馃攳
How to Calculate and Interpret R Square in SPSS; Regression; Correlation - YouTube
Channel: unknown
[0]
In this video I want to show you how to
[2]
calculate R squared.
[4]
Now here I have an example with three
[6]
variables. If I'm trying to find R squared
[9]
for just two variables, one way we can go
[11]
about doing that is just running a
[13]
correlation. Let me show you what I mean
[14]
here. So if I go to analyze and then
[17]
correlate and then bivariate let's say
[21]
we want R square between SAT and
[24]
college GPA, so I'll move those two over and
[28]
then I'll click OK. Now this gives me not
[31]
R squared but it gives me r so the
[34]
correlation between SAT and college GPA
[36]
is .65 and that is in fact
[40]
significant at the .01 level. Now
[43]
that's r; so if I want R squared what I
[46]
can do is just simply square that. So .65
[50]
and then squared here is point .4225 so R
[57]
squared here is .42. Now
[61]
another way to do this is that if I have
[64]
more than two variables that I'm working
[67]
with, or if I just don't feel like
[69]
calculating R squared manually by
[71]
squaring r, what I can do is I can go
[74]
to analyze and then regression and
[77]
select linear. Now I have to decide here
[81]
what my dependent variable is, or what it
[84]
is that I'm trying to predict. Now SAT was
[87]
measured in high school and college GPA,
[89]
as the variable sounds, is GPA in college
[91]
during the first year, after one year
[94]
college. So it makes sense that SAT
[97]
would predict college GPA. So we'll put
[100]
the thing we're trying to predict in the
[102]
dependent box and then we'll put SAT and
[106]
the independent. OK let's go ahead and
click OK. Now
[110]
if you recall from our earlier analysis, when
[112]
we squared that correlation we got
[114]
.4225.
[116]
So R square was .4225. And here
[121]
to find R squared we want to go to the
[122]
Model Summary table and
[126]
here's r this is the correlation
[128]
.65, we saw that in our previous
[130]
analysis.
[131]
And then R squared is right next to r,
[135]
notice .422. And that's exactly
[139]
what we got before within rounding error.
[141]
So we can run regression to calculate R
[145]
squared. Now in case you're not familiar
[148]
with what R squared is, it indicates the
[151]
amount of variance in the dependent
[154]
variable that is accounted for or
[157]
explained by the independent variable. So
[161]
since we're using SAT here to predict
[163]
college GPA,
[166]
What that means is. if we know a person's
[168]
SAT score, we can account for I can
[172]
convert this to a percentage, about
[174]
42% of the variance in
[178]
college GPA, which is pretty good.
[181]
OK so that's what R squared means, it's a
[183]
measure of how much we explained in one
[187]
variable using one or more other
[190]
variables. And one last thing here, if you
[193]
want to calculate R squared and you have
[195]
more than two variables at once, then you
[198]
really need to use this regression
[199]
approach here to find that. So let's say
[202]
we want to use both SAT and social
[206]
support to predict college GPA and we're
[209]
doing this two try and get our R square.
So we're doing this to try and
[211]
see how well, overall, these two
[213]
predictors combined how much of college
[217]
GPA they account for or explain, which you
[221]
may recall is what R square really means.
[223]
How much of the variance did we account
[225]
for in a given variable using one or
[228]
more other variables. So go ahead and
[230]
click OK. And then here notice our R squared
[233]
increased, and it will when we add
[235]
another predictor in almost all
[237]
cases. And here our R squared is .511.
[241]
So using social support and SAT we
[246]
can account for about 51%
[248]
of the variance in college. So the
[253]
GPA in college after their first year.
[256]
OK that's it. Thanks for watching.
Most Recent Videos:
You can go back to the homepage right here: Homepage