Standard Error - YouTube

Channel: Bozeman Science

[4]
hi it's mr. Andersen and in this video
[6]
I'm going to talk about standard air
[8]
standard air is one of my favorite
[10]
statistics and the reason why is that it
[12]
shows you how good your data is and let
[16]
me show you what I'm talking about so
[17]
let's say I collect two data sets so
[19]
this is data set 1 and this is data set
[21]
2 if you were to look at the data you
[23]
would say well this one over here seems
[25]
to show more precision it's going to
[27]
have more data and it doesn't seem to be
[29]
spread out as much but if I were to
[31]
graph this data so of these two data
[33]
sets if I were simply to create a graph
[36]
showing the mean or the average of the
[38]
data both of them would be 12 and so if
[40]
I had a bar graph it would look like
[42]
this for both data set 1 and data set 2
[45]
and so you should never have a bar graph
[47]
if you're using it to represent a lot of
[48]
data unless it has error bars on it and
[51]
so how do you figure out what the
[53]
standard error is how do you make the
[55]
standard error bars well let me show you
[57]
so basically I'll work through data set
[60]
two and then I'll ask you to work
[62]
through data set one I'll put the answer
[64]
down in the description down below and
[65]
so to figure out standard error the
[68]
equation is really ingenious basically
[71]
all they do is they take the standard
[73]
deviation and we divide it by the square
[75]
root of the sample size why is that
[77]
important well what a standard deviation
[79]
measure standard deviation is going to
[81]
measure the spread or how far the data
[84]
is spread out and so if you have a lot
[86]
of spread then you're going to have a
[88]
large standard error and the other thing
[90]
it uses is the sample size so the sample
[92]
size is how much data are you collecting
[94]
if you collect a lot of data and we're
[97]
going to have a large number here then
[99]
that's going to decrease my standard air
[101]
likewise if I have not much data at all
[103]
that's going to increase my standard
[105]
error so let's get to standard deviation
[107]
if you don't know how to calculate
[109]
standard deviation or even what it is
[111]
and how it measures spread you're going
[113]
to want to go watch a video and so
[115]
here's a little link I'll put to a video
[117]
I made on standard deviation so you
[119]
could simply click here or I'll put a
[121]
link to the standard deviation in the
[123]
video description down below but you
[125]
want to figure out how to do standard
[127]
deviation I show you how to do it by
[129]
hand and then I'll show you how to use a
[131]
spreadsheet to calculate standard
[132]
deviation
[133]
very very quickly and then what is
[135]
sample size well sample size is going to
[137]
be the number of trials that you do or
[140]
the number of data points that you
[141]
collect and so if we were to look at
[143]
data set two there are 10 numbers over
[146]
here and so that means I'm going to have
[147]
a sample size of 10 and so keeping those
[151]
in mind let me switch over to the
[152]
spreadsheet and show you how I would
[154]
figure this out so right here I'm in
[156]
Excel this is going to be data set 2
[158]
again I'm going to show you how to do
[160]
that one and then here's data set 1 so
[162]
what I'm going to do is since I've got
[163]
this column right here I'm going to
[164]
first of all calculate the average of
[167]
that or the mean of that if you were to
[169]
type equals MEA n for the mean nothing
[172]
will show up and that's because
[174]
spreadsheets don't use the word mean
[176]
they're going to use the word average so
[178]
you can see that that pops up I'm going
[179]
to put a left first I hit an equal sign
[182]
then I'm going to write an average and
[183]
then parenthesis and now I'm going to
[185]
tell Excel which kind of a data I want
[188]
to figure out the average from I put it
[190]
in parenthesis and it simply calculates
[191]
for me that it's 12 now here's a quick
[194]
little trick so I've done the average
[196]
for this column if I select this average
[199]
and then I move my cursor down here
[201]
watch how it changes so it changes from
[203]
that to that and so I can click it now
[205]
and then just drag that over here and
[207]
Excel is smart enough that it's going to
[210]
figure out oh if you want the average of
[211]
this now you want the average of this
[213]
column so you can see that the average
[214]
is going to be the same now let me show
[216]
you how to do the standard deviation if
[218]
you did that by hand it takes forever
[220]
and so it's really important that you
[221]
use either a calculator or a spreadsheet
[223]
to do this I'm going to hit equals and
[225]
then I'm going to start typing stdev
[227]
it's going to be standard deviation I'm
[230]
going to hit a parenthesis then I'm
[231]
going to select the data and then I'm
[233]
going to hit end parenthesis and there's
[236]
my standard deviation now if we were to
[238]
compare that to the standard deviation
[240]
of data set 1 so again let me do that
[241]
trick I'm going to select this move this
[243]
over here you can see the data set 1 has
[245]
a much larger standard deviation that
[248]
means it's going to have a greater
[249]
spread in the data but I'm going to
[251]
figure out the standard error this first
[252]
one so this is 0.8 to 5 and so how do
[256]
I'm going to remember these numbers here
[258]
and then switch back and figure it out
[260]
so let me write down the things that I
[262]
need to know so my X bar or my mean I
[265]
said is going to
[266]
12 the N or the sample size we set is
[270]
going to be 10 in data set 2 because I
[272]
have 10 values over here on the right
[275]
and then the standard deviation that I
[278]
just calculated is going to be 0.8 to 6
[282]
and I did that in Excel so how do I
[284]
figure out the standard DVR standard
[286]
error
[287]
so the standard error of the mean is
[289]
simply going to be the standard
[291]
deviation which is 0.8 to 6 and then I'm
[295]
going to divide that by the square root
[296]
of the sample size and the square root
[298]
of the sample size is going to be 10 and
[300]
so this is equal to 0.8 to 6 divided by
[304]
the square root of 10 I'm going to need
[306]
my calculator right here and that is
[310]
going to be 3.6 and then I have to
[313]
figure out so let me write that down
[314]
so that's three point one six and then
[318]
if I take point eight to six divided by
[322]
three point one six I get point two six
[326]
one so 0.26 one is my standard air or
[332]
standard air of the meet so you can
[335]
figure out the standard error of the
[337]
data set one remember in data set one we
[342]
simply had two values 9 and 15 and so if
[346]
you want to figure out the standard
[347]
error of those and so you make sure
[349]
you're doing that right you could do
[350]
that and I'll put the values down below
[352]
but what does this all mean well let me
[354]
show you why this is super important so
[356]
if we go over here and look at our graph
[357]
again so I just figured out the standard
[361]
error of data set 2 and so what that
[364]
means is data set to my mean is 12 and
[367]
so the average is 12 but I could add an
[370]
error bar on the top like that and then
[374]
I could add an error on the bottom and
[377]
my standard error is going to be we said
[381]
point 2 6 1 in other words it goes from
[386]
12 up to 12 point to 6 1 and then from
[390]
12 down to 12 minus 0.26 1 so we're
[393]
going to have an error bar on each side
[394]
what you can do is go figure out set 1
[397]
set 1 you're going to
[399]
much larger standard deviation you're
[401]
going to have a smaller sample size and
[402]
so I would imagine you're going to have
[404]
pretty big error bars and so what's neat
[407]
about that is the bar graph now with the
[410]
standard error bars is going to show me
[412]
um kind of the story it's going to tell
[415]
me about the data behind that mean or
[417]
the data behind the average and so
[419]
that's standard errors it's pretty
[421]
straightforward it's simple you're
[422]
probably going to need a spreadsheet to
[423]
do it and I hope that was helpful