Literature Research Methods
Psychology selection round I
Experiments
Experiments can be done in any setting (o.a. laboratory and field experiments).
An experimenter manipulates something -> he makes something happen and observes
it’s effects.
Experiments specify the directions of causality
- in a correlational study you might find a correlation. But does A cause B or does B
cause A? Or does the third variable Z cause both?
example: experimenter manipulates A. If B changes and the experimenter rules out other
factors. A caused B.
That experiments make causality clear is a big advantage!
- another advantage is that they can rate out other factors
If a factor is relevant -> it makes a difference
manipulation: making the effect of a factor visible by making it different where nothing else
is, and seeing if the difference influences what the subject does.
note: there are various experimental designs.
The anatomy of an experiment
There is a skeletal structure that all experiments have; they only differ from how they put
flesh on the skeleton.
Independent and dependent variables
An experiment is defined by manipulation
manipulation: the experiment intervenes in nature to make something happen; that is ->
causing something.
independent variable: what the experimenter causes to vary
a variable must have at least two values otherwise its it’s not a variable
So there will be at least two values of the independent variable
- often referred to as different experimental conditions or experimental treatments
(because they specify how subjects are treated under various conditions)
The independent variable is the variable that is the intervention - what we make happen
The dependent variable is the data - what we observe (the effect)
, Control variables
An experiment, we say, asks how the independent variable affects a dependent variable.
But the world isn’t perfectly made for experimenters. There are a lot of other influences
that also affect the subject.
So in real world research our subjects behavior may be affected by our independent
variable, but our subject will also have to do with other influences.
To best see what effect or interdependent variable has, we want to get rid of these other
influences if we can. If we can’t do that, we want to equalize them, so that they affect all
experimental conditions in the same way and then drop out of our comparison between
them.
(Om zo goed mogelijk te zien welk effect of welke onderling afhankelijke variabele heeft,
willen we deze andere invloeden zoveel mogelijk kwijtraken. Als we dat niet kunnen, willen
we ze gelijk maken, zodat ze alle experimentele omstandigheden op dezelfde manier
beïnvloeden en dan uit onze onderlinge vergelijking verdwijnen.)
- deze hinderlijke variabelen heten control variables.
2 manieren om control variables constant te maken;
1. Control variables (variabelen die ook het object beïnvloeden naast de independent
variable) can be held constant directly, by the way we set up the sitaton.
2. Other control variables cannot be so directly controlled, we can not make all our
objects equally altruistic, or alienated, or distracted. What we can do is try to make all
groups of subjects equally altruïstic on average. ANd one way to do that is to assign
each subject strictly at random the ‘alone or to the ‘other onlookers condition. Then,
if we have enough subjects, the randomization procedure is likely to assign roughly
as many altruistic subjects to one group as to the other.
The larger the groups, the more unlikely they are to be very different on any variable.
The fundamental principle of experimental control: we control what we can; and what we
don’t control we randomize.
notice; what we controle, what we mainplate and what we measure, depends on the
question we ask. (An independent variable in one experiment may be a control variable in
another experiment).
There is a popular superstition that every experiment needs a control condition or control
group. Not so. But it is true that every experiment involves a comparison between something
and something else. What do we compare with that? It depends on the question we are
asking.
Any experiment will incorporate a comparison. There will be an independent variable
that takes on different values for different experimental conditions or experimental
treatments. And we will see how our subjects say or do depending on the value of the
independent variable. SO we measure what they say or do as the dependent variable.
Psychology selection round I
Experiments
Experiments can be done in any setting (o.a. laboratory and field experiments).
An experimenter manipulates something -> he makes something happen and observes
it’s effects.
Experiments specify the directions of causality
- in a correlational study you might find a correlation. But does A cause B or does B
cause A? Or does the third variable Z cause both?
example: experimenter manipulates A. If B changes and the experimenter rules out other
factors. A caused B.
That experiments make causality clear is a big advantage!
- another advantage is that they can rate out other factors
If a factor is relevant -> it makes a difference
manipulation: making the effect of a factor visible by making it different where nothing else
is, and seeing if the difference influences what the subject does.
note: there are various experimental designs.
The anatomy of an experiment
There is a skeletal structure that all experiments have; they only differ from how they put
flesh on the skeleton.
Independent and dependent variables
An experiment is defined by manipulation
manipulation: the experiment intervenes in nature to make something happen; that is ->
causing something.
independent variable: what the experimenter causes to vary
a variable must have at least two values otherwise its it’s not a variable
So there will be at least two values of the independent variable
- often referred to as different experimental conditions or experimental treatments
(because they specify how subjects are treated under various conditions)
The independent variable is the variable that is the intervention - what we make happen
The dependent variable is the data - what we observe (the effect)
, Control variables
An experiment, we say, asks how the independent variable affects a dependent variable.
But the world isn’t perfectly made for experimenters. There are a lot of other influences
that also affect the subject.
So in real world research our subjects behavior may be affected by our independent
variable, but our subject will also have to do with other influences.
To best see what effect or interdependent variable has, we want to get rid of these other
influences if we can. If we can’t do that, we want to equalize them, so that they affect all
experimental conditions in the same way and then drop out of our comparison between
them.
(Om zo goed mogelijk te zien welk effect of welke onderling afhankelijke variabele heeft,
willen we deze andere invloeden zoveel mogelijk kwijtraken. Als we dat niet kunnen, willen
we ze gelijk maken, zodat ze alle experimentele omstandigheden op dezelfde manier
beïnvloeden en dan uit onze onderlinge vergelijking verdwijnen.)
- deze hinderlijke variabelen heten control variables.
2 manieren om control variables constant te maken;
1. Control variables (variabelen die ook het object beïnvloeden naast de independent
variable) can be held constant directly, by the way we set up the sitaton.
2. Other control variables cannot be so directly controlled, we can not make all our
objects equally altruistic, or alienated, or distracted. What we can do is try to make all
groups of subjects equally altruïstic on average. ANd one way to do that is to assign
each subject strictly at random the ‘alone or to the ‘other onlookers condition. Then,
if we have enough subjects, the randomization procedure is likely to assign roughly
as many altruistic subjects to one group as to the other.
The larger the groups, the more unlikely they are to be very different on any variable.
The fundamental principle of experimental control: we control what we can; and what we
don’t control we randomize.
notice; what we controle, what we mainplate and what we measure, depends on the
question we ask. (An independent variable in one experiment may be a control variable in
another experiment).
There is a popular superstition that every experiment needs a control condition or control
group. Not so. But it is true that every experiment involves a comparison between something
and something else. What do we compare with that? It depends on the question we are
asking.
Any experiment will incorporate a comparison. There will be an independent variable
that takes on different values for different experimental conditions or experimental
treatments. And we will see how our subjects say or do depending on the value of the
independent variable. SO we measure what they say or do as the dependent variable.
