What are the uses of observational methods?
- only 1 species have language so questionnaires have limited applicability
- apparatus limits generalisability to the species/individuals who are familiar with it
- behaviour can be difficult to replicate in controlled environments eg. hunting
define ‘operational definitions’
to specify the physical requirements for the coding behaviour
define ‘ostensive definitions’
operationalising the coding behaviour through providing examples through diagrams/written descriptions
define an ‘event’
a coding behaviour that is occurs in a short duration of time
define a ‘state’
a coding behaviour that is occurs in a long duration of time
what is an ethogram?
a complete table of all coding behaviours observed in the study species
what are the types of measures?
- latency (how long it takes to respond to a stimulus)
- frequency (the total number of observed behaviours during the observation period)
- rate (the frequency of the coding behaviour per unit of time)
- duration (the total amount of time that a single occurrence happens during the observation period)
- proportion (the proportion of total time that a behaviours occurred of the proportion of total behaviours that occurred)
what are the scales of measurement?
- nominal (categorical)
- ordinal (order)
- interval (0 is arbitrary)
- ratio-interval (continuous)
what are the different sampling rules?
ad libitum - observer writes down everything that seems interesting
:( underestimates less conspicuous events
focal sampling - an individual is isolated for observation
:( time consuming
scan samples - a number of individuals samples at the same time
:( overestimates conspicuous events
behaviour sampling - all occurrence sampling
:( overestimation of conspicuous events
what are the types of recording methods?
time sampling
:(can underestimate behaviours of short duration
continuous recording
:( underestimate long duration behaviours
what is a coding scheme?
a way of categorising behaviour
what is intra-observer reliability?
the degree to which measurements taken by the same observer are consistent with one another
what is inter-rater reliability?
the degree to which measurements taken by different observers are consistent with one another
what are some problems of coding schemes that might cause poor inter-rater reliability?
- poor behavioural record
- coding scheme not detailed enough
- the target behaviours may be too subtle
how can you ensure inter-rater reliability when using a coding scheme?
- train coders
- get 2 coders to code the same data and compare them for inter-rater agreement
what is the formula for inter-rater agreement?
number of times raters agreed/total number of observations
how do you work out percentage agreement?
times the IRA by 100
why is cohen’s kappa used?
because it controls for probability of agreement by chance
what is the formula for cohen’s kappa?
p(observed) - p(expected by chance)/1-p(expected by chance)
what is the p(observed) value?
the inter-rater agreement, the amount of times the codes agreed divided by the total number of observations
what is the p(expected by chance) value? What is the formula
probability of observers saying yes by chance - (a+c)x(a+b) + probability of observers saying no by chance (b+d)x(c+d)
how can you interpret cohen’s kappa value?
can range from -1 to 1
0=bad 1=perfect
negative numbers = agreement is below that expected by chance
