nm0658: this is the [0.2] last of five sessions [0.5] which are [0.3] specially 
for your group [0.4] to look at the design analysis of experiments [1.0] er 
from next week [0.6] we [0.2] are going back to the general group [0.5] and [0.
2] we're going to look at advanced methods [0.4] which i think are equally 
applicable [0.5] for everybody [0.3] so we're then going to look at advanced 
modelling ideas [0.5] which i think are the same sort of models you need to 
consider [0.3] whether you're doing a survey or an experiment [0.9] the two 
main problems we will tackle there [0.3] will be [0.2] what happens if your 
design is unbalanced [1.1] and some experiments are unbalanced and most surveys 
[0.5] have unbalanced data [1.3] and the second thing we will tackle [0.2] is 
what happens [0.3] when your data may not come from a normal distribution [0.7] 
the traditional statistics [0.3] says that [0.3] if your data come from a 
normal distribution then everything is fine [0.4] and if they don't come from a 
normal distribution [0.4] then [0.2] you first panic [0.4] and then you 
transform your data and then you hope your panic is over [0.7] and there are 
modern methods now [0.4] where you can analyse data from a non-normal 
distribution [0.5] much more flexibly [0.3] than was possible before [0.6] so 
that's what we're going to do together [0.4] with [0.2] the er [0.2] students 
from wildlife management [0.4] er and also from vegetation [0.3] surveys [1.3] 
so what i want to do today [0.3] is to [0.2] review [0.2] the ideas [0.3] er of 
experimental design [0.3] and analysis [0.2] and then go through one more 
advanced topic [0.4] which is a very common topic and which is the subject of 
repeated measures [0.4] so that's where [0.3] you measure repeatedly on the 
same unit [0.4] be it the same animal [0.2] the same tree [0.3] or the same 
plot [0.3] you go back to it perhaps repeatedly through the season [1.0] and 
what i want you to do [0.3] with this [0.3] single example of a more advanced 
method [0.4] is to see whether the weapons you have learned through the last [0.
3] four weeks [0.4] how much can they help [1.0] because what i would like you 
to get to by the end of this course [0.9] is [1.1] er so that the subject of 
statistics is not always expanding [0.6] you've got to give yourself a 
framework [0.6] so that when 
you have a new problem [0.4] you say well what parts of the problem are new [0.
2] and what parts can fit into the the subject that i already know [0.4] and i 
want to show you [0.2] that although repeated measures appears to be a new 
subject [0.5] once you understand some of the basic ideas we've tried to cover 
[0.5] then [0.2] there are only a few [0.2] extra [0.3] changes [0.3] to be 
able to follow the ways you could analyse a repeated measures example [1.6] so 
[1.0] what i'm going to do this morning [0.8] is [0.5] to review some of the 
ideas of designing experiments [1.0] bringing in [0.2] what Genstat can do to 
help you [0.8] look very quickly at data management [0.7] er and then go into 
[0.3] as part of the analysis [0.2] how to analyse a repeated measures 
experiment [3.5] [cough] [1.6] you should have just two handouts [0.3] although 
it says here you have three [0.2] i couldn't think of anything extra to write 
for the worksheet so this is the one week where there isn't a worksheet [0.3] 
so you have a lecture note [0.2] and you have the practical exercise for [0.4] 
the practical [0.3] er [0.3] which is as usual in the Met department at 
eleven o'clock [6.6] [cough] [0.2] now we're going to look at [0.4] the ideas 
of designing experiment [0.7] and [0.4] i've kept repeating [0.3] that you 
start with your objectives [0.2] they lead to the treatments [1.0] the way 
you're doing your experiment leads to the layout [0.2] you must decide on your 
measurements and we've now moved for the last four weeks to do the analysis [0.
4] and we're going to consider these components in turn yet again [0.6] and see 
at the same time how Genstat can assist [0.3] with the randomization of an 
experiment [5.0] so [0.3] the menu that you're going to use [0.3] here we see 
Genstat it's not very clear but it's probably clearer on your slide [0.5] er [0.
2] that [0.2] you're on the usual stats menu [0.8] but instead of going [0.2] 
to analysis of variance you go up one [0.8] to an option called design [1.0] 
and there are a variety of things you can do [0.2] and all we're going to do [0.
4] is look at one or two standard [0.3] designs [1.3] and we're going to use 
them first to show you how to randomize an experiment [0.6] but also to review 
some of the ideas of different experiments [5.6] here's the 
sort of menu you get [1.6] and almost all the experiments that we've been 
discussing [0.4] you can randomize [0.3] using this menu [1.0] for the very 
simplest experiment you might consider [0.2] is a one-way design where you've 
just got treatments [0.6] and the [0.5] experiment is laid out in randomized 
blocks [1.2] so then in this menu [0.2] you have to say [0.2] what do i call 
the blocks [0.2] we've just left it with an A block [0.3] and how many blocks 
are there [1.3] what are we going to call [0.5] the units well we'll call them 
plots [0.7] and what are we going to call the treatment factor which might be 
variety here i've just called it treat [0.5] and how many levels [1.1] 
sm0659: one question please [0.3] what is the criteria for one-way or two-way 
[0.2] two-ways design [1.7] 
nm0658: you mean this 
sm0659: yeah what will be the criteria for this to choose one-way or [0.2] 
nm0658: two-way designs 
sm0659: yeah i believe there must be a two-way 
nm0658: this is where [0.3] on the second component [1.0] i've said that [0.2] 
you first think of the treatments [0.8] if you're [0.4] if you have one 
treatment factor [0.7] like variety [0.9] that is a one-way design [1.3] if you 
have two treatment factors [0.7] let us say [0.2] three varieties [0.4] and [0.
3] two levels of fertility [1.0] that would be two treatment factors [1.3] and 
then you would have [0.3] a two-way 
design [0.3] in randomized blocks [0.6] 
sm0659: different ways for the [0.4] 
nm0658: well when you have 
sm0659: bias in the [0.3] experiment [0.8] 
nm0658: when you have a one-way design it's simple [0.5] whether you call it a 
factor or a treatment it's just the same thing [0.7] when you have a two-way 
design take the example i've given you [0.5] where you have [0.2] three 
varieties and two levels of fertility [0.5] you can either think of this [0.2] 
as a two-way design [0.8] for the treatments [0.8] or you can think of it [0.2] 
as a one-way design [0.5] where that one way has six treatments [0.7] because 
it's you can think of it as three by two [0.3] or you can think of it as the 
six [0.3] and it's up to you [0.5] whether you want to think of it [0.3] as 
just six treatments [0.4] or whether you want to say well i would like to know 
right from the beginning [0.3] the layout of my experiment in terms of both 
factors [0.8] so when it's more complicated than a one-way design [0.3] you can 
choose whether to take all combinations of the treatment factors [0.2] and call 
them one big [0.8] factor with all those levels [0.3] or [0.2] to split it up 
[0.3] by the [0.5] components [2.7] [cough] [0.3] 
sf0660: do you ever have three-way design [0.4] 
nm0658: ooh 
lots of times [0.6] 
sf0660: so it's like er [1.4] on the er [0.3] er dialogues within 
nm0658: yes 
sf0660: and you just have one and two-way [0.3] and then generalize [0.8] the 
model [0.3] 
nm0658: this is this parallels here [0.2] what you would have for the analysis 
sf0660: yeah [0.4] 
nm0658: and [1.2] you can either choose [0.2] to do the analysis [0.2] thinking 
what is my design let me give it a name [0.7] or and i prefer [0.4] to say it's 
a general design 
sf0660: mm 
nm0658: how many treatment factors do i have [0.6] and this menu system [0.9] 
because it's for simple designs only [0.2] it only allows you up to five 
factors but five is probably enough [0.2] so [0.3] if you go to the general [0.
6] then it would say how many treatment factors do you have [0.4] and you could 
have up to five different treatment factors [2.8] there's no limit in Genstat 
on the number of treatment factors [0.3] but we find most experiments [0.3] 
five is enough [0.2] i would like one of the main bits of work i did when i was 
working in er [0.4] in Niger was to try and encourage people who only had [0.4] 
experiments with one or two factors [0.6] er to include more factors [6.4] once 
you click on okay [1.4] this is what you will get [1.5] so you will now get [0.
3] a Genstat spreadsheet [1.4] and [0.5] you can see here [0.4] er let me 
return to 
this previous slide [0.5] just to remind you what i did [2.4] this particular 
example [0.7] i have [0.5] four blocks [0.6] and [0.2] three levels of the 
treatment [0.6] so it's a simple design i have twelve plots so there's a twelve 
down at the bottom [0.5] which Genstat works out automatically [0.2] just 
multiplying the four by the three [0.2] that this was an experiment with twelve 
plots [1.9] so when i click on okay [1.0] it will produce the randomization [0.
