One way ANOVA in SPSS (repeated measures) Video Lecture | SPSS: For Beginners - Data & Analytics

hey how you doing everyone this is still
from Statistics mentor.com in this video
we're looking at the 1 way within
subjects and over this is where we wish
to compare means within the levels of
one factor where each item being
measured experiences each level of the
factors the example I've got here is
time taken to drive around the racetrack
we've got 8 participants each
participant drive around the track using
cars 1 2 3 and 4 and the DV is the time
taken to go around the track and the
question is is there a mean difference
in the difference in the mean times
taken to go around the track between
these cars so basically speaking we want
to see if if there's a difference in the
performance of these four cars now how
is this different - one way between well
notice that each level of this factor
being type of car as column is calling
for each level we're asked
for the one way between underwear you've
got two columns one containing the
levels of the factor just label one two
or three four or five so on and one
containing the D V so go back to the
within you can see that the DVS separate
columns one for each levels of the
factor now the question is going to be
the same again the main null hypothesis
is we're testing the nulls that there is
no difference in the means of the DV
versus the alternative that there is and
should there be evidence of a difference
we can look at pairwise comparisons
tests now to set up the one-way within
we go to analyze we go to general linear
model not generalized linear model go to
general linear model and go to repeated
measures repeated measures is another
name for within and over within subject
name factor name we give that the name
which we desire here they've got four
types of Qasim is a car type number of
levels got four levels because the four
types of cars then we go at and then we
go define now notice that in this box
goes levels one two three four is these
numbers are say labeling the levels so
we've got four levels we highlight our
four levels of a factor and we transfer
over
and so those question mark now converted
into names of our levels at this stage
we could ask for pairwise comparisons
tests will go to options fact
and fact interactions we take card type
over its display means and we wish to
compare what they call compare the main
effects click on that here we're given
an option one for only LSD
let's go for bonferroni that's a popular
option continue and then it's ok now
there's a lot more output here than for
between an over this will multivariate
test we can ignore that for one way
between which what we need first is to
conduct this martially test of
sphericity the null hypothesis is that
this condition of veracity is satisfied
versus the alternative it isn't if you
want to know it more technically what
saying just read under here right to
mostly is w 0.57 to it has a
significance level cigarettes because i
p-value of point six seventy four so we
do not reject null
hence the condition is satisfied that
means we can continue to look at the
tester within effects and what we do the
source here is split into the one for
the factor card type and the other
watches error ignore the error will
connect the factor look at sphericity
assumed because that condition was
satisfied the null is we look at f here
the null is there's no difference in the
means of the levels of that factor i
know means in the time taken to around
racetrack each of the four cars do
reject a null yes since p value is less
than not point not five
so next we want to see where the
differences lie it's a wind contrast
ignore
we don't need that don't need that so
estimated marginal means cost one to
four you see the mean times here that
takes mean is lot time is longest for
type for 209 whatever seconds is and the
fastest is for type 377 you see that
looks like some difference is there
under started I was quite small as well
relative to the size and means
suggesting there is a difference in
means but we found formally that there
was difference so look at the pair wise
comparisons we asked for the bonferroni
I car type and Jake I type and got mean
difference I - J so mean one compared to
two that's the mean difference of going
from one to two
Carter / - so it looking at one seven
eight point six boxing minus one eight
nine so it's going to be negative figure
indeed it is look at the figure is it
significant yes like things there is a
difference between evidence of a
difference between the mean travel times
one car type one and two how about one
and three no p1 he'll at one and four
yes okay likewise two to one Willis same
as one to the other way around two to
three is a difference
yes p-value point not seven bigger not
four not five two to four yes the
difference and so on
so the basic level it's as simple as
that
do that first of all check martially
condition is that satisfied then go and
read off the F value testing if it is
the difference across the levels of the
of the factor and if there is evidence
of that we look at pairwise comparison
tests
let's go back at the beginning Marshall
is test for severe city suppose it's not
satisfied then what do we do well what
we can do is read off choose one of two
types of tests to find the F instead of
reading Christie assumed we may choose
to read this greenhouse skeezer or this
I won't even attempt to say HF test
anyway just to say that you know some
lecturers prefer just to go tell their
students just go for the this GG one why
is because this GG one is more
conservative than this HF one in other
words if you reject the gg you're going
to definitely reject for HF so it's a
more strict stricter some people like to
make a choice between the G G and the HF
by looking at the sizes of the epsilon
okay but if you want to be conservative
if your score for the Q G this stage you
know just look at what your lecturer
prefers you to do okay so that's a very
very simple look at the repeat at one
way within subjects ANOVA good have a go