First 896 words of the document:
AS Research Methods
TOPIC ANALYSIS & EVALUATION
Types of research Lab experiment Strengths of lab experiment:
A true experiment has three key features: Extraneous variables can be controlled we can infer
methods An independent variable (IV) manipulated by the cause & effect
Experiments researcher in an attempt to produce a change in Greater control allows ability to replicate study thus
(lab, field & the dependent variable (DV) ensures reliability
natural All other variables which might influence the Weaknesses
experiments) results (called extraneous variables ) are held Artificial conditions may lead to low ecological validity
Correlations constant or eliminated. Variables that cannot be may not generalise to real life situations
Naturalistic controlled are called confounding variables . Higher risk of demand characteristics etc results may
Participants are randomly allocated to the lack validity
experimental conditions or take part in all Results more likely to be biased by sampling e.g. people
Questionnaire conditions willing to participate in study
surveys Note: a lab experiment occurs when conditions are May raise ethical issues of deception
Interviews controlled does not have to be in a science lab! Total control over all variables not possible
Field experiment Strengths of field experiment
The IV is still manipulated but the study is Being carried out in natural environment means less
carried out in the participants' natural artificial & higher ecological validity
environment Less bias from sampling, demand characteristics etc
Natural (quasi)experiment Weaknesses
The IV is changed by natural occurrence (e.g. More difficult to control all variables may not be able to
sex) infer cause & effect
More difficult to replicate may not be reliable
Correlations Ethical problems e.g. privacy, consent etc.
These measure the corelationships between Strengths of natural experiment
variables. high ecological validity
A correlation can be Very little bias or demand characteristics
positive or negative. If Weaknesses
when one variable Hard to infer cause & effect due to little control
increases the other also Very difficult to replicate
increases then the Ethical problems e.g. privacy, consent etc
correlation is positive. Strengths of correlations
If when one variable Correlations are ethical as there is no manipulation of
increases the other variables e.g. would be unethical to cause stress in an
decreases then the correlation is negative. If experiment but can measure existing stress levels in a
there is no relationship between the variables correlation
then there is no correlation. Can be used to explore relationships in complex situations
The correlation coefficient is number between 1 and can suggest ideas for further research
and +1 that shows the strength of the relationship Weaknesses
between the two variables. Because no IV being manipulated & extraneous variables
+1 = perfect positive correlation not controlled it is impossible to establish cause and effect
0 = no correlation using correlational analyses we cannot say that one
1 = perfect negative correlation causes the other only they are linked
0.3 = weak correlation Non linear relationships cannot be measured by
0.5= moderate correlation correlation.
0.8= strong correlation We may detect spurious relationships that do not exist but
A scatter gram can show correlation have just occurred due to coincidence
Strengths of naturalistic observations
Naturalistic observations Realism behaviour is observed in a natural setting with
Observations involve watching and recording behaviour. no attempt to change it it is likely that people will
Observations may be disclosed (when the participants know behave normally.
they are being observed) or undisclosed (when the Low demand characteristics and experimenter effects
participants do not know they are being observed). when people are unaware that their behaviour will be
Naturalistic observation refers to the observation of observed it is unlikely that it will be influenced by demand
behaviour in its natural setting. The researcher makes no characteristics and experimenter effects.
attempt to influence the behaviour of those being observed. New ideas can provide new ideas for research that can be
The aim is to study naturally occurring behaviour.
studies more systematically later.
Researchers may hide themselves or use video cameras.
Testing existing findings can be used to discover whether
In observations data can be sampled by:
Time sampling findings of lab experiments are reflected in real life
here you record what behaviours are occurring at certain situations.
time intervals e.g. what a child is doing every 5 minutes Weaknesses
Event sampling Lack of control the researcher has no control over the
Here you record how many times a certain behaviour or variables & cause and effect cannot be established.
behaviours occur in a set time period e.g. observe a monkey Meanings observations can describe behaviour but are
for 30 minutes and count how many times it shows unlikely to be able to tell us why it took place.
aggression, grooming behaviour, eats etc Observer bias observers are not passive, they select what
they see and tend to see what they expect to see.
Observer effects If people are aware they are being
observed they may wish to be seen in a favourable light
and alter their behaviour
Other pages in this set
Here's a taster:
Questionnaire Surveys Ethical issues in many cases the people have no idea that
A questionnaire survey involves asking participants they are being watched. This means ethics dictates that
questions about, for example, their attitudes, observations can only be carried out in public places.
behaviours or intentions. The use of a questionnaire However, people may still object to being filmed without
allows the researcher to gain information from large their knowledge.
numbers of people relatively quickly and efficiently.…read more
Here's a taster:
When sources can be retained or accessed by others (e.g.,
used. Once the items to be recorded have been chosen back copies of magazines or video's of people giving
a checklist can be developed to record the speeches) findings can be replicated and so tested for
observations from the sample. reliability.
weaknessess of content analysis
Observer bias reduces the objectivity and validity of
findings because different observers may interpret the
meaning of the behavioural categories differently.…read more
Here's a taster:
Participant variables can never be matched in all respects
Matching participants is time consuming and difficult
More participants are required
Pilot studies Pilot study can find out any possible problems with the
Pilot studies A small scale trial run of the actual study that is done procedures before invest huge amounts of time & money.…read more
Here's a taster:
Don't use more than 100 participants but generalise
Selection of findings to whole population. The participants
participants selected are called a sample .
Sample is taken from a larger population called the
target population , which consists of all members of
the group from which the sample has been taken.
In order to be able to generalise the sample must be
representative i.e.…read more
Here's a taster:
Reliability & Reliability Assessing & improving reliability
validity Results are reliable if they are consistent. E.g. if an Partial or complete replication of studies
Results of studies experiment is repeated then if the results are the same In observation studies, behaviour should be observed by
each time it is reliable more than one observer. If both observers agree then the
need to be both
data is reliable (interobserver reliability. Behaviour
reliable & valid Validity checklists help with this.…read more
Here's a taster:
Graphs & charts Choose a bar chart if data is in categories that are distinct from each other and not continuous (called nominal data)
Bar charts, Choose a histogram if the data is in categories that are continuous with each other (e.g. age 1120, 2130 etc)
Choose a frequency polygon if data is like histogram but you want to compare two sets of data
polygons, Chose a scattergraph if you are doing a correlation (i.e.…read more