3] and automatically [0.3] you will now get [3.7] i've called it data but it's 
really just the design [0.5] you will now get [0.9] this structure [0.2] in a 
Genstat spreadsheet [3.1] now i want you to notice and we will emphasize this 
later on [0.6] that [0.5] this way of laying out the data [1.0] is part of the 
simple data management [0.7] and i've seen one or two examples this term [0.3] 
because [0.2] you are all addicts of Excel [0.8] you will find that Excel gives 
you too much freedom in laying out your data [0.3] and that can cause lots of 
problems later on [0.7] so i emphasize what we discussed in [0.3] session three 
[0.6] that [0.2] when you have twelve plots in an experiment [0.4] the 
layout of the data for a statistics package [0.3] is to have [0.2] twelve rows 
of data [0.2] one for each plot [0.6] you will have the label for the plot [0.
6] and you will then [0.3] have another column which says which block [0.9] am 
i using [0.4] which [0.4] plot do i have and which treatment [0.3] and these 
treatments have been randomized [0.9] so there's your randomization [0.9] so 
this is your data laid out [0.2] in the sort of field order [0.3] ready [0.7] 
to [0.2] add further columns if you like with your measurements once you take 
them [0.9] and so [0.2] this is all ready [0.2] to be a data collection form [1.
7] and this is in Genstat [1.3] so the next part of your strategy you see [0.3] 
is to say well [0.3] after [0.3] i haven't got any measurements yet because i'm 
just designing my experiment [0.3] but when i've got my measurements [0.2] they 
will make more columns here [0.9] do i want to enter these measurement data [0.
5] into [0.3] Genstat or into Excel [1.0] i can choose [0.7] so if you choose 
to save [0.2] this as an Excel file which i did [0.5] then [0.2] here is the 
same information [0.2] saved in Excel [1.3] so now as you see you've got [0.3] 
A B C and D in Excel [0.3] and 
once you measure a few things [0.2] the height of the plants or the mean height 
of the plants [0.3] er the yield and various things [0.2] those just become 
more columns [0.3] which you can enter [0.4] into [0.4] back into Genstat for 
the analysis [2.5] and you will remember from the er [0.2] session on 
management [0.2] we have said [0.7] please enter your data straight from the 
field in the randomized order [0.5] and now you see that's a very easy routine 
[0.7] you can start in your stats package [0.2] to randomize your experiment [0.
7] if you choose to do your data entry in Excel then that's fine [0.6] you 
export [0.2] all the details [0.3] before the season [0.4] here [0.6] you now 
collect data and as you collect the data [0.4] you just type in your extra 
columns of data [0.3] and your data are ready entered [1.1] 
and in many of the courses that i give [0.4] i try and encourage scientists [0.
3] not to enter data at the end of the season when they've got everything 
collected [0.5] but if [0.5] after [0.5] a few weeks [0.2] you've measured [0.
3] the data fifty per cent flowering [0.3] then the day you measure it [0.2] 
you just type it in [0.3] it's very quick [1.1] and so it's there in your field 
book [0.3] and in it goes same day [0.5] and then if there's a problem [0.4] 
you can [0.3] probably have a look very quickly and you've seen how easy it is 
[0.3] to having [0.5] even if you enter in Excel [0.3] to then get your data 
into Genstat for the analysis [0.4] so the very day you collect your data [0.3] 
there is no problem doing your first analysis [0.6] and the procedure works [0.
4] very nicely [3.3] any [0.2] questions so far [1.3] [cough] [2.6] the second 
example [2.8] this now can follow up your initial question [0.2] this is now [0.
4] an experiment [0.5] where i have got [0.2] three factors [2.1] so i've got 
three treatment factors i'm assuming irrigation with four 
levels [0.3] i'm assuming fertility with two levels [0.2] i'm assuming variety 
with three levels [1.5] so i'm assuming a much more complicated experiment [0.
5] er how many treatments [2.8] 
sm0661: twenty-four [0.3] 
nm0658: twenty-four treatments [0.4] so we're just taking [0.5] we're going to 
have an experiment with all levels of these this would be a simple experiment 
[0.3] which would mean the total number of treatments is twenty-four [0.6] so i 
could enter this as a one-way [0.4] design [0.4] with a treatment factor going 
from one to twenty-four [0.3] and associate each number [0.3] with a 
combination of irrigation fertility or variety [0.3] or i could choose [0.7] to 
enter straight away and say i want to do this and [0.4] er randomize it [0.5] 
straight away for my irrigation fertility and variety [0.9] i've decided to 
have four replicates [1.1] and i assume also i've decided to have a split plot 
design [2.5] and in my split plot design i'm going to have my level of 
irrigation and fertility on the main plots [0.8] and i'm going to have my 
variety on the subplots [1.0] i assume i've decided this [0.3] maybe you decide 
this in 
conjunction with a statistician [1.1] before [0.3] you come to the experiment 
[0.9] now you decide [0.2] that's the design i would like to try [1.4] [cough] 
[0.2] so how many [0.5] subplots in this experiment [0.9] 
sm0662: [1.4] 
nm0658: how many how many plots altogether [2.6] you've already said there are 
twenty-four treatments [3.9] 
sm0663: ninety-six [0.3] 
nm0658: there are ninety-six [0.4] plots [0.4] just twenty-four by four [0.5] 
because i've got four replicates [0.7] and how many main plots [10.2] 
sm0664: eighteen 
sm0665: thirty-two 
sm0666: sixteen [0.7] 
nm0658: sixteen [0.3] let's have a look [6.8] well the corresponding design [0.
2] dialogue [3.3] this is an answer to your question now can you have more than 
two i now go to the general design [0.3] i've chosen to make it split plot [0.
3] so i'm going to the general split plot design [1.0] and it will now say to 
me how many treatment factors do you have [0.3] so i must understand [0.6] 
these questions what do i mean by a treatment factor [0.2] which i hope [0.3] 
you understand quite clearly from the course [1.4] and how many of those 
treatment factors [0.2] are on the subplots [1.8] well my design that i was 
considering had two treatment factors on the main plot [0.2] and one on the 
subplots [1.8] how many replicates 
four [0.4] on my main i'm going to call the main plot factor M plot i'm going 
to call the subplot factor subplot [0.9] whole plot treatment factor one i'm 
going to say is irrigate [0.8] whole plot treatment factor two i'm going to say 
is fertility [1.2] with those number of levels [0.3] and subplot treatment 
factor one variety has three levels [0.9] 
sm0659: why is that [0.7] 
nm0658: sorry [0.2] 
sm0659: why is that [0.5] 
nm0658: i chose that [0.6] i decided that was my experiment you can have any 
number [2.0] so [0.2] you have to decide how many varieties do you have to 
compare [1.4] 
sm0659: er [0.3] why did you call those subplots variety [1.5] 
nm0658: because i [0.3] thought that that would be a factor [0.6] which [0.3] i 
didn't need big plots for [0.4] that's part of the design process [0.3] you 
have to [0.9] you have to decide [0.3] do you need the same size plots for all 
your factors [1.8] if you don't [0.7] then [0.6] are there any factors that 
need larger plots than the others [0.3] here irrigation [0.2] obviously often 
needs large plots [1.3] sometimes fertility [0.7] levels need large plots 
because you get leaching from one plot to another [0.9] and often varieties [0.
4] you can have quite small plots [0.2] breeders have 
very small plots [0.8] so i'm assuming [0.2] that [0.2] either because of your 
expertise or in discussion with a statistician [0.3] you decide that [0.2] you 
can get away with small plots for variety [0.3] but you need larger plots for 
here and that's why you chose a split plot design [1.1] if that had not been 
the case [0.4] then as we've said before [0.7] i would recommend that you don't 
have a split plot design you've have a randomized plot design [0.2] and then 
you [0.4] you'd just have a different menu [0.4] and you just say these are the 
three factors [2.3] and what i want to show you in a minute is that we'll come 
and check [0.4] well was it a good idea [0.2] our design [1.5] 
sf0660: can i just ask what's the randomization seed is that the way of 
generating random numbers [1.2] 
nm0658: that's correct er yes i haven't explained these things down at the 
bottom [0.4] the randomization seed [0.6] means that that's the point at which 
it starts its random number generation [1.0] so [0.6] if you were [0.3] to make 
a note of this [1.6] then you could always regenerate the same randomization [0.
2] yet again you don't have to keep everything [0.4] by [0.6] typing this 
number in yourself so usually you allow the 
computer [0.2] to choose this and it's different every time [1.2] but if you 
chose [0.2] to make it the same you'd get the same randomization [1.1] and 
that's quite a good way of keeping a record of the randomization you had [0.3] 
without noting everything down [0.4] if somebody would say please could you 
print me another copy of that [0.4] well you can always [0.3] print the 
resulting [0.3] spreadsheet [0.5] but if you said well [0.8] sending the 
information all you have to do is fill this in this way and give that random 
number seed and you'll get the same [0.2] randomization [7.1] and this is the 
randomization that follows [1.7] so this is the output when you press okay [3.
0] it goes on of course [0.4] down to ninety-six plots we only have fourteen 
plots here [0.4] but you see it gives you a column which says which block [0.2] 
which main plot [0.2] which subplot [0.3] which level of irrigation which level 
of fertility and which variety [0.8] these of course are not randomized [1.5] 
but [0.8] the this [0.5] which is the level of irrigation fertility and variety 
[0.4] that is associated [0.2] is randomized [0.6] to a 
certain extent [0.6] what you should notice [0.4] is that the variety [0.9] is 
randomized on the little plots [0.9] so here's the first main plot which goes 
one one one [0.8] while the subplots go one two three [1.5] you will notice [0.
