Please enable JavaScript.

Coggle requires JavaScript to display documents.

PS906 Observations - Coggle Diagram

- - - - appears to measure what it is intended to measure
  - - - require solid theory
      - good description of what we want to study
    - - reliability
      - validity
- - - - design of modern workplaces
        
        natural light
        
        ergonomic workstations
        
        collaborative spaces
        
        hot desking
        
        reduced sense of belonging
        
        disrupted workflow
        
        increased stress from uncertainty and lack of control
    - - norms, values, behaviours
    - - effective communication / trust
      - shared goals in team performance
- - - - observer records in consistent manner
      - agreement between different observers
    - - measure what intend to measure
      - reliability between and among observers is a necessary but not sufficient criterion for validity
  - - - thoroughly describe what you want to study
    - - all observation is biased - never neutral
      - guided by theory
      - made explicit
      - emic vs etic
        
        emic
        
        inside
        
        functional contrast
        
        etic
        
        outside
        
        quantitative
        
        measure discrete behaviours
    - - field or laboratory
      - observer bias
        
        expectations and knowledge that observers have about participants might influence
        
        the objectivity of their observation
        
        the ways they treat participants
        
        data derived from the observation
        
        how to overcome?
        
        AWARE
        
        Double-blind procedures
        
        neither observer nor participants know the treatment / experimental condition
        
        Replication
        
        get same result using same measures with similar participants in similar situations
      - when to observe?
        
        impossible to observe all participants all the time
        
        preliminary observations are necessary to determine the best times for specific behaviours to be observed
      - how much to observe?
        
        more is more
        
        larger samples (of participants / observations) more closely represent the population
        
        allows for generalization
        
        it depends on theory and RQ
    - - driven by
        
        RQ
        
        the way the question interfaces with the specific situation in which the observation is made
      - you can
        
        develop your own to reflect specific RQ
        
        adapt a coding system already developed by someone else
      - properties
        
        homogeneity
        
        mutual exclusivity / independence
    - - measuring behaviour
        
        different ways in which categories can be quantified
      - sampling behaviour
        
        decision on how much to observe
    - - Factors affecting reliability
        
        Observer Fatigue
        
        how to manage?
        
        shorten length of observation session
        
        rotate observers
        
        record behaviour in the field and then code later
        
        Observer drift
        
        over the course of a study, same behaviour is coded differently
        
        how to manage?
        
        monitor & retrain observers
        
        make inter-observer checks
        
        Category Definition
        
        Clearly defined categories are more reliable than ambiguous categories
        
        how to manage?
        
        use objective category definitions where possible
        
        define a category in terms of core dimension and borderline dimensions