human research Flashcards
what is quantitative data
data that can be counted or expressed numerically eg. number of aggressive acts observed
what is qualitative data
cannot be expressed numerically but is concerned with words, texts ideas and themes
what is tally sampling
not down in the appropriate cell each time they see a person in a certain category, categories need to be clear and unambigous
what is time sampling
making observations at different time intervals, may not always be representative as certain behaviours can be missed
what is event sampling
recording a certain behaviour everytime in happens, researchers may not be able to record all instances
what are naturalistic and structured observations
naturalistic - behaviour of participant within their own environment
structured - carried out in an environment in which the researcher has some control, subsequent behaviour can be observed behind a one way screen, numerical data generated, few observers to validate study, more reliable
what is a participant and a non participant observation
participant - observer is part of the situation, also a participant, may be hard to record notes of the observation
non participant - observer not part of the situation, limted in terms of validity as presence may change participants behaviour, concentrate on recordings
what is an overt and a covert observation
overt - person being observed is aware they are, informed consent and right to withdraw
covert- participants are unaware, ethical issues, less demand characteristics and exerimenter effects so more valid
what is a content analysis
-observational study is observed indirectly in written or verbal materia
- quantitative data when recording units such as themes, tone or individual words counted
- qualitative data is excerpts from texts used
- a hypothesis generated, decisions about a sampling method, decisions of coding units/frames, operationalise, pilot study
evaluate observations
+ high ecological valididty, direct observations of real media which are current and relevant
- researcher bias reduced objectivity and validity, coding units may be open to interpretation, researcher decides if it fits a coding unit
+ when resources are retained and can be assesed by other findings can be replicated and tested for reliability
- if there is more than one observeer and they are inconsistent in application of coding units, the inter rater reliablity reduced
+ if qualtitative data is collected, inferential stats tests can be carried out, increase objectivity
+ inter rater reliability by using more than one coder to code info used, increase objectivity
- issues with internal validity, cant be sure categories represent what they intend to measure
- likely to be culture dependant because interpretation of verbal or written content will be affected by language and society
