Please enable JavaScript.
Coggle requires JavaScript to display documents.
research methods - Coggle Diagram
research methods
experiment types
lab: a controlled setting. increases the replicability of the exepriment as the extraneous variable can remain fixed and so won't influence results: e.g. little albert- classically conditioned due to rat and loud bang. but is arteficial and not reflective of real life.
field: occurs in anatural setting but with the IV manipulated. allows for participants to be in a natural setting, reduces demand characteristics. not controlled so may be influenced by exernal factors, conditions may be uncomfortable or distracting
quasi: the IV is naturally occurring. e,g differences between groups. easily comparable, independant groups harder to find so many participants.
observation: where the researcher is watching and recording behaviour. good if there are multiple observers, reduces bias, demand characteristics if participants know they are watching- may not behave normally. doesn't prove cause and effect, and possible lack of consent.
self report- participants record own thoughts and behaviour, non observable traits can be noted usually in the form of questionaires and interviews. researcher bias- may interpret answer in the wrong way, or participant might lie to appear better in research despite it being anonymous.
sampling techniques
-
The volunteer sample involves posting an advertisement or asking for volunteers. risk of volunteer bias, as they may vary from non volunteers, do they have something they want to show or a specific reason behind volunteering- e.g. milgrams
random sampling involves drawing names of participants from a hat or a wheel spin to select a sample. this can come from a data base. everyone has an equal chance of being selected. hard to organise
oppotunity sampling means that a target population may be approached and asked to participate, these are people who are already available. maybe unrepresentitive
stratified sampling means participants are selected in proportion to the population targetted. requires population info
self report methods
questionaires: a set of written questions answered by participants. open questions are ones which the participants writes their own answers and allows them to elaborate on why. closed questions have set answers and the participant must choose one which resonates the most (quanitiive data). ranked scale questions =rate something out of 5. likert scale = agree or disagree and to what extent.
collects data quickly, standardised questions are easy to compare
questions could be misunderstood so answered wrong, socially desirable answers
interviews: this is where a researcher/ experimenter sits with the participant to ask them questions about the specific topic.
structured interviews involve set questions so the answers of particpants are easy to compare but theyre inflexible.
a semi structured interview has set questions, but follow up questions can be asked. more detail, less standardised
an unstructured interview starts with the same question but allows for free conversation. a lot of detail but hard to analyse.
how to improve: unambiguous questions, avoid double questions, pilot the questions to make sure they're easily understandable, keep questions relevant to the aim.
experiment design
indpendent groups: there are separate ssets of participants for each condition of the experiment. time consuming, less risk of order effects and participant variable may effect the result
repeated measures means that the same participants are used for each of the conditions of the experiment. it runs the risk of participants being influenced by previous conditions or potentially fatigue, but does mean participant variables are controlled
matched pairs: participants take part in relevant conditions but are matched to the conditions based on their charcteristics. takes time, but reduces participant variables.
reliability vs validity
RELIABILITY = CONSISTENCY
test-retest= same test to same participants to compare results
inter-rater= check if two researchers agree
standardisation= consistent instructions
VALIDITY = DOES THE STUDY MEASURE WHAT IT INTENDS TO
internal= does the IV cause the observed change in DV?
ecological= do the findings reflect real life
population= can it be generalised to the target population
temporal= applicable to different times
types of data
nominal- categories with no meaningful order
ordinal- ordered or ranked data
interval- data with equal intervals between values.