WEBVTT Kind: captions; language: en-us
NOTE
Treffsikkerhet: 89% (H?Y)
00:00:00.000 --> 00:00:06.300
In this video, we're going to talk about the distinction between dependent and independent
00:00:06.300 --> 00:00:13.500
variables. This is a very important distinction in research design, but is also very important in
00:00:13.500 --> 00:00:21.000
statistics in order to properly set up the analyses. So it is useful to understand what these terms
00:00:21.000 --> 00:00:28.000
actually mean. Let's go back to the fundamental of setting up a research question, which as we have
00:00:28.000 --> 00:00:30.150
said, always must concern a relationship
NOTE
Treffsikkerhet: 88% (H?Y)
00:00:30.150 --> 00:00:37.700
among variables. This means that in order to ask a question that can be answered by research. We
00:00:37.700 --> 00:00:46.100
have to narrow the question down and precisely define everything about measuring the variables that
00:00:46.100 --> 00:00:53.400
go into this question. So everything has to be cast in terms of variables, their measurement, the
00:00:53.400 --> 00:01:00.349
context, and the conditions under, which they are measured. In general we usually are
NOTE
Treffsikkerhet: 91% (H?Y)
00:01:00.349 --> 00:01:08.500
interested in what effects some Target. Some outcome. This target variable is therefore called a
00:01:08.500 --> 00:01:16.600
dependent variable because by design, by the nature of our question, we are interested in what it
00:01:16.600 --> 00:01:25.050
depends on or whether it depends on something. This leaves out the things that it might depend on
00:01:25.050 --> 00:01:30.449
and therefore all the other variables that are used essentially as predictors for it
NOTE
Treffsikkerhet: 91% (H?Y)
00:01:30.449 --> 00:01:38.400
are called independent variables. So this is just designed terminology. By calling a variable
00:01:38.400 --> 00:01:44.600
independent we're not really saying that it's independent of anything else. It's one more poor
00:01:44.600 --> 00:01:47.900
choice of words in statistical terminology.
NOTE
Treffsikkerhet: 91% (H?Y)
00:01:47.900 --> 00:01:55.400
Independent variable means we use it to predict another variable. Dependent variable means this is
00:01:55.400 --> 00:02:01.900
our outcome of interest and the variable that we are assuming or asking whether it is affected by
00:02:01.900 --> 00:02:11.700
others. Let's go through some examples to understand this better. So, one silly example, is to
00:02:11.700 --> 00:02:18.300
ask the question, does coffee help studying, but this is a very vaguely phrase question.
NOTE
Treffsikkerhet: 83% (H?Y)
00:02:18.300 --> 00:02:25.000
In fact, it's not possible to answer this question because it's not well specified. In order to be
00:02:25.000 --> 00:02:30.200
able to answer it we have to define specific variables that we can then measure in. There are many
00:02:30.200 --> 00:02:35.500
many options will not going to go through many of them, but we'll just see, two completely different
00:02:35.500 --> 00:02:44.900
examples. One possibility is to define one variable as the number of espresso cups per day and
00:02:44.900 --> 00:02:48.050
another variable as the number of hours spent reading per day.
NOTE
Treffsikkerhet: 72% (MEDIUM)
00:02:48.050 --> 00:02:48.900
NOTE
Treffsikkerhet: 89% (H?Y)
00:02:48.900 --> 00:02:56.500
So essentially this is like asking does how many espresso a day, does how many espresso we
00:02:56.500 --> 00:03:03.500
drink in a day affect how many hours we study in a day? That's what we're asking.
NOTE
Treffsikkerhet: 91% (H?Y)
00:03:03.500 --> 00:03:09.200
And if we put it this way, then this means that
NOTE
Treffsikkerhet: 91% (H?Y)
00:03:09.200 --> 00:03:15.200
the number of espresso cups would be our independent variable and the number of hours spent reading
00:03:15.200 --> 00:03:17.650
would be our dependent variable.
NOTE
Treffsikkerhet: 88% (H?Y)
00:03:17.650 --> 00:03:25.300
Now, you might object that, this isn't what you really interested in when you are asking, whether
00:03:25.300 --> 00:03:30.400
coffee helps. But that just means that you're going to have to define your variables differently.
00:03:30.400 --> 00:03:37.200
And before we go on to another definition. Let me remind you that this is still way too vague.
00:03:37.200 --> 00:03:44.200
We have to specify whom we ask what they do, what else they might be doing if they drink any other
00:03:44.200 --> 00:03:47.750
things. When we measure these, how often
NOTE
Treffsikkerhet: 91% (H?Y)
00:03:47.750 --> 00:03:52.600
exactly what the observations or questions are and so on.
NOTE
Treffsikkerhet: 91% (H?Y)
00:03:52.600 --> 00:03:58.900
So this is just the first step in defining variables so that they can be designated as dependent or
00:03:58.900 --> 00:04:07.550
independent. A different approach to the same general question might be to define the variable as a
00:04:07.550 --> 00:04:15.399
the property of drinking coffee versus drinking soft drinks, like Coke, or bruse or what else some
00:04:15.399 --> 00:04:17.200
people might be drinking.
