Please enable JavaScript.
Coggle requires JavaScript to display documents.
RESEARCH METHODS PT2 (scientific processes) - Coggle Diagram
RESEARCH METHODS PT2 (scientific processes)
hypothesis
directional - hypothesis based off of previous research findings
non - directional - predict their will be a difference or a relationship, can't give a specific outcome as there is no previous research.
Sampling
random - no researcher bias (s) , freak sample (very similar people therefore findings) lower population validity (L)
systematic - chance I will end ups with very similar people (L), objective way to collect samples (S)
stratified - most representative, mini representation of target, can be generalised (S), time consuming - difficult to match the percentages of each group (L)
opportunity - quick and easy to carry out (S), end up with a bias sample (L)
volunteer - volunteer bias, get certain types of people (L), read and willing to take part (S)
Pilot studies
trial run at a smaller scale, aim to check the materials are correct, that you are following all the ethical guidelines, study is worth carrying out on a larger scale, to check you have controlled as many extraneous variables as possible
Experimental designs
matches pair - pair someone up with another who has similar characteristics as them, person A does condition 1, person B does condition 2
likely to be participant variables that affect results (L), less likely to have demand characterises (S)
repeated measures - same participants do both conditions (measures)
more valid, no participant variables (S), more chance of order effects, based on order in which they do the conditions (L)
independent groups - group A would only do condition 1 and are unaware of another condition and vice versa
cant have the order effects effect (S), participant variables will effect results, different people in each group (L)
Observational Design
behavioural categories
used for structured observations
create a preplan list of behaviors they will observe, these need to be operationalised, thus allows for consistency to make sure everyone is observing the same thing, also ensures the observation can be repeated in a standardised way
event sampling
note down everytime you see a behaviour on your list at a change of event (entirety of a Childs break time for example)
good representation - not taking snippets (S), time consuming, not everything will be relevant to the study (W)
time sampling
observe in time intervals (ex. every ten minutes) note down everything that happens then
miss out of important information in the time not recorded (L), less time consuming so more precise in observing behaviours (S)
Questionare contrusction
closed question - fixed responses they can choose from
open question - they can give as much or as little info as they want
Variables
dependent - one that you are measuring
IV - one you change
extraneous - other factors that have the possibility to effect results
confounding - any variable that cannot be controlled and WILL effect results
operationalise - make variables measurable, how you are going to measure something
Control
random allocation - used in independent groups design, way of assigning the participants to a condition without the researcher having any biases
counterbalancing - used with repeated measures design, a way of reducing order effects, they will split participants in half - half will do conditions 1 first then other half will of condition 2 first
randomisation - insuring everything in your study is free from research bias, bu using things like random sampling and random allocation
standardisation - making sure every participant experiences the study in exactly the same way, making sure everything is the same for everyone, done by giving same instructions, using the same confederates
demand characteristics, changing answers to fit what they think the researcher wants to hear
investigator effects - anyway the investigator has effected the way the participants feel, indirectly or directly