2] that irrigate [0.5] is at the same level one [0.2] for all those three [0.4] 
because that's at the main plot level [1.0] and so is fertility [0.7] but the 
varieties [0.3] go [0.3] one two three [1.2] because they are randomized on the 
subplot level [0.8] and when we go to the next main plot [0.6] here [0.8] we 
have main plot two [0.5] subplots one two three these two again remain the same 
because it's a split plot design [0.3] and these three this is a complete 
repeat of all the values here [0.5] so each main plot [0.8] is a total repeat 
[0.6] for all the levels of variety [0.6] but it just is one plot from here [7.
8] 
sm0667: [1.0] er [2.2] if they're completely randomized will it be numbers one 
to ninety-six or [0.8] 
nm0658: er sorry [0.3] 
sm0667: if they're completely randomized [0.8] 
nm0658: if it were completely randomized [0.2] 
sm0667: mm 
nm0658: without even any replicates [0.6] then it would randomize everything [0.
3] for the numbers one to ninety-six [0.3] so you can choose [0.3] do i have 
replicates [0.6] which is a sort of blocking idea [0.5] er [0.2] do i have 
treatments [0.2] and how do i randomize [0.5] you probably wouldn't use this 
menu if you just wanted a set of random numbers there's an easier way of doing 
that [0.4] but that would be the extreme [0.3] for the randomization that's 
right [1.1] now [1.1] you asked just now about this randomization seed [0.4] 
but there are other options down at the bottom [0.7] that you could also ask 
for [0.9] and i've used one or two of these [1.2] er i've asked could i have a 
dummy ANOVA table [1.2] which might answer a bit more of your question [0.2] 
why did i choose to put varieties at the subplot level [0.8] what i would like 
to see occasionally is [0.2] is the resulting design a sensible one for me to 
continue with [1.6] and so i could learn a little bit about that with a dummy 
ANOVA table [0.3] and i will show you in a minute what that is [1.3] i could 
ask [0.3] for trial ANOVA with random data [2.7] what it does there [0.3] 
please don't make too much of that the random data [0.2] is only there to show 
you what will the results look like [0.8] from this sort of experiment [0.2] 
when you have 
some data [0.3] how will the results be laid out [0.7] so it can show you [0.3] 
the way in which the results will be laid out and i think that's quite useful 
ahead of the experiment [0.4] to say is that a sensible thing so [0.2] this is 
the sort of tool that i would like also for discussion [0.3] with people such 
as yourself [0.3] when you're saying i'm thinking of doing this sort of 
experiment [0.4] and i can say well [0.2] this is the sort of result you will 
get [0.8] are you happy A with the design [0.5] and B [0.2] that you can then 
understand the results [2.1] so let's show you what this [0.2] works out [1.3] 
because i ticked the dummy ANOVA table [3.2] here i have [0.2] the dummy ANOVA 
[0.4] from Genstat [6.1] i asked you how many main plots [0.3] and you see [0.
2] this is straight print out from Genstat [0.3] and here you see here is the 
main plot [0.2] section [2.7] so this is all revision [0.3] but you can see [0.
2] that from the point of view of these two factors [0.6] this is a simple 
experiment [0.2] not with ninety-six [0.2] plots [0.5] but with ninety-six 
divided by three plots [0.8] so there are [0.3] thirty-two plots [0.8] thirty-
two main plots [0.3] 
which is my [0.2] twenty-one plus three plus one plus three [0.2] plus the 
three [0.2] plus the extra one [0.3] so here we are at the main plot section [1.
7] so when you want to evaluate if this was a good experiment [0.5] from the 
point of view [0.3] of the main plots [0.6] you can start having a look just up 
here was it a good experiment to study irrigation and fertility and the 
interaction [1.7] 
and then when you come along to the subplot section [1.2] you come along here 
[0.4] and you have all the degrees of freedom [0.5] down here [0.2] ninety-five 
is the total because that's ninety-six minus one there's your [0.3] n-minus-one 
[2.1] [cough] [0.6] i haven't given this but you'll try this in the practical 
[0.3] Genstat also provides a sample ANOVA with random data to show you how the 
results will look [0.4] and i find that quite useful in teaching and we're 
asking you to get this in the in the practical to have a look at that [3.1] and 
the sort of question you can now answer is [0.4] how would the degrees of 
freedom change if you moved fertilizer to the subplots [0.9] if instead of 
having [0.2] two main plot factors and one subplot factor [0.3] you said well 
fertilizer could go on the little plots i wonder w-, how that would look [1.1] 
and you have now two ways of doing that [0.3] you can do that from first 
principles yourself [1.2] or [0.2] you could run the randomization again [0.3] 
and move one of the factors from the main plot 
level to the subplot level [4.2] i hope you can see from here if we move [0.2] 
fertilizer from here [0.4] down to here [1.3] then this would come down here [1.
0] and also this would come down here [1.1] so we would now be doing an 
experiment [0.5] with [0.7] three [0.2] degrees of freedom there [0.6] and [0.
2] just three values here [0.5] because this stays [0.8] and so this now [0.9] 
the intera-, the residual [0.2] would have nine [0.2] degrees of freedom [0.3] 
we've got ourselves rather a small experiment at the main plot level [0.3] 
that's not such a good idea [1.4] so it's those sorts of things you can now do 
[0.2] either by studying this [0.3] or [0.2] by running through the design 
again [0.9] and that's what we've asked you to do in the practical [5.2] any 
more comments or [0.4] questions [1.4] on this design part i'm now moving [0.3] 
to data management [1.0] 
sf0660: er i'd like to ask something on the number of plots [0.3] that you have 
for your [0.8] er main [0.4] blocking practice the main treatment practice [0.
3] 
nm0658: yep 
sf0660: er [1.6] from [2.0] [0.6] where he's given us the three factors [0.4] 
nm0658: yes 
sf0660: er [1.6] i don't know [0.3] if you've got irrigation as four levels 
fertility two levels [0.3] how do you then [0.4] get [0.7] can you not say that 
you had nine [1.0] plots for your [1.2] two [2.0] 
nm0658: with the design as randomized [0.8] that's what you've got [4.8] 
sf0660: so for your main plot from that did you say you've got how many plots 
[0.6] 
nm0658: for my main plot i said i've got [0.3] thirty-two [0.6] main plots [0.
8] which is the ninety-six that i had [0.7] divided by three [2.7] 
sf0660: but if you didn't go through that ANOVA table [0.4] er is there any way 
of working that out just from what you get at the very beginning er structure 
[0.4] 
nm0658: yes [1.6] if i didn't have that at the beginning [0.5] i would say [0.
4] that 
at the main plot level [1.0] what i've got [0.3] is [1.3] four blocks 
sf0668: [0.5] 
nm0658: and four levels of irrigation which is four times four 
sf0668: [0.5] 
nm0658: times two levels of fertility [0.3] which is four times four is sixteen 
times two is thirty-two [1.3] 
sf0668: [2.9] 
nm0658: any more [0.7] questions [2.6] the second topic [0.5] 
sm0669: er excuse me [0.2] er er did [0.3] and the design in the in in reality 
[0.4] 
nm0658: yeah 
sm0669: i was thinking how about three factors [0.4] plot [0.5] that's to say 
maybe to have [0.2] the er irrigation out the way [0.6] then the fertility as a 
subplot [0.4] and then the variety as the sub-subplot [0.5] in that case we 
will have been having [0.4] more [0.2] of a [0.6] how will i call it degrees of 
freedom than irrigation and the fertility [0.4] as a main plot and then having 
the variety as a [0.5] subplot how would you [0.5] [0.8] 
nm0658: okay [0.7] can anybody [0.4] er [0.6] let me repeat the question [0.3] 
and then i'm interested if any of you can now [0.2] provide an answer you are 
now you've had fifteen weeks of statistics so you're all semi-statisticians [2.
2] [laughter] now the question was [0.9] that this was [0.4] er deliberately 
something a little more complicated as an experiment [0.2] than you've had 
before it had three factors [2.5] and the question was well if you have three 
factors [1.0] why did we just have a single split [1.2] or if i understand your 
question correctly why didn't we have [0.3] two splits [1.0] where perhaps we 
have irrigation on the main plots [0.7] and then we have [0.4] er [0.2] 
fertility on [0.4] the sort of middle-size plots [0.5] and variety on the 
little plots [1.3] and that is quite common [2.7] does anybody have any 
comments on that 
suggestion [0.2] supposing that [0.3] when you're back home [0.3] and starting 
to work somebody says i'm doing a three factor experiment [0.3] so i'm doing it 
on a split-split that's called a split-split plot [0.4] design [0.6] very 
difficult to say if you've had a few drinks [0.9] [laughter] does anybody have 
any [1.3] anybody have any comments [0.5] on [0.2] whether they would encourage 
that whether they would like that [1.2] 
sm0670: [0.4] 
nm0658: ah [laugh] 
sm0670: and it was [0.5] how would we differentiate the effect of [1.2] the two 
that you combine in that one [1.6] like the [0.4] the variety [1.2] the 
irrigation in the block [0.4] 
nm0658: yes [2.7] here 
sm0670: combine [0.3] make the main [1.3] 
nm0658: i had [0.4] here i had irrigation and fertility on the main plots 
sm0670: yes [0.3] 
nm0658: yes [0.3] 
sm0670: how would you s-, [0.5] 
nm0658: how [0.4] 
sm0670: how would you differentiate the effect on your [0.3] experiment [0.6] 
of [0.9] 
nm0658: well [0.2] if you look at the ANOVA table [3.1] that's the same [0.7] 
question as though [1.6] you just did [0.2] a randomized block design this is 
like a randomized block design [0.5] and you'll have a set of treatment means 
which give you [0.2] the mean for each level of irrigation [0.8] and you will 
have another one [0.2] which gives you the mean for each level of fertility [0.
