Week 10 (Observation)

Question 1

Observational methods

Answer 1

Observing a relatively unconstrained segment of individual’s freely chosen behaviour within a particular setting
-No need to request info
-No intervention or manipulation of behaviours

-Trained individuals record activities, events, or processes as precisely and comprehensively as possible.

Question 2

Pros and Cons of observational design

Answer 2

PROS
-Produces data on real behaviour rather than possibly distorted self-reports
-Can gather data on behaviour to amenable to experimentation (e.g, playground aggression)
-If conducted in field settings , can often gather data on unforced behaviour

CONS
-People’s behaviour can be affected y awareness of being observed (reactivity effects)
-Can be very time-consuming compared with the experiment or questionnaire study
-May not be possible to identify cause and effect

Question 3

Controlled vs Naturalistic observation

Answer 3

Controlled: Observation in controlled setting, often a laboratory or observation room.
-Artificial, potentially inhibiting atmosphere
-Limits external validity: how well the results of the study be expected to apply to other settings

Naturalistic: Observation of people without intervention in their own environment
-Access to genuine behaviour, high level of external validity

Question 4

Participant vs non-participant observation

Answer 4

Participant
-The observer acts in the observed group
-Qualitative approach to observation

Non-participant
-Observation from a distance. No interference with the behaviour being observed.
-Structured and qualitative approach to observation

Question 5

Disclosed vs Undisclosed observation

Answer 5

-Distinction based on whether or not the researchers disclosed to the participants that they are being observed of research purposes
-Info given can be partial, or can involve some amount of deception, in order to encourage more genuine behaviour.

-Undisclosed (covert) observation has ethical issues
-As participants don’t give consent, and don’t have the option to opt out
-But protects researchers from harm and allowing access to sites that otherwise would not have been possible to study

Question 6

Reactivity effects

Answer 6

Tendency to alter behaviour in response to awareness of being observed
-One way to reduce this is to become a predictable and familiar part of the environment.

Question 7

Structured vs non-structured observation

Answer 7

Structured (or systematic) observation
-Relies, ideally, on representative samples of behaviour
-Define behaviours before starting observation
-Observed behaviours are coded using these pre-established categories
-Typical in quantitative studies

Unstructured observation
-Recording takes place after the observation
-No restrictions placed on what the observer will note
-Typical in qualitative studies, all aspects of the phenomenon that seems to be relevant are recorded.

Question 8

What is needed for defining behaviour categories

Answer 8

Need to be
-Mutually exclusive: behaviour an only be in one category, no overlap
-Exhaustive: All behaviours must be covered by the codes

Question 9

Types of observational sampling

Answer 9

Event sampling
Interval time sampling

Question 10

Event sampling (Observational)

Answer 10

Record previously defined behavioural pattern (e.g, child crying, picking up a toy, going to a parent ) every time it occurs

Question 11

Interval time sampling (Observational)

Answer 11

-Divide observation period into intervals
-Time intervals usually equal in length
-Record presence or absence of a behaviour within or at end of each interval
-The intervals can be consecutive or separated

Question 12

Prepping for structured/systematic observations

Answer 12

-Arrange for participants to be observed in specified settings for periods of time (we usually call these sessions)
-Establish systematic procedures or transforming the observed behaviour (live or recorded) into data (quantitative)
-Train observers in the use of the coding system
-Prepare measurement procedures to apply to behavioural records.

Question 13

Cohen’s Kappa Coefficient (K)

Answer 13

-Measure of inter-rater reliability. Takes into consideration the probability f chance agreements.
-Sometimes coders can agree by chance , artificially inflates the agreement rate
-Cohen’s Kappa takes chance agreement into consideration

Question 14

Cohen’s Kappa equation

Answer 14

K = O - E divided by N - £

N: Number of observations
O: Number of times observers agree
E: Number of agreements by chance

Question 15

Interpretations of Cohen’s Kappa

Answer 15

<.40 Unacceptable
<.40 to <.60 Fair
<.60 to <.75 Good
>.75 Excellent

Question 16

Inter-rater reliability

Answer 16

-A measure of reliability (consistency) among coders (observers)
-Checks how much coders “agree” on how behaviour is categorised

Question 17

Why inter-rater reliability is important

Answer 17

-Observer bias: when observer’s characteristics influence what and how behaviour is coded
-Expectancy effect: coder’s expectations (for instance about the hypothesis) influence how they interpret behaviour
-Information bias: error resulting from systematic, selective, incomplete, or inaccurate recall of observed behaviour
-Need to consider potential sources of subjective bias when recruiting coders and creating coding terms.