SPSS for newbies how to get mean; median; mode; standard deviation etc

hey Phil from Statistics mental calm so
maybe with the help of my video you
spent hours and hours entering data your
survey data and SPSS and now finally to
ready to analyze it where to begin well
what you could begin with is to do plots
I graphs and also report the descriptive
stats at this stage many students or
think ah okay I need to report just the
mean and standard deviation right the
answer is not quite so I've got three
bullet points here that I'm going to
look at today basically think of them as
brother-sister pairs you've got a report
both of them together you use a report
for your variable mean and standard
deviation median is a quartile range
mode and range what these things are and
how you do the maths pisser yes we're
going to look at now an example okay so
here we have three variables you can't
be odd V nom now which pair we report in
terms of measuring for the central
tendency and dispersion depends on our
variable type okay so remember that
measure of central tendency you just
kind of measure is a one number tells
you what the typical value is of your of
your variable and the options we have
our mean median mode for dispersion that
gives us an idea of how spread out your
data is and the options we've got here
today we're going to be standard
deviation
our interquartile range and range got
three variables and five observations
for each okay first column V count for
continuous continuous variable what do
we report we report the mean and
standard deviation there is a remark
about what I just said which I leave to
the end of the video if you could just
hang on to the end of the video do so
okay v ordinal the odd for ordinal
remember an ordinal variable is a
categorical variable where these numbers
here have a ranking in other words the
ranking is what gives an imaginary so
two is lower than one is lower than two
for example
okay for all the verbal categorical
variables we reports the median and
alongside with the interquartile range
and again there is a remark here I'm
gonna leave to the end of the video
finally we have the normal one this is
perhaps the easiest one Callisto remark
for this nominal categorical variables
that is one where the numbers here
they're just labels mean to say that one
is loaded to the one is looking to that
is not consequence between nominal
variables because these are just labels
a college left and right okay
with nominal variables we report the
father for the mode for the central
tendency and the range for dispersion
there's many ways now on how to get
these measures only to show you I'll
show you very much many of them okay
I'll do start with the most often used
methods first so continuous we want the
mean and standard deviation basically
everything isn't that we want is under
analyze button so analyze descriptive
studies okay there shouldn't be anything
in here but if there is press press
reset we take the continuous variable
over into the variables boss and then we
click on to statistics
notice here we've got some options we've
got four tables central tendency
percentiles values dispersion
distribution we want first from the
central tendency because the enter
there's got mean median mode and some
because they're the options we have for
measures to call value we want mean
dispersion we've always said that the
standard deviation goes with the mean so
let's click that and then okay and then
uncheck display frequency tables just
because it will dump out if you don't do
that we'll dump out another table which
is not very useful and then press ok and
so we'll see here that we've got a mean
of three hundred forty three point six
understand deviation of 160 21 6 96
those numbers well if you're if the
variable has units for example for
measuring these things as say kilograms
these don't forget to report the units
kilograms beside both of these counts
good
next or no we want the median and the
interquartile range in the woman know
what interquartile range means so
analyze descriptive statistics or reset
it and then we'll go to ordinal fiord
take it over click on statistics central
tendency we want the median the
interquartile range is you don't find a
dispersions that go with the percentile
values I will click on quartiles and
then continue if I press ok let's just
kind of get a bit more technical don't
scare you but it's not that technical
it's probably you know already remember
that to find the median of ungroup data
the first list the numbers in ascending
order so here this variable I've got 5
values 1 2 3 4 5 BC so sample size is 5
so we'll list them in ascending order
and then we'll pick the middle one on
the fifth one high the third value so
think about it what value do you think
you'd get for the meeting
okay I think you should hopefully you
should get three in your head why I said
region these large in ascending order
well one is a smallest second a lot
slices - so come next to I drag down 1
comma 2 comma and next one is 3 comma X
1 is 4 comma next one is 55 what a third
value so it just so happens that the Y
values 1 to 5 here 1/3 value after that
list that's a three just three should be
the median if you didn't get that since
this is implied video don't worry about
okay uncheck the display frequencies
table press ok so here you go median yes
indeed it's 3 percentile 25th percentile
otherwise known as first percentile on
behalf second percentile of the median
is 3 third percentile center-field
percentile is one half okay I said even
though if we don't understand well that
we need to be reported that's gonna be
fine
finally nominal so we want for this the
mode and the range now get a quick recap
mode is the most frequently occurring
reservation so what are the observations
here I've got one two one appears twice
in my list two appears three times in my
list so the mode must be yep two range
okay range isn't the way used in a
natural language of being over would say
the ranges between one and two you know
that's what students often write in
their kind of homework or exams in
statistics range means it's got this got
an equation range is equal to a maximum
minus the minimum value so here the
maximum is two and the minimum is 1 so
the range is equal to 1 okay
so it's actually the range is a single
number it's not you don't see it's not
like you can't say oh it's one to two so
let's get that analyze descriptive
statistics case over to the variables
box statistics central tendency mode
dispersion range and we're looking for
the answer of two for mode one for range
okay and check that well let's keep it
checked so I can just show you that why
we should then that should convince you
why shouldn't check it all right so here
you go
mode is two and the range is one great
now the frequency box basically just
tells you a list of a number of things
that happen a number of currents or
things for the value so here you got one
appearing twice two appearing as three
terms okay and it gives you the
percentages and stuff like that which is
fine why I said in general is best not
to do
is that if you've got a big long list of
many very numbers in the variables then
this table could be long you don't want
to do this
hmm there's a there's a kind of case for
doing it when you're categorical
variables but not when it's continuous
ie for the first variable you wouldn't
