Other slides in this set
Here's a taster:
Reliability is the consistency of the results: are they trustworthy? Do they have a high dependability?
Reliability indicates that the results are accurate- so firm conclusions can be drawn. There are two key
measures of reliability:
· Test-retest reliability: Will the results stay the same over time?
· Inter-rater reliability: Will the results stay the same no matter who's measuring them?
· Validity means that we're testing what we set out to measure; this allows to generalize the results.
There are also two key types of validity:
· Internal validity- Am I measuring what I think I'm measuring? It is all about the level of control the
experiment has over other variables, (Extraneous variables) that might impact our results.
· External validity is the extent which results can be generalized (externally) i.e. to the world/ a country/
to certain types of people. There are two types of external validity:
Ecological validity: Can it be applied to other settings? For example, a laboratory controlled experiment is
less likely to have ecological validity because the measures are artificial, and may not represent real
life tests of memory or attachment, for example.
Population validity: This is the extent which the results can be generalized to people, i.e. the population. Is it
a test of real life?
A balance of external and internal validity needs to be established: you can't have too much control (internal
validity) because it becomes artificial and lacks in mundane realism. It also can't have too much
external validity, because we aren't keeping other factors in control that might impact the result.
The experimental method is all about establishing a causal relationship: what the underlying impact of a
variable is on another.…read more
Here's a taster:
This is when variables other than what you set out to manipulate (i.e. the independent variable, the
amount of alcohol for example.) effect your results: the dependant variable. Therefore you need to
consider what else is effecting what you're measuring. For example, when measuring alcohol on the
effects of memory, we need to consider what else might effect memory: e.g.- amount of sleep,
caffeine, units of alcohol, etc.)
· It's important we control these Extraneous Variables, because it can reduce our internal validity. EVs
include participant, and situational variables:
· Participant variables are about who the person is: their age, gender, intelligence, and experience that
could alter the results. Everyone is different, so there's a need to control participant variables. To
control this, experimenters can randomly assign people to each condition to `spread out' participant
· Situational variables, are when demand characteristics are provoked. This gives participants ideas in
the research environment leading participants to work out what the experiment hypothesis. This can
ruin the hypothesis, as it cause these demand characteristics. In order to control this; experimenters
use the single blind technique so participants don't know what condition they're in. For example,
using the same methods in both controls, such as a placebo model in the control. This stops demand
characteristics from occurring.
· Investigator effects- this is when the investigator behaves in a way which increases the chance of
their hypothesis being correct: this makes the participants to misinterpreting results, and makes the
experiment bias. To control this, double blind is used- the researcher and the participants don't know
what each condition is, reducing both sets of demand characteristics.
· Order effects is when participants get better or worse between both conditions: either from rehearsal
or fatigue. In order to control this, counter balancing is needed: so half do one condition, while the
other half does the other conditions, and then these swap.…read more
Here's a taster:
There's the general population; the target population, and the sample. Therefore, you need to make sure
that this sample is a representative of your target population, which may, or may not be generalizable to the
· If the sample is not a representative of the target population it's bias. For example, the Strange Situation
(Ainsworth) was bias, because it was not representative of target population: it was bias to westernized
· External validity is the extent finding can be generalized to the population: so having the correct sample
size, will increase the experiments external validity: it should have a high population validity.
· There're three, key sampling methods:
· Opportunity sampling, is when people are selected who are the most easily available. (For example, the
first 20 people that pass you on the street.) This is good, as it is widely used; is less time consuming, and
easily able to gather people. However, it is does have cons. It's bad because the sample can be bias. You
also don't have access to the target population; and you're most likely to approach certain people.
· Volunteer sampling, this is when psychologists ask for participants through advertising. This is potentially
high in representative participants for the target population. It is also an effective way to gather participants;
however, people who volunteer might have a certain personality type. (participant variables) For example,
they might be more motivated (hence for volunteering) or it could be based on something they're interested
in, making them not a good choice to the population.
· Radom sampling aims all people in the target audience, where they all have an equal chance of being
selected. This is potentially unbiased and well controlled, however is difficult and time consuming.…read more
Here's a taster:
There're three main sampling types:
Naturalistic experiments: this is an experiment which takes place in a natural environment; with natural independent variables which is not
manipulated by the researcher. The experimenter makes the most of these naturally occurring IVs; for example, monitoring people in
institutionalization against those who are with their biological parents: you wouldn't take the person out of their family and put them into
· These experiments have high ecological validity because it can easily be applied to life situations: meaning results can be generalize.
· Allows research into areas of study that would otherwise be unethical- it is more ethical; we couldn't examine the effects of extreme
privation, for example.
· Higher mundane realism and external validity.
· It is hard to replicate as IV variables are out of the experimenters control.
· A lack of control means it is open to many extraneous variables that might effect the outcome.
· Can not fully demonstrate causal relationships as the IV is not directly manipulated.
· Participants may be aware of the study; meaning the act differently as they're being monitored (demand characteristics.)
Laboratory experiments- take place in a controlled laboratory condition; to investigate causal relationships under high levels of control.
· Extraneous variables are and participant allocation allows accurate establishment of causal relationships.
· A high systematic methodology means it's easily repeatable.
· There's a high internal validity due to the levels of control.
· Demand characteristics can occur, this is a different environment type than the real world.
· It is artificial so lacks ecological validity and external validity- meaning it can not be generalised.
· Prone to investigator effects, and these factors means it lacks mundane realism.
Field experiments take place in an environment where behaviour is naturally occurring ; however the independent variable is manipulated. It
therefore combines both aspects of natural, and laboratory based ideas; meaning it can establish a balance of internal and external validity.
· There's a high ecological validity.
· Reduction in demand characteristics, participants are not often aware of participating in a study.
· Good mundane realism: it creates a good balance in validity, so causal relationships can be established.
· Not as easily repeatable, and also not a full control of extraneous variables.
· Difficult to assign participants to each condition on a random bases- lack of control.…read more
Here's a taster:
Experimental designs is the way in which participants are allocated to the conditions in your study.
· Repeated measures: Getting participants to do conditions 1 and 2, using the same people to compare the
· This is good because: it increases accuracy as participant variations won't effect the study.
· However, order effects occur, and could increase demand characteristics by doing both variables.
· This is solved through counter-balancing, and using the single blind technique so demand characteristics don'
· Independent measures: this is when you have two separate groups for each conditions.
· This is good because there are no order effects, and demand characteristics are less likely to occur.
· However, independent measures are bad because it can result in participant variables between the two
groups. Twice as many participants are also needed.
· To control this, random assignment in used to balance out differences.
· Matched pairs: this is when experimenters recruit people by indentifying the target population, and basing the
second condition of people to match participant variables/ needed and match able qualities that might effect
the results. This can then be used to compare conditions. It's more accurate than the other experimental
· This is good because it decreases demand characteristics, no order effects, and also relevantly matched
participant variables that could affect the results of the study.
· However, it's bad because you can rarely find exact matches: no two people are the same, it's time
consuming as you need to find participants that match in lots of aspects.…read more