Hi. Anderson Smith. In this section of the course,
we're going to talk about threats to internal validity.
Remember, internal validity is that confidence that we have
in the causal relationship between the independent variable and dependent variable,
whether the manipulation we're making is
really causing the changes we see in a measurement.
And that validity that
we really do have a difference is threatened by certain things that happen.
We're going to look at those threats in this section of the course.
First, we are going to talk about maturation in history,
the fact that people change and that change might be
observed in differences that are not related to the independent variable.
So, in the experimental method,
the ability to make conclusions about the effects really depend upon internal validity,
and there are situations in which internal validity is
threatened because many experiments include
several measurements over a period of time and people can
change in that period of time and the environment can change within that period of time,
and measurements might be due to these changes and not to
the independent variable that we're assuming is being manipulated.
For example, people change.
It's called maturation.
They change simply with time.
At age 70 I'm different than when I was age 50.
When I was age 50, I was different when I was age 20.
In fact, there's big differences in development between age seven and eight,
between 18 and 20.
So you might have changes that occur if I'm doing,
for example, in my ageing research,
a study of people how their memory changes between age 50 and age 70,
I might be looking at the fact that maturation changes are also occurring.
They're simply changing biologically.
And then there are things in history that can change the way in which we- so,
it can be a person changes,
they can be environmental changes.
The maturation changes, some of them can be very short term.
That is, the way we are as people can have a very short term change over time,
and that could greatly influence experiments
where we're using measuring behavior over time.
Things like mood, I get happy,
I'm sad, and I can change over an hour or a few minutes.
Fatigue, I might start out the experiments fully invigorated
in doing this experiment but with time I simply become tired of this.
And so it's going affect the measurements that I
make over early in the experiments versus late in the experiment.
Time of day, yes,
I can measure somebody in the morning and if there are
morning people they're going to be very different when I measure in late afternoon.
There are people who are late afternoon people
and they do better in late afternoon than if I measure in the morning.
So time of day has a huge effect on psychological processes.
Stress, I might start
out an experiment sort of stressed about what's going to happen to me,
but then stress is reduced.
Or stress might be increased in the experiment,
but it's not the independent variable.
It's the person that's changing my level of stress.
Or I can be bored,
this experiment is so boring.
I see how fast I can put these little pegs in these little holes.
I started out thinking that was interesting but then after an hour of doing that,
please, let it stop. I might be hungry.
When I start the experiments I'm not hungry,
when I finish the experiment I am hungry.
Those changes in these biological states might be
influencing the measurements that I'm taking in the experiment.
And there are many many others that I could list,
I could fill this page with many possible
person changes that could occur over the course of an experiment.
They can be longer term changes like for example,
aging that I mentioned earlier or educational attainment.
I might measure somebody at the beginning of their freshman year in college,
and has to be very different when I measure them
in their final year of college if in fact I'm measuring cognition or motivation.
There can be socioeconomic status changes.
There can be attrition,
people can drop out.
And so, when I'm studying conditions at the beginning of this experiment it
might be very different from studying people
at the end of an experiment because they're different people.
People have dropped out and they are the ones that
probably are very different from the ones that stayed in.
And like I said, there are many other things I could list,
I'm just giving you some examples of people changes
that occur on a short term or over the long term.
The longer the experiment over time,
the greater the probability that these maturation and people
changes are going to affect the experiment.
Now there are other things like learning effects,
the fact that as I go through
the experiment I'm learning how to do the tasks that are being
given to me and so there can be carry over effects from
things that happened early in experiment
and the things that happened late in the experiment,
having no relationship to the independent variable I'm actually manipulating.
Like understanding the format of the experiment.
If I know what the purpose of the experiment is,
I might learn more about that as I go along throughout the experiment.
Familiarity with the testing procedures and becoming less anxious about the experiment.
All these are changes that can occur simply because it is practice over time.
I might develop a strategy to do better or to do worse,
knowing what the expectations of the experiment are.
I might learn how to be better or how to be worse if that's what I'm expected to do.
All these are biases.
They are threats to the internal validity of the experiment and
some of these are simply habituation and it's as I go through the experiments,
I'm habituating to the experiment.
Habituation is seen in all organisms.
The first time you see a stimulus you become aroused by that stimulus,
but if I see this stimulus over and over and over again,
then my arousal is reduced because of the experience I have with that stimulus.
In addition to time between the observations,
other things happen during the course of the experiment and that
is for example, history. History happens.
So events outside the research study itself can alter research participants performance.
I might be sick today,
I wasn't sick yesterday that's going to greatly affect performance.
So what we have to do is we have to carefully select subjects
in an experiment so we minimize the history effects,
the fact that this person has
a different history from the other person in the experiment.
If I have a large sample or I'm very careful in my sampling procedures,
then that can reduce that particular threat.
So we want to try to make sure that everybody has that same historical experience,
both within the experiment and before they get to the experiment.
The threats to internal validity include maturation,
these biological processes that change due to simply the passage of time,
and history, events occurring during the experiment or before
the experiment that are not part of the treatment but
can cause threats to internal validity. Thank you.