do it just because then this table be
very would be as long as probably as
long as the number of observations you
have I say probably if you have yeah
okay good
they're the basic kind of things so now
we know that at least we know that it's
not a case of simply reporting the mean
and median or more on the report
descriptive stats because it depends on
the type of variable we have okay and
the type of variable have could be
different for each column that data
right I did say earlier on at the
beginning and I hope you still with me
that and if you are well done that I had
remarks about this continuous and
ordinal continuous I said that we use
the mean and median observed mean and
some deviation generally that's true
unless unless you reckon or you think or
there is outliers in the data because if
there are outliers in the data you'd be
better off reporting the median and the
interquartile range just because the
mean wouldn't be a great representation
of the typical value if you've got
outline in your data set of course then
you have but could have the option of
removing me outlier if you found one or
more outliers and then you can report
the mean and median as usual and mean
and send the Ovation as usual okay so I
just pointed out something there that
where you might see you could report the
median as opposed to the mean even
though it's continuous
next ordinal my previous videos we've
discussed this ordinal variable in more
detail just to recap ordinal is a bit
hazardous character it's bit like
Jacqueline height so sometimes it's the
treaty as nominal sometimes it can be
treated as continuous and sometimes it's
just known right but it's usually
sometimes it can be nominal sometimes
you continues whether we treat it as
continuous depends on the type of
analysis you want to do with it so for
example this ordinal could be a Likert
scale from a scale of say one two seven
one two five one two ten even right for
a certain analysis you might want to
treat that scale is there being ordinal
you might want to treat it as continuous
for example you might want to do a
regression and you want it to be treated
as a scale i continuous in that case you
would since you're going to treat it as
continuous you report the mean along
with the standard deviation all right so
that's it for those two remarks another
remark I said that there's more than one
way of getting these numbers out for
some of them okay well let's just show
you
we'll go analyze descriptive statistics
button but these frequent this
descriptive and the frequency button
they're kind of similar but the
descriptives is like subsets of the
smaller brother of the frequencies
because you can get less out of it so if
you press on descriptives you can see
under the options box if we click on it
that you've got mean and then you've got
measures of dispersion all right
so you might see in your lectures we
might read in some textbooks there use
this for continuous so this is only
people continuous variable
next mmm we might want to we might have
like you have done an experiment and
you've got two measures to scores right
and you want to calculate the average
score that the person scores in the test
so why did I just do something off the
cuff here just listen to my music or
listen to something while I do this and
we just change this
okay forget about these two I've got I
know this is gonna be useful for some of
you especially if you have to do in
psychology or you've created
questionnaire social science where
you're measuring somebody's kind of
attitude on something different facets
of it so go be score be score - they're
measuring something right and I want to
take the average of these two so in this
case I don't want to just take the
average of this guy by itself
and some deviation of this guy by itself
like I said not the sky by itself so
what to do well if I just want the mean
was to waste let's just say it's two
ways first of all so let's do them both
say I want the mean and the standard
deviation because I can let's just say
the continuous right so continuous is
cause I'm just going to add them
together I can create a new variable
first this is first way so go to
transform compute variable and I clear
my box and I'm gonna say score this
being the aggregate score right it's
going to be made up of V score 1 plus V
score 2 and then I go okay and then what
you should see is a new con score and
it's just as we've just created it 2
plus for the first person 6 4 plus 2 6
okay and now we can go ahead just like
before it's like a new variable forget
about these two now to compute the mean
and standard deviation for this score so
analyze descriptive statistics this time
score over there options mean clicks
already dispersion checked already
addressing you definitely notice it is
okay
to score I have a mean of 6.8 send
deviation of three point three five say
to personal places that's method 1
method 2
I'm going to show you how to use it now
[Music]
using calculator so compute again
compute variable oops sorry about that
okay I want to do something different
what I've done here is computed the
aggregate score for each person and then
on based on the new score calculated the
mean of that less than deviation of that
but say I wanted to calculate the
average score for each person so this is
person 1 his this person's got to on the
first test for the second test so the
average of that or the two tests would
be 4 plus 2 divided by 2 so that be 3
okay how would I get the average scores
per person I would do transform compute
variable that shifts curve score
[Music]
score score meaning solid and then what
we do is statistical no double click on
mean and then this box here that tells
you more about how to use this mean
function but over here you see the
question marks is two question marks I
enter the mean of enter the two
variables of which I want the means I
want V score and the other question my
confi score
- I'm gonna press it okay a new column
comes up score me and can you see how
it's different so I calculate the mean
of these two guys for the first person
score two of score for is mean of three
for the final person it's just scored
zero on this first test two on the
second testers that's got to be an
average of one and C so this is
different to doing this
the thing I've just done there computing
the average like this if you have more
two scores you can definitely there's
nothing stopping you doing comma and
then putting in more variables you score
three comma score for right just because
it comes out with two question marks
doesn't mean that it just works with two
two variables also we didn't have to
under the function group we didn't have
to click statistical and then look for
the mean you could have typed in by hand
here straight away mean score you know V
and then the variables and it's not
case-sensitive so we didn't have the
right dude capitals like he's showing if
one standard deviation as well for each
person to score you could do that that
would be SD and then that'd be
[Music]
Scott dan I want to score SD and say
that's close he got standard deviation
so here you got pairs of mean and
standard deviations from each person's
score all right okay well there's a pile
of stuff there and so I hope just work
for it and find some that's useful for
you there more discussion anything like
that you can always get hold on me on
statistics mentor calm okay cheers guys