types of experiment
Lab experiments-Involve IV being manipulated in a controlled environment. EVALUATION :( in an artificial contrived enrironment so lack realism and ecological validity (not true to the real world). Can produce demand characteristics-were participants adjust their behaviour to what they think the investigator wants.People may also change the way they act in order to be socially desirable. :)because the experiment is standardized it can easily be replicated to test for reliability of results (see if similar results are obtained when someone else does the test). Because of the environment it is easy to control extraneous variables in order to avoid confounding variable which affect change in the DV.Natural experiment- the IV occurs naturally because it is unethical to manipulate it ( for example studies on privation) the environment is natural. EVALUATION :( there is lack of control over extraneous variables because of the environment.making it difficult tpo establish cause and effect because experimentor is unsure wether the IV caused changes in the DV or wether the change was due to confounding variables.:) there is a reduction of demand characteristics as participants may be less aware that they are taking part. because of natural environment there is high ecological validity.Field-the IV is manipulated but the environment is natural. EVALUATION :( lack of control because of the environment. Difficult to replicate due to the lack of control. :) has higher levels of ecological validity compared to lab experiment becasue it is in a real life setting.there is less demand characteristics as people are less aware that they are taking part in an experiment.
Correlational analysis-is a technique wich measures the relationship between two variable. +Correlation- as one variable increases so does the other. -correlation- as one variable increases, the othere decreses. Correlation coefficient gives a value which tells you the strength of the relationship and wether it is positive or negative ( eg. -0.5 is a weak negative correlation, + 1.0 is a perfect possitive correlation. 0 is no correlation) EVALUATION-:)gives a quantitive measure of relationship:( can not establish cause and effect.
Inderpendent measures design: each participant takes place in only one condition.:( Results may be due to parti and not the conditions.cipant variable/ individual differences. having more participants reduces this
Repeted measures design:each participant takes place in both condition :( the difference in data from the different conditions may be due to order effect. Either fatigue or practice. this can be overcome by counterbalancing. half of participants do condition A first other half do B then they swap.
Matched paris design:each participant takes place in one condition but is matched to another participant on key charachteristics such as age. :( it is time consuming
measures of central tendancy and dispersion
MEASURES OF CENTRAL TENDANCY
MEDIAN-excludes extreeme score:) , only takes 1/2 middle values into account:(
MEAN-takes all scores into account:), however this means it can be distored by extreeme values.-making it unrepresentative:(
MODE-uneffected by extreeme scores:) , can be affected by change in one score making it unrepresentative:(
MEASURES OF DISPERSION
RANGE-easy to calculate:) , easily distorted by extreeme values:(
STANDAR DEVIATION- takes all score into account so it sensitve:) difficult to calculate :(
Content Analysis- from qualitative data you draw out a theme and count how many times it occurs. for example in magazine articles see how many times sexist references appear.?
Pilot study-preliminary study on a smaller scale. it enables researcher to test their procedure or design before investing money and time into the actual study. allows improvements to be made.
Ethical issues-informed consent-participants agree to take part with knowledge of what they will have to do in the study.However participants that agree may not fully do so.Also some studies will work best if participant does not know every detail as it may increase demand charachteristics. to overcome this the researcher can gain presumptive consent.
Deception -if a researcher decieves participant it should be aproved by ethics commite prior to study. participants should be debriefed after study so that if they wish to they can withhold their data.
Protection from harm- participants should avoid any risk that are greater then what they would experience in everday life.however researchers can not always predict the risks of taking part in the study.
Confidentiality-false names shouls be used when refering to the participant . may not be possible if information reveals location which is obvious to participant.
Right to withdraw- people can withdraw at any time.