4] just the two means 'cause there's [0.4] only two levels [0.3] and then 
you'll 
have another table [0.8] for the [0.3] interaction [3.8] and that is one reason 
[0.3] if you're not sure how that's going to look in practice [0.5] that's one 
reason [0.2] why Genstat gives you [1.1] a dummy analysis [0.2] not just with 
the degrees of freedom which you got here [0.9] but also [0.2] with [0.2] 
random data [0.2] so you can see how all the means will look and you can see [0.
3] i wonder if that will give me sufficient information [0.5] to understand [0.
2] all the components [0.3] of the treatments that i've applied to my 
experiment [2.1] i think it's a similar question [0.3] that [0.9] most people 
feel that [0.2] if they've got many factors [0.4] they're much happier [0.2] if 
each factor's at a different sort of level [0.6] which leads you [0.3] towards 
having two factors in a split plot experiment and three factors in a split-
split plot [0.6] and i hope you never have five factors [0.4] because then 
you've got [0.5] huge plots and [0.2] and so on [1.9] does does anybody have 
any thoughts about would you encourage [0.6] 
sm0671: well [0.8] just a general stab in the dark [0.2] 
nm0658: yes [0.2] have a s-, 
sm0671: you split your plot [0.6] into [0.5] levels of fertilizer and split it 
again [0.4] and they're very small plots [0.5] the smaller the sample [0.6] er 
[0.4] if you have a larger sample variations [0.4] tend to be absorbed in a 
large sample rather than a small sample [1.7] is that 
nm0658: that that's [0.2] that's almost there [0.2] 
sm0659: [0.2] depends on the way degrees of freedom [0.2] 
nm0658: degrees of freedom will come 
in [0.7] er you would [0.4] it would be like having the top level it would be 
like just having irrigation at the top level [0.9] does anybody have any 
general feelings [0.2] about whether [0.5] they would [0.3] encourage 
experiments to be at lots of different levels [0.9] or whether they would re-, 
prefer information [0.3] to be at a single level [0.7] 
sm0672: [0.9] 
sf0673: wouldn't it depend on what you're looking [0.6] 
nm0658: i-, 
sf0673: at what you were interested in [0.7] 'cause if you want to to find [0.
3] er equally [0.2] the effect on irrigation and part of the effects of 
fertilizer and variety and on interactions [0.6] then by splitting it up into 
several levels you're going to lose a lot of degrees of freedom for your [0.5] 
upper levels [0.9] 
nm0658: right [0.4] 
sf0673: er [0.2] so surely [1.4] the information that you'll obtain [0.2] will 
be er [0.7] you won't have [1.6] [0.6] er [3.6] i don't know how to finish that 
sentence [1.1] 
nm0658: er let let me try and finish it for you well maybe try and ask somebody 
else 
sf0673: mm 
nm0658: because i think you were voting for the side of not having too many 
levels 
sf0673: mm [0.9] 
nm0658: er there were other people who i think instinctively said let's have 
more levels is there anybody who would [0.7] who approves of the idea of having 
lots of levels [2.6] 
sm0671: in blocks [0.4] why didn't we use [0.2] one of those factors [0.5] 
nm0658: right e-, each one going down [0.2] 
sm0671: yeah that's the way they [0.2] 
nm0658: right [0.5] okay [1.2] having [0.5] your idea of having [0.2] two 
factors in the split block design [0.3] and three factors therefore in a split-
split block design [0.3] is extremely common [2.1] my view is [0.6] your other 
intervention which is to say [0.3] let's not have too many levels unless we 
need to [0.8] so the general view i have is that [0.2] 
lots of levels [0.2] causes complication [0.7] if you can have your experiment 
at a single level [0.7] life is simpler [0.7] all your tables are compared at 
the same level all your plots are the same size [0.5] and the analysis the 
design is simpler [0.2] and i think the analysis is simpler [0.7] so the split 
plot analysis has lots of different levels [0.3] and you will find when you 
look at the standard errors [0.6] that that indicates that the analysis becomes 
very messy and complicated [0.9] so i would prefer [0.2] not to have [0.2] 
split plot experiments [2.6] the only reason i would have a split plot 
experiment is if [0.2] some factors need [0.2] large size plots [0.2] like 
irrigation [0.6] and other factors don't [1.3] and then [0.2] if the l-, if 
irrigation needs large size plots and we want to have irrigation and variety [0.
6] if we want to have no [0.3] different levels [0.2] then all plots have to be 
very very large [0.9] whereas because variety only needs small plots [0.2] we 
can choose to split the large plots that we need [0.3] for irrigation [0.2] 
into subdivision [0.2] that seems a good reason [1.0] i find 
there is no other good reason [0.3] for having split plot [0.4] designs [1.6] 
we will see later today [0.3] that [0.2] the whole nightmare of repeated 
measures analysis [1.4] is a similar argument that the repeated measures [0.8] 
are [0.2] repeated within the plot [0.5] they're like a sort of split plot [0.
6] where time [0.3] is each measurement [0.5] and you will find [0.2] as soon 
as you have lots of different levels in your data [0.6] your analysis is 
getting messier [1.0] so i'm not very comfortable [0.2] with the idea [0.2] 
that if you increase the number of factors [0.5] you also increase [0.6] the 
layout problems [0.3] remember [0.3] at the very beginning [0.6] we've said 
that [2.0] when you look at design [7.0] please think of your treatment 
structure [0.3] we said we want three factors [0.2] because that ties in with 
our objectives [0.9] and then please think of your layout [1.2] now what the 
people who do split-split plot all the time are doing [0.4] is they are 
thinking of these two together [1.2] 
the treatments and the layout [1.0] and they keep [0.2] confusing them together 
[0.9] and i would like people to think of the treatments first and say [0.6] i 
could satisfy my objectives by having three factors and there'd be lots of 
objectives i could satisfy [1.0] now can i have a very simple layout [0.2] a 
simple layout [0.2] is a randomized complete block [0.6] sort of layout [0.4] 
with twenty-four plots [0.3] in this case [1.0] for each block [0.8] and so [0.
5] there are no different levels [0.3] for the treatments [1.2] and if you can 
manage that [0.2] please do it [0.8] if you can't manage that [0.3] you say 
well [0.5] maybe i'll have to go to a split plot [0.2] sort of layout [0.2] 
with two levels [0.8] but don't volunteer for it and say because i've got two 
factors i will automatically go for two levels [0.7] i think that causes many 
problems [0.4] in the analysis of experiments [0.2] and is part of the reason 
people 
don't exploit their data [0.2] as much as they could [0.6] 
sm0669: er my last question er would you be able to [0.4] tell what a [0.2] 
reasonable portion is statistically speaking [0.7] of the effects of say 
irrigation [0.2] confidently [0.4] then confidently [0.4] them [0.6] i was 
thinking [0.4] when you split them and you could do these things [0.4] see i 
would have gone in for irrigation as the main [0.3] and then the er [0.4] er 
fertility i mean the maybe [0.4] the fertility [0.2] that and then gradually 
move on [0.5] to the variety [0.4] and have an idea whether even there is an 
interaction between [0.4] variety i mean irrigation [1.0] 
nm0658: if you [0.2] lay it out in a randomized block with just one level [0.4] 
you can answer all those questions [1.2] you can askn-, answer the questions 
about irrigation about variety and about the interaction [0.7] whether it's 
laid out as a split plot [0.2] or whether it's laid out as a randomized block 
[0.5] so [0.2] the questions you can answer about the treatments are the same 
[1.8] there are some people who would argue [0.5] that when it's laid out [0.5] 
as a split plot [0.7] compared to random and let me [0.2] come back a step [0.
2] when it is laid out as a randomized block [0.5] you are sort of treating 
each factor equally [0.4] they are all on plots of the same size [1.3] one of 
the arguments given in the textbooks [0.6] for having [0.2] a split plot [0.6] 
is where [0.2] you want more information [0.3] in this case let's say on 
variety [0.2] so you're going to put them on the little plots [0.2] and then 
they're always close together [1.2] and you ye-, you want less information [0.