NOTE
Treffsikkerhet: 91% (H?Y)
00:04:17.200 --> 00:04:25.000
And another variable as the average grade in semester courses. So this might be more interesting for
00:04:25.000 --> 00:04:33.000
some. These are very different variables. The first variable is actually a grouping variable. It's a
00:04:33.000 --> 00:04:38.900
person who drink coffee versus persons, who don't drink coffee, but drink is another category of
00:04:38.900 --> 00:04:44.299
drinks and that would be our independent variable, the group,
NOTE
Treffsikkerhet: 85% (H?Y)
00:04:44.299 --> 00:04:51.500
and the dependent variable would be a grade. So asking if there are differences in grades between
00:04:51.500 --> 00:04:59.000
these two groups of people and we still need to define a whole bunch of other details before you can
00:04:59.000 --> 00:05:05.200
actually produce an answer to this question by running the measurements in the analyses. So you see
00:05:05.200 --> 00:05:06.300
that
NOTE
Treffsikkerhet: 87% (H?Y)
00:05:06.300 --> 00:05:12.800
the general questions does coffee help studying is very far from being answerable. The many
00:05:12.800 --> 00:05:19.450
different ways to make it specific enough. And in the process, we have to define specific variables
00:05:19.450 --> 00:05:25.100
and one of those variables is going to be our target and that's our dependent variable that we want
00:05:25.100 --> 00:05:31.000
to study if it's affected and the other variable, the predictor, is called the independent one.
NOTE
Treffsikkerhet: 91% (H?Y)
00:05:31.000 --> 00:05:34.200
Let's look at some more examples.
NOTE
Treffsikkerhet: 91% (H?Y)
00:05:35.000 --> 00:05:44.700
So, one question might be does music affect aggression. This is very vague. Of course, we candDefine
00:05:44.700 --> 00:05:50.400
it in many ways, for different groups and for different purposes. So, for this illustration, I'm
00:05:50.400 --> 00:05:58.400
just keeping it very simple. And let's say that one way to define it would be by groups such that
00:05:58.400 --> 00:06:04.400
people in one group are asked to listen to Kpop all day and people in other group, listen to
00:06:04.400 --> 00:06:05.800
NRK Jazz all day.
NOTE
Treffsikkerhet: 87% (H?Y)
00:06:05.800 --> 00:06:12.550
And other people are asked to listen to death metal music all day, and they do that for some time,
00:06:12.550 --> 00:06:19.000
which is part of what needs to be defined. And then we ask them to fill out an aggression
00:06:19.000 --> 00:06:20.350
questionnaire.
NOTE
Treffsikkerhet: 90% (H?Y)
00:06:20.350 --> 00:06:27.000
And in this case, the group which kind of music, they are asked to listen to the group would be our
00:06:27.000 --> 00:06:32.500
independent variable. And the score on this aggression questionnaire would be our dependent
00:06:32.500 --> 00:06:38.600
variable. And, of course, we would have to define a whole bunch of other things.
NOTE
Treffsikkerhet: 91% (H?Y)
00:06:38.600 --> 00:06:46.900
Another completely different conceptualization of this question would be to note, how long each
00:06:46.900 --> 00:06:54.000
person listens to music per day and we have to define for how many days you
00:06:54.000 --> 00:07:01.100
want to do this for one day for a week or how long or every Monday for a year, whatever we decide.
00:07:01.100 --> 00:07:08.150
And we also at some point need to define some measure of aggression. So
NOTE
Treffsikkerhet: 73% (MEDIUM)
00:07:08.150 --> 00:07:16.200
another variable might be an aggression assessment checklist in which an observer checks some
00:07:16.200 --> 00:07:22.800
items. And, for example, this can be used with special populations or with children and so on, and
00:07:22.800 --> 00:07:30.100
we need to define the all the conditions for it. In this case number of our listening to music
00:07:30.100 --> 00:07:36.700
would be our independent variable and the score on the aggression checklist will be our dependent
00:07:36.700 --> 00:07:38.300
variable.
NOTE
Treffsikkerhet: 91% (H?Y)
00:07:38.900 --> 00:07:42.100
Different examples.
NOTE
Treffsikkerhet: 80% (H?Y)
00:07:43.000 --> 00:07:52.200
A very interesting example in our context is does intervention for something work; for reading for
00:07:52.200 --> 00:08:00.300
emotional difficulties, for behavior problems, for language, for whatever, for X does intervention for
00:08:00.300 --> 00:08:08.700
something work. We need to define variables. One possibility is to have an intervention group and a
00:08:08.700 --> 00:08:13.650
control group. So the intervention group receives intervention the control group
NOTE
Treffsikkerhet: 91% (H?Y)
00:08:13.650 --> 00:08:24.300
One way to define these is to have an intervention and a control group. The intervention group
00:08:24.300 --> 00:08:30.000
receives the intervention and the control group is engaged in something else.
