
• Effects matrixdisplays data on one or more
  outcomes, in as differentiated a form as the
  study requires. Focus on dependent variables.

13
Analysis

• Conclusion Drawing and Verification
• As one creates and views displays, the salient
  components of meaning and activities become
  apparent.
• In descriptive analysis, the researcher tries to
  represent the data (meanings, observations) to
  readers in such a way that they will understand
  what the researcher sees in the data.
• In causal analysis, the researcher tries to link
  concepts in the data together to explain observed
  meanings or phenomena, and to represent that in
  such a way that readers will understand what
  the researcher sees.
• This stage relies very heavily on logical
  evaluation and systematic description

14
Analysis

• Conclusion Drawing and Verification
• The researcher must describe what he or she sees
  in the data.
• The researcher always refers back to the data
  displays and raw data as descriptions or causal
  statements are made.
• Systematic, organized, and good coding and notes
  will really pay off at this point, allowing
  efficient, accurate access to data
• Conclusions are made through the process of
  writing up (describing) what is in the data