2] on the irrigation so you put those on big plots [0.2] that are [0.2] by 
definition further apart on average [1.0] and [0.2] so when you compare a 
randomized block [0.3] with a split plot [0.6] in the split plot [0.7] 
according to the textbooks [0.6] you get more information on the subplot 
factors [0.7] and less information on the main plot factors [1.8] the problem i 
have is that [0.7] you get much less [0.3] on the main plot compared with a 
very little gain [0.4] on the subplot [0.5] and also [0.2] that does not 
account for the fact you also [0.3] add in [0.3] unnecessary complication in 
interpreting the results [0.8] which means that [0.3] i don't like that as a 
reason [0.6] for doing a split plot experiment [0.2] the only reason i like [0.
4] is the fact that you have to for practical purposes [0.4] and there are many 
of those [5.7] okay [0.3] the next subject [2.8] then we'll have a break [0.3] 
and we'll discuss repeated measures after the break [2.9] now [1.4] what i hope 
to show you [0.5] is [0.8] that [0.2] the ideas of data management [0.6] which 
we [0.3] distributed in session three which was [0.6] very early last term [0.
7] that was data management [0.2] for any sort of data [1.3] when we translate 
those ideas into experimental data [0.4] 
experimental data are quite simple [0.3] usually [0.3] and so you shouldn't 
have any problem [0.5] if [0.2] you manage the data sensibly and if you keep 
the principles [0.6] so we're going to review the standard ways of entering 
experimental data [0.3] which i think follow automatically if you've understood 
the ideas of design [2.1] and [0.3] what we're also going to show you [0.3] is 
what happens [0.4] if your data have been entered differently [0.5] and this is 
where we're hitting a new problem now [0.2] because of all these people as i've 
said before [0.3] who are maniacs for Excel [0.5] and that means you can enter 
data in all sorts of crazy ways doesn't it [0.6] er [1.0] and and then you can 
have problems reorganizing your data [0.3] so you can do the sensible analysis 
[0.6] and i have to tell you that as statisticians now [0.4] this is serious [0.
6] because [0.5] in the olden days we found that we spent all our time helping 
people on the analysis [1.0] now [0.7] most of your time seems to be spent [0.
3] on rescuing your data from poor data management [0.3] the analyses are very 
quick [0.5] you just click on the 
ANOVA button [0.2] in Genstat [0.5] so the analysis step [0.2] is very very 
quick [0.3] and you've done it many times this term [0.7] the step which isn't 
quick [1.5] is reorganizing your data [0.3] because they were entered in a 
funny way [1.5] so there are two ways to avoid this [0.3] and i want to [0.2] 
indicate both of them [0.3] the first is please enter your data sensibly [0.2] 
and then you won't have this problem [1.3] and the second [0.2] is if you 
haven't entered your data sensibly [1.0] please don't re-enter [0.7] please use 
the computer [0.2] to reorganize your data [0.5] but accept [0.3] that [0.2] 
don't get annoyed at either Excel or at Genstat or Minitab [0.2] because that's 
taking the time [0.3] i'm afraid it is the data manipulation [0.2] that does 
take the time [0.4] so [0.6] allow a little bit of time for that [2.0] so we're 
going to show you both [0.6] and [0.7] you either need to reorganize the data 
because you've entered it in an odd way [0.9] or [0.2] because alternative 
analyses [0.4] require different layouts [0.2] so [0.2] sometimes [0.2] when 
the data aren't so simple [0.4] you might have three different analyses [0.2] 
and for one analysis it was good to 
enter the data across [0.3] and for another analysis [0.3] it was good to enter 
the data down [2.5] so you need to become a little adept [0.2] at data 
manipulation [0.7] or you're going to waste a lot of time [1.3] and a lot i do 
mean it [4.3] perhaps i'll [0.5] er tell you one [0.7] er [0.4] not horror 
story but close [1.9] the subject of data management [0.6] is not well reported 
in the textbooks [0.9] when i arrived back a couple of years ago [1.7] one of 
the first advisees i saw [0.8] was somebody [0.2] who was just finishing his 
PhD at Reading [1.5] and he had been on this course [0.9] but we hadn't been 
discussing much about data management in the olden days [0.8] and he'd followed 
a little bit of the notes on how to organize the data [1.3] and it was 
discussions with him [0.4] that clarified to me that we must include [0.6] data 
manipulation data management in this course [0.3] which is why we've changed 
this course to include this [1.1] he had an experiment which was done here [0.
8] and he'd done his experiment [0.3] at [0.3] two different places for each of 
the three years of his PhD [2.6] and [0.2] each experiment 
he had measured ten different things [1.8] so he'd had ten measurements [0.8] 
it's all very simple [0.3] these were very simple experiments [1.3] he had 
forty-eight plots in each experiment [1.5] he had now looked in a textbook [0.
9] and he looked at his notes [0.3] to see how to enter the data [1.6] and he 
had entered the data quite nicely organized [0.5] with three columns [1.3] the 
first column [0.3] was the yield [0.2] or the measurement [0.9] the second 
column [0.3] was the block [0.5] and the third column was the treatment [0.4] 
there were twelve treatments [0.3] and there were four [0.5] blocks [0.6] so he 
had [0.2] measurement [0.2] block [0.4] treatment [1.8] because he'd looked in 
a textbook [0.6] and textbooks [0.4] only seem to deal with experiments when 
you have a single measurement [1.6] which is unlike the real world [0.4] where 
you always have lots of measurements [1.4] he decided to enter his ten 
measurements [0.4] in ten different files [2.0] so he now had [0.2] ten little 
files [0.2] each one with three columns [1.0] the block and the treatment 
columns were the same [0.3] and the measurement was different [2.2] and then he 
had six experiments [1.9] 
so he now had sixty files [1.9] each one [0.2] with three columns in [1.1] and 
each one followed precisely the method of analysis he'd been taught in his 
course [0.7] and he'd finished his analysis [0.6] and he was three weeks away 
from submitting his PhD [2.3] and his supervisor [0.6] looked at his 
information and said [0.8] there are two interesting things i would like you to 
do in addition [1.6] the first is [0.2] that i've noticed that you have 
measured [0.7] the yield [0.5] in two different ways [0.2] i'd like you to do 
what's called an analysis of covariance where you adjust one measurement for 
the values of another measurement [1.7] could you please do a bi-, a simple 
analysis of covariance [0.6] for each of your six experiments [0.5] and he gave 
him a very simple thing [0.2] and showed him it was in the textbook [2.4] the 
second question [1.0] was even more perplexing to him he said could you do a 
simple combined analysis where you combine the information for the six 
experiments together [0.2] to see how [0.7] the [0.2] results you have 
reinforce each other [2.0] and so he came to 
statistics [0.4] to do this [0.7] he knew no [0.4] computing he only knew how 
to turn [0.4] the computing handle [0.7] so he was now being asked [0.8] a 
simple question of data management [0.4] well it would have been simple [0.3] 
if his data had been organized sensibly [0.6] but can you see that [0.2] this 
idea of covariance [0.9] which [0.2] would have been trivial with Genstat [0.2] 
normally [1.0] meant that he had two data files [0.2] which he had to merge 
together because the two columns were in different files [1.6] nobody'd taught 
him about merging files [2.8] and [0.6] the combining of the experiments the 
six experiment [0.2] meant not just that you merged the files [0.4] but you 
then had to put them end to end [1.0] because he had forty-eight but now he 
wanted forty-eight times six [0.5] with another column which said which 
experiment it was [1.7] so when he came with three weeks to go [1.4] we 
explained [0.3] these ideas to him [1.0] but because he had no concept of data 
management [0.3] or of [0.3] computer ideas such as merging files [0.7] he 
never succeeded [0.2] the only way it could possibly have happened [0.7] is if 
somebody else had taken 
his data and done it all for him [0.9] that's not his PhD [0.8] or if he'd 
learned a little about data manipulation [0.5] and these were very very simple 
tasks [1.1] i have to say these are much easier tasks [0.4] now you've got 
Windows [0.3] than they were [0.2] when you had to merge files using DOS 
commands [1.4] 
but [0.2] nevertheless [0.7] this was not possible for him to do [1.5] and [0.
3] that's data manipulation [0.4] taken to its illogical extreme [1.2] but i 
have to say [0.3] that the way he was encouraged to enter the data first [1.2] 
was exactly what was recommended [0.5] in a very popular textbook [1.0] which 
we use on our courses [0.5] says [0.2] when you're entering your data into the 
computer this is the sort of layout [1.2] and it encouraged the layout [0.3] 
which caused his disaster [0.3] because it didn't say of course in practical 
experiments [0.2] you will have more than one measurement [0.3] and just put 
them end to end [0.4] don't make them in a different file [6.5] so let me just 
remind you [1.9] this was the yield data [0.8] that we discussed in session 
three [1.8] you have nothing new you need to learn if you understood session 
three [0.3] notice [0.2] that the way that we have laid this out [0.2] with the 
block numbers [1.3] is exactly the same [0.4] as the way Genstat has just 
randomized your experiment at the beginning [0.7] so here [0.5] are the blocks 
[0.2] the repetitions [0.2] and the treatments [0.7] and here are the 
measurements [0.3] that you just type in when you get them [1.3] so as long as 
you don't mess things up [0.8] from where you started [0.8] the rules are very 
simple [0.5] how many plots do you have [1.2] each plot [0.6] becomes a row [1.