NOTE
Treffsikkerhet: 80% (H?Y)
00:08:30.000 --> 00:08:39.299
And after the intervention, after some time of doing this. Then both groups are assessed with a test
00:08:39.299 --> 00:08:45.600
and we measure their performance. So, here the independent variable is group in the dependent
00:08:45.600 --> 00:08:54.600
variable is the score with the performance on this assessment test. That is a valid instrument for
00:08:54.600 --> 00:09:00.350
measuring X, whatever it is. We want to improve with the
NOTE
Treffsikkerhet: 79% (H?Y)
00:09:00.350 --> 00:09:05.750
And, of course, there's lots of other details that need to be defined in our research design.
00:09:05.750 --> 00:09:14.400
Another approach to this question, is to define a different variable, the variable of time. So, that
00:09:14.400 --> 00:09:20.650
time would have the value before intervention and the value after intervention.
NOTE
Treffsikkerhet: 91% (H?Y)
00:09:20.650 --> 00:09:29.400
And another possibility for the outcome here would be an assessment checklist. So for some things,
00:09:29.400 --> 00:09:38.650
it may be preferable or only possible to assess the x that we're trying to improve by observation,
00:09:38.650 --> 00:09:46.200
in which case you have to use the checklist. So we define our outcome accordingly and this means
00:09:46.200 --> 00:09:51.250
that we basically assess the participants twice.
NOTE
Treffsikkerhet: 79% (H?Y)
00:09:51.250 --> 00:09:59.550
Once before the intervention and once after the intervention. So the independent variable is time
00:09:59.550 --> 00:10:06.400
and the dependent variable is the score on this assessment checklist.
NOTE
Treffsikkerhet: 88% (H?Y)
00:10:06.400 --> 00:10:14.700
So you see two different design approaches. One is to have an independent variable of group. The
00:10:14.700 --> 00:10:20.900
other is to have the independent variable of time. And in real research want to do this properly. We
00:10:20.900 --> 00:10:26.700
actually use both independent variables. So we have two groups and we says both groups before, and
00:10:26.700 --> 00:10:32.500
after, for reasons that you're going to learn in the research design part of the course.
NOTE
Treffsikkerhet: 86% (H?Y)
00:10:32.800 --> 00:10:42.100
And some final examples. We can ask for, for example, are girls better than boys in math?
NOTE
Treffsikkerhet: 90% (H?Y)
00:10:42.600 --> 00:10:49.300
Of course, this needs to be specified like in what age, what math are we talking about and so on.
00:10:49.300 --> 00:10:55.500
But anyway, we can define the variable of sex. This is obviously our independent variable and the
00:10:55.500 --> 00:11:01.750
dependent variable, could be performance on some math assessment test. For example, the state test,
00:11:01.750 --> 00:11:06.950
the national assessment on math performance in a given age.
NOTE
Treffsikkerhet: 91% (H?Y)
00:11:06.950 --> 00:11:17.100
Or we can ask something like, are children with dyslexia better in art things. So, here we would
00:11:17.100 --> 00:11:24.300
have a group variable distinguishing children diagnosed with dyslexia from children, typically
00:11:24.300 --> 00:11:26.849
developing in their reading skills.
NOTE
Treffsikkerhet: 86% (H?Y)
00:11:26.849 --> 00:11:33.100
And the two groups would probably have to be matched in some other things. So they should be as
00:11:33.100 --> 00:11:40.450
similar as possible except for the dyslexia diagnosis. And that would be our independent variable,
00:11:40.450 --> 00:11:46.950
the group or the diagnostic category. That's our independent variable and the dependent variable
00:11:46.950 --> 00:11:53.100
would be the score on some instrument for assessing artistic skills.
NOTE
Treffsikkerhet: 91% (H?Y)
00:11:54.300 --> 00:11:57.750
And one final example.
NOTE
Treffsikkerhet: 91% (H?Y)
00:11:57.750 --> 00:12:08.000
We could ask do minority language speakers have for Norwegian vocabulary. So, in that case, again,
00:12:08.000 --> 00:12:15.100
our independent variable is a grouping variables. It's a group. We would have children with the
00:12:15.100 --> 00:12:20.600
minority home language, in one group and otherwise comparable children with the majority home
00:12:20.600 --> 00:12:26.100
language. So they would have to be matched in other things, you know, from the same neighborhoods
00:12:26.100 --> 00:12:28.300
and circumstances and so on.
NOTE
Treffsikkerhet: 86% (H?Y)
00:12:28.300 --> 00:12:35.700
That would be our independent variable and our dependent variable would be some score on
00:12:35.700 --> 00:12:39.400
appropriate vocabulary test for the age.
NOTE
Treffsikkerhet: 91% (H?Y)
00:12:39.400 --> 00:12:47.500
So I hope this concept is now clear and please go through more examples, yourselves. You connect the
00:12:47.500 --> 00:12:53.300
nature of a research question to the concept of what is the dependent and what is the independent
00:12:53.300 --> 00:12:54.900
variable.