2] each column [0.3] is a measurement [1.7] i can't think of anything [0.4] 
that's more simple [2.8] and that's what you have to do [4.0] you will find [0.
3] that there's one complication which we do discuss in session three [0.6] 
which is [0.2] what happens if some of the measurements [0.4] are made at the 
plant level or the plot level [0.4] here sorry the plot level [0.2] and other 
measurements are made at the plant level [1.9] so for example [0.2] it's very 
popular to measure the yield by harvesting at the plot level [0.4] but you 
might measure the height of twenty plants in each plot [0.6] i wonder how you 
would enter that [0.3] to which the answer is [0.2] either in Excel or in 
Genstat [0.3] you have one sheet [1.0] for your plot level [0.3] and you have 
another sheet for your plant level [0.4] and we've covered that [0.5] both in 
[0.4] the discussion in section three and in the practical that you did [1.3] 
so as long 
as you're happy with those [0.5] you don't have a data management problem [0.9] 
you should enter your data in this way [2.9] i can leap ahead a little bit [0.
4] to [0.4] these repeated measures ideas [0.3] supposing that [0.3] here's an 
example [0.2] i'm sorry it's a bit in French [0.3] but this is the weight [0.2] 
of small potatoes middling potatoes et cetera [1.4] we're going later on to 
consider repeated measures [0.6] which is [0.2] measuring things [0.3] after [0.
2] twenty days twenty-five days thirty days and things like that [0.3] how 
should you enter those data [0.2] answer [0.2] they are just measurements [0.2] 
so enter them across [0.8] just as though [0.6] you were entering different 
measurements [0.2] so just because you made the measurements which differ in 
time [0.3] don't get more complicated [0.2] just treat them as a measurement [0.
6] and so [0.2] if you have [0.2] six repeated measurements [0.6] just [0.6] 
they're six measurements [1.0] enter them across [0.2] they haven't changed the 
number of plots in your experiment [0.4] so [0.2] they go across [5.8] this [1.
6] is how not [0.2] to enter experimental data [1.9] this is the same data [4.
6] here we have [0.9] the 
treatment information [1.1] and here we've entered the data [0.2] for replicate 
one replicate two replicate three for one measurement [4.2] that is very 
popular [2.3] isn't it [0.8] [laugh] [0.3] i i'm sorry i'm i'm looking at you 
because [0.4] i-, i-, you know it's not you that's made this as a [0.3] this is 
encouraged in many departments [0.8] and it seems very obvious [0.9] and it 
works quite well for simple problems [0.2] also [0.5] notice that if you were 
doing the analysis by hand [1.6] in the olden days [0.7] this is exactly what 
you would do [0.4] because you could work out the mean of these [1.3] and you 
could get [0.3] the mean for that treatment [0.9] so you could get all the 
treatment means down here [0.3] and all the block means across here [2.3] so if 
you are in the habit [0.9] of confusing entry with analysis [0.5] this is 
wonderful [0.4] you can use Excel [0.5] to work out calculated columns [0.2] 
and it all seems to work quite well [0.2] until you try and do a full analysis 
and then it all falls apart [1.2] do not confuse analysis [0.4] with entry [1.
3] if later on you want to analyse your data [0.8] either with Genstat or Excel 
[0.5] you can 
enter the data in the proper way [2.5] that's like this [2.8] if you would now 
like [0.7] to look at the data [0.3] like this [1.6] i've said [0.3] this is 
how not to enter experimental data [0.5] if you want to look at the data like 
this that's terrific [0.2] it's called tabulation [0.8] you enter the data [0.
2] in long columns [0.2] and then you say [0.2] please tabulate the columns [0.
2] and across the top i want one factor [0.2] and down here i want the other 
factor [0.5] and then i'll put some summaries i don't have a problem looking at 
the data like that [0.2] i have a problem entering the data [0.3] it's 
confusing entry [0.2] with analysis [1.1] in Genstat [0.2] that's called 
tabulation [0.9] in Excel it's called pivot tables [0.2] i don't care h-, 
whether you do it in Excel [0.3] or whether you do it in Genstat [0.6] so enter 
your data in Excel [0.3] the [0.2] proper way [0.3] enter your data in Genstat 
the proper way [0.2] if you want to look at the data like this [1.3] tabulate 
the data present the data very nice you can see what's going on [0.8] you can 
see what a nightmare this is for your data entry [0.9] by looking [0.4] and 
explaining to 
somebody [0.2] how you would enter [0.2] these data [0.3] in that format [0.7] 
because that only works if you've got one measurement [0.4] but in the real 
world we have [0.2] here we have one two three four five six measurements [0.3] 
are you going to enter them on six different sheets [0.3] it's all getting 
messy [2.1] this [0.4] is very simple [1.6] and if ever you want to transform 
your data [0.2] and get [0.3] the total [0.3] it just becomes another column [0.
5] down here [3.6] 
sf0668: why doesn't he have repeated measurements of so many different factors 
[0.3] [0.4] 
nm0658: if you've got repea-, lots of repeated measurements they still keep 
going across [1.3] 
sf0668: [0.9] 
nm0658: and you might have sets of repeated measurements you might measure [0.
2] lots of things after twenty days another set of things after thirty another 
set of things after forty [0.2] they can go across [1.1] so [0.3] you can go 
way across [0.3] here [0.6] er [0.8] 
sm0669: what if experiment where you take in data [0.3] from de-, for instance 
you do maize and [0.2] pea or whatever it is [0.6] and then you have all 
similar columns like this will it make safe [0.2] to try to box them together 
or have separate because [0.3] on them really [0.3] i mean they're independent 
[0.2] 
nm0658: okay 
sm0669: [0.7] sort of shoot up into [0.4] for say [0.3] maize height [0.4] 
nm0658: okay 
sm0669: 
nm0658: right th-, we'll that can be the last [0.2] question before we have a 
quick break [0.5] 'cause you have such a nice new coffee place downstairs [0.6] 
er [0.2] th-, [0.3] er the question i hope it's understood to everybody [0.2] 
there are some experiments [0.2] which are called mixed cropping experiments 
does anybody not know [0.2] what is a mixed cropping experiment [0.9] you you 
all happy that [0.7] this is a an experiment where you might have [0.2] maize 
on sub-, [0.2] some plots [0.3] and beans on other plots [0.4] and then [0.2] 
the aim of the experiment is to see how the maize and the beans mix together [0.
3] so some plots have both maize and beans together [2.2] and my simple rule 
for this [0.6] is [0.2] please stay simple [1.2] so that is [0.5] that please 
do exactly the same [2.5] down here you have all your plots [1.0] your 
treatments will say is it maize sole or bean sole or a mixture [0.5] and then 
here are all your measurements [1.4] and in your measurements you will have 
lots of blanks [0.9] for the observations [0.4] that [0.3] don't the maize 
observations don't have any beans leave it blank or put it as missing it 
doesn't matter [0.9] keep life simple [1.0] in the past [0.4] there has been a 
problem [0.2] particularly related [0.3] to [0.2] a package [0.2] called Mstat 
[0.3] that is very poor at data manipulation [0.6] and that has caused people 
to say [0.3] well maybe i'll enter [0.2] the maize data in one file and the 
bean data in 
another file [1.0] or maybe i'll enter [0.3] the maize sole data in one file 
and the bean sole data in another and the mixed data in a third [1.0] and then 
they get very confused [0.6] er [0.3] so to avoid all that confusion [1.2] keep 
my simple message [0.7] don't confuse [0.4] the entry with the analysis [0.3] 
the reason people choose the different files [0.2] is to simplify the analysis 
[0.6] simplify the analysis afterwards [0.6] but for the entry [0.2] the entry 
is simple [0.3] if you say [0.2] how many plots did i have [0.8] and i will 
enter observations [0.2] and if there are some observations i don't have [0.2] 
they're like missing values [0.8] the reason is different [0.4] but you just 
leave it blank [0.5] or you put a missing value code in [1.7] let's have a nice 
simple life [0.2] and then you will find [0.2] that the analysis is also simple 
[0.8] after the break [0.3] er [0.5] we'll look at [0.2] how to manipulate the 
data [0.2] and then [0.2] quickly on to repeated measures [0.3] okay we'll have 
a break 
nm0658: 
i'm realizing as usual that luckily [0.4] you have the slides for everything [0.
3] because i don't think i'm going to finish yet again and i do want to discuss 
repeated measures [0.5] so i may have to leave out a little bit of [0.2] one of 
the topics on data management [2.8] i just want quickly to review the ideas [0.
6] to compare these two formats you've seen [1.2] our recommended way of data 
entry this is all revision now [0.2] has one row for each plot [0.2] or unit [1.
0] and that's the lowest possible level [0.3] if it's a subplot it's one row 
for each subplot [0.5] so [0.2] in the example [0.2] in lecture three there are 
twenty-seven rows [0.3] because there were three reps by nine treatments there 
were twenty-seven plots [0.6] in the randomized one just now there were ninety-
six rows because there were ninety-six plots [1.0] therefore there's one column 
for each measurement [1.8] the other way [0.2] looks like a textbook example 
for a hand analysis [2.0] also unfortunately [1.0] and they've been written too 
many times but because they're all powerful they don't have to listen to 
anybody [0.5] er in Excel [0.3] it's the way you need to lay out the data for 
Excel to do an analysis of variance [0.2] we do not recommend you use [0.5] 
analysis of variance in Excel [0.2] if you're going to do analysis of variance 
[0.5] use it in Genstat or Minitab or anything [0.3] but Excel's analysis of 
variance is not very good [0.9] 
sf0674: statistically it's not very good [0.7] 
nm0658: it's it doesn't do enough it doesn't show you residuals 
sf0674: 
nm0658: it doesn't encourage a critical looking at data [0.2] it only goes up 
to two level factorials [0.5] and it's much better [0.3] if you're going to use 
Excel for your statistical analysis [0.3] you go up to [0.2] simple description 
and tabulation [0.3] maybe you do graphics in Excel [0.4] but don't get into 
ANOVA and regression in Excel [0.2] you've gone off the end of Excel [0.4] use 
the proper tool for the task you have [0.5] when you have complicated [0.7] 
analyses of data [0.4] there are many statistics packages [0.3] use the one 
that's most appropriate [0.9] for you i think for experimental data [0.2] it's 
absolutely clear [0.2] that 
Genstat is the most appropriate [0.3] but the important thing is that you use 
[0.2] one that is appropriate [0.5] and [0.2] you've fallen off the end [0.3] 
of a very general purpose tool [0.2] which is a spreadsheet [1.8] [cough] [0.7] 
so [3.6] as i've said before this lecture [1.0] try to use the standard format 
for your entry it will save time on the analysis later [0.9] but if somebody 
has entered the data differently [0.3] it's usually quicker to reorganize the 
data [0.2] than to have to retype [1.3] so [0.2] don't say oh gosh you've made 
a big mistake [0.2] you'd better start again [1.2] use the computer to help 
reorganize the data [0.6] use either a statistics package or a spreadsheet [2.
1] i find [0.5] that Genstat [0.6] the competition between Genstat and Excel [0.
3] for reorganizing the data [0.4] for you [0.3] would be quite a good one [0.
8] because [0.2] i think for reorganizing [0.3] if you knew neither package it 
would be quicker [0.2] to learn and use Genstat [0.8] because it's built to do 
that sort of reorganizing [0.8] than to use Excel [1.0] but [0.2] life is not 
equal [0.2] most of you know Excel very well and you hardly know Genstat [0.4] 
so some of you will 
prefer [0.3] to reorganize your data in Excel [0.4] and other people would 
prefer [0.4] to learn a bit more of Genstat to do the reorganizing [0.3] you 
should choose the method that's most appropriate for you [0.4] and that will 
depend [0.2] on your work in the future [0.4] but if you're thinking of using 
Genstat seriously [0.3] in the future [0.3] then [0.4] try and learn a bit 
about its methods [0.2] for data manipulation and i'll show you [0.3] a little 
bit about what that means [2.7] first you have to understand what you're trying 
to do [2.2] here is an example [0.4] where i've taken the one from before [0.3] 
where i assume we've entered the data in the wrong way [2.0] and we want to go 
[0.2] from this way [0.2] to [0.2] this way [0.7] that's the way we would have 
had if we'd entered it correctly [1.0] with [0.2] the replicates and the 
treatments and the dry matter [1.0] so you ought to see [0.2] what it is [0.2] 
we are trying to do [0.3] we want to go [0.2] i've said from this wide format 
[1.3] to the long [0.2] format [2.6] we are trying [1.0] to take these and i 
think you can see i hope you can see [0.3] we're trying to stack these [0.2] 
one below another [1.3] 
so what goes across [0.3] becomes stacked [2.3] it's not er only that [1.5] you 
could picture that's very easy in Excel [0.8] but we want to do a little more 
than that [0.5] because we want to take this [0.2] which are already in the 
right form [0.9] and we want to repeat that [1.5] so [0.2] this one on the left 
hand side [0.4] gets repeated [0.2] three times [1.4] once [0.2] for each 
column [1.8] so it stays opposite [1.5] not just this measurement but this 
measurement [1.1] and [0.3] these reps one two and three [1.0] we want to have 
a new column called rep [0.3] which goes one [0.2] two [0.2] three [3.1] so [0.
5] it's not just stacking [2.2] it's a bit more [0.7] what do we mean in 
practice usually [0.3] you want to stack [0.2] the measurements [1.3] and you 
want to carry [0.2] the factors [1.8] let's have a look at how that works in 
Genstat [2.5] Genstat [0.8] has a dialogue [0.3] called stack [3.8] you just 
have to go [0.4] you should be getting to the stage with Genstat that you 
become curious and you think well there must be there somewhere [0.4] and you 
go to spread [0.5] manipulate [0.4] and you will find stack [1.3] within stack 
[1.6] it will say how many columns [0.3] do you want to 
stack together [1.3] well it should be quite easy for you to say i want three 
columns because i have rep one rep two rep three i want those [0.3] stacked [0.
5] one below the other [1.4] i want to record the column source i want to 
record where they came from [0.3] in a column called rep [0.3] that's a new 
column [1.9] and then you put [0.2] your observations [0.5] which were called 
rep one rep two rep three which were actually the yields [0.6] and you put them 
there [0.2] you notice it puts a one [0.3] beside [1.1] to say they're all 
going to be one column [1.3] you might have had [0.2] more things as well [0.9] 
and then it would have twos beside [0.5] because that would now be a second 
column [2.3] and you have a repeat column [0.2] which is treat [1.6] you want 
treat [0.2] to be repeated [0.2] each time that's what we said was needed [0.9] 
and when you click on okay [3.4] then you will find [0.7] that it will produce 
[1.1] exactly [0.4] that [0.2] spreadsheet [2.6] so it will take that [0.7] and 
produce that [4.5] i don't think that's too difficult [2.2] so that is stacking 
in [0.4] Genstat [8.4] once you've got that far and you realize that's quite 
easy [0.5] i [2.0] just 
occasionally [0.3] you want to go the other way round [4.9] not often [0.7] but 
sometimes [0.2] you have it long [0.4] and you want to go wide [1.6] and if you 
ever need that [0.7] there is you should now not be surprised if there's a 
stack dialogue [0.3] there is also an unstack [1.1] and [0.2] this is like 
stacking things on a shelf [0.6] you you can either stack them up [0.3] or you 
can take them from there [0.2] and you move them [0.6] sideways so that they're 
now unstacked [0.7] so that also exists [6.4] that is one sort of data 
manipulation [2.0] the second one [0.4] i'm going to [0.2] mention that it 
exists and then leave it [0.5] because [0.6] i think there is no not time for 
the repeated measures [0.4] so let me just mention [0.3] that [0.2] the next 
it's in your notes [0.9] that the next is that you sometimes have data [0.3] at 
two levels [0.7] and you must summarize one level [0.3] to move it to another 
level [2.0] the example that [0.2] you have in the notes [0.3] is data that 
were measured [0.7] yields were measured on the plot level [0.7] but tuber 
number was measured on the plant level [0.4] and you want to summarize the 
tuber number 
onto the plot level [0.3] to analyse with your other data [0.7] and you'll be 
doing that in the practical [0.3] one of the examples does that [0.3] and that 
is called [0.2] data summary [0.3] before the analysis [1.1] and that's very 
common [0.3] and is usually where measurements are made at a lower level [0.4] 
than the one where the treatments were applied [1.2] so here the varieties were 
applied at the plot level [0.4] but you measured something [0.2] on plants 
within the same plot [2.5] and so i just give you the three examples that 
you've had already [0.9] in session three [0.2] there were tuber measurements 
on twenty plants in each plot [2.2] in session in the last session [0.3] you 
had tree measurements [0.8] this was what namex [0.4] described [0.3] on four 
trees in each plot [0.4] so you have [0.2] a plot was four trees and you 
measured [0.4] the girth [0.8] and other measurements [0.3] on each tree [0.2] 
within the plot [0.6] but you want to analyse at the plot level [0.5] because 
that's where you applied your treatments [1.1] and in the Genstat guide [1.2] 
the very simple example that you will have seen in the 
first part of the guide [0.2] had four replicates and three treatments [0.9] 
and there were twelve pens [0.7] but there were two sections to each pen [0.5] 
and that's the example in the Genstat guide [1.8] so [0.2] there we have [0.2] 
twelve [0.2] plots [0.4] and two sections in a pen so we have twenty-four 
sections [0.4] and we want to summarize from the twenty-four [0.2] up to the 
twelve [1.9] now [0.2] i've put at the bottom [0.2] check you recognize these 
[0.3] these three as the same problem [0.9] because in the Genstat guide [0.4] 
we show you in detail how to solve this problem [1.0] you will meet it very 
often and if you recognize that this is the same problem as this and this and 
many others [0.6] then you can use Genstat [0.4] to do [0.2] that initial 
summary [2.0] and [0.3] let me just show you the [1.1] the dialogue [0.2] that 
you get with Genstat [2.3] without giving you [2.4] there's your data at your 
low level [0.4] you want to get the total or the mean [0.2] at the higher level 
[0.2] and you want to carry other things along [0.2] so your analysis can 
proceed at the higher level [0.4] and so again [0.4] you should find [0.3] 
there is a summarize the 
spreadsheet [0.3] which is another dialogue [1.0] which will help you [0.3] 
when you've got very detailed data at one level [0.2] and you want to move up 
[0.2] to another level [4.6] so that's in your notes [0.4] and it'll be in the 
practical [2.6] does anybody have any [0.3] i haven't covered it because [0.6] 
i'm a little behind and i would like to tackle the repeated measures [1.1] does 
anybody have any comments or questions on that [2.2] yes [0.3] 
sm0675: case we have er four trees er in each plot [0.6] we know for example 
the directional reach of the trees 
nm0658: yes [0.2] 
sm0675: so we wanted to know the production for the block [0.2] separately then 
[0.5] we add all the production of the four [0.7] trees and then er we'd like 
to know [0.5] er for the whole plot [0.5] 
nm0658: that's correct that [0.2] that that's very common [0.5] er [0.4] you 
you don't have to and in the notes [0.2] we describe you can do the analysis at 
the tree level [0.7] but you must accept that you applied [0.2] your treatment 
at the plot level [0.5] so the usual thing is to say [0.2] what was the 
production for the whole plot [0.5] and then you must say well do i want the 
mean production per tree [0.3] or the total production [0.2] and that depends 
on the problem [1.9] 
on er some experiments i analysed [0.4] er on disease [0.7] there were [0.7] 
measurements [0.3] a an [0.6] an insecticide [0.2] was applied [0.4] at the 
whole plot [0.9] and then ten plants were measured [0.7] to see how diseased 
they were [1.4] and [0.4] the idea was now [0.2] for each plot you wanted a 
measure of disease [1.8] in the olden days people always used [0.7] the mean [0.
2] disease score [0.5] as a measure of the disease per plot [0.8] we decided [0.
5] that because we wanted to see the most effective [0.4] insecticide [0.2] we 
should also calculate [0.3] the maximum disease score [0.2] namely the disease 
score of the worst [0.3] plant in the plot [0.3] as a summary number [0.4] to 
say [0.2] that characterizes [0.6] the worst [0.2] that could possib-, if that 
worst is pretty good it's a good insecticide [0.6] so [0.2] it's not always [0.
2] usually you have the total or the mean in the plot [0.3] but it can be the 
maximum or the minimum [0.4] that is also useful [0.3] so you calculate a 
summary statistic for the plot [0.6] which you then analyse at the plot level 
[6.3] last subject [1.9] repeated measures [3.3] these are very common in 
designed 
experiments [1.1] where measurements are repeated on the same unit [2.6] they 
can be in time [0.4] or in space [1.9] and the problem [0.3] from a statistical 
point of view is the same [1.2] so [0.5] animals weighed each week [0.8] would 
be [0.3] i'm assuming the animal receives a particular treatment [0.7] at the 
beginning of the experiment or has having a diet as you go through the 
experiment [0.3] and is weighed each week [1.0] so rather than just an ordinary 
yield experiment where you measure just once at the end [0.3] you record 
successively in time [1.3] that is a repeated measure in time [1.8] most [0.2] 
repeated measures are measures in time [2.1] occasionally here's another 
example tree diameter is recorded every six months [0.4] to see about the 
growth [1.9] in human experiments [0.4] you often do lots of measurements in 
time [0.3] you measure the weight of babies [0.3] every month [0.3] for a year 
[1.7] you measure [0.3] the effectiveness of a treatment for cancer [0.4] by 
recording [0.2] every three months [1.3] the effect on patients and so on [0.4] 
so this is very common [0.2] in many fields of application [0.9] just 
occasionally [0.2] these repeated measures are in space [0.6] an example [0.3] 
would be if you have a hedge [1.0] and you want to see [0.2] the effect of the 
hedge on the plants [0.3] that are [0.3] close to the hedge you might have [0.
3] a hedge [0.2] er with a certain tree species [0.5] and then [0.2] you have 
some rows of maize [0.3] which grow [0.4] and you have six rows [0.5] going 
away from the hedge [0.6] and now you want to measure each row [1.3] so you 
applied your plot [0.2] consists of your hedge [0.4] with six rows each side [0.
6] and now [0.2] you want to repeat the measurement namely the yield [0.4] but 
not for the whole plot [0.3] but for each row [0.4] to see the effect in space 
[0.3] as you go further away [0.2] from the hedge [1.1] and that is a problem 
of repeated measures [0.3] in space [3.7] those repeated measures introduce 
problems of data management and analysis which we're going to look at [1.1] and 
it reviews many of the ideas [0.5] from this part of the course [3.0] we'll 
have a very simple example [0.5] and so you have more information it's in the 
Genstat guide [1.9] there are five [0.5] replicates [0.2] this is an example 
for you because 
it's example where there are no blocks [0.2] one or two [0.3] er so there are 
just [0.2] fifteen petri dishes little dishes [1.6] and [0.5] there are three 
isolates of a fungus [0.7] and they're repeated [0.2] five times [0.4] but 
there's no [0.2] group of five here group of five here there are just fifteen 
of them [2.1] and there were six measurements made [0.6] on days three four 
five six seven and eight [2.3] [cough] [0.3] how many plots [0.3] how many 
measurements [1.0] is it clear about the normal data and how many plots [4.8] 
how many plots [0.3] easy question [6.3] fifteeen 
sm0676: fifteen 
nm0658: it is [0.4] i'm sorry you you probably were confused 'cause it's too 
easy [0.5] er it depends whether you think of your your confuse your 
measurements with your plots you see if you don't [0.3] it should be obvious 
you have a question [0.3] 
sf0677: [0.2] destructive [0.7] measurements on the same plot as the 
nm0658: the des-, [0.5] tha-, that's a very good question [1.1] when you are 
measuring on the same plot [1.4] you have the choice of measuring let's say 
height [1.0] which is [0.4] something you can go back to exactly the same plant 
[0.2] and measure [0.6] or [0.8] harvesting a few plants in the plot [1.1] from 
a stats point of 
view [0.4] it's roughly the same thing [0.4] because [0.5] they are [0.3] 
they're still within the same plot [0.7] but from a precision point of view [0.
4] it's much better [0.4] if you can [0.6] i-, i-, it isn't quite the same 
because [0.2] if you go back to the same plant [0.7] then your repeated measure 
[0.6] is at the level plant [0.4] if you're measuring the height [1.1] whereas 
if it's destructive [0.6] and [0.2] you're measuring let's say the height of 
four plants and then you throw them away [1.0] then [0.4] and you measure the 
height of four more plants [0.2] then you're still repeating the measure [0.3] 
but the level is the plot level not the plant level because you can't go back 
to the same plant [0.2] because it's harvested [0.7] so where it differs [0.2] 
is the level [0.2] at which you're able to do the repeat [0.9] usually we find 
[0.3] that the lower the level you can do the repeat the better it is [0.9] so 
[0.2] we often find [0.3] that non-destructive measurements [0.4] are 
tremendously useful [0.5] and last week [0.2] i [0.3] was examining somebody 
whose thesis [0.4] is on taking aerial photographs [0.4] of plots [0.5] where 
[0.2] you can measure [0.4] the [0.7] the area 
roughly of each plant [0.5] repeatedly [0.8] very very easily by taking a 
photograph [1.1] and that is [0.5] non-destructive [1.0] and [0.3] was shown to 
be a very good way [0.6] compared to these small harvests that people often 
take [0.4] where you get exactly what you want namely the harvest but it 
destroys it so you can't measure the same plants later on [5.0] here's the data 
[0.6] well sorry [0.4] er i haven't [0.5] i'd asked you how many plots there 
are [0.3] but it hasn't [0.4] answered [0.9] you should be saying [0.4] there 
are six measurements so that's six columns [1.1] the fact they're on days three 
four five six seven eight doesn't affect [0.2] and there are fifteen [1.2] 
plots [0.4] therefore i'm going to have [0.2] fifteen rows of data [1.2] and so 
[0.2] the data [1.1] are going to look [0.3] well here's an example of the way 
the data could look [1.5] where [0.2] there's the unit [0.2] there's the 
isolate there's the rep [0.3] and there's my measurements on day [0.2] three 
four five six seven eight [9.5] [cough] [3.4] now [0.6] you now have to think 
of your strategy [0.4] for the analysis [4.3] and here we begin [0.2] with a 
slight [0.3] problem [2.1] that is [0.2] i would like you to 
look constructively at the data [1.4] exploratory analysis [0.2] we said [0.2] 
is very important [0.3] i wonder how you would like to explore [0.5] these data 
[2.7] well a very common way that people would like to explore the data [0.8] 
is [0.2] to see [0.8] what's the change in observation [0.4] over time [2.3] 
sort of [0.3] notice this goes [0.4] on the first petri dish i go three-point-
seven five six-point-one [0.2] seven-point-five eight-point-three nine-point-ei 
[0.2] seem to going up with time [1.5] to understand my data [0.5] maybe it 
would be nice [0.8] to look at that sort of graph [0.5] as a function of time 
[0.9] that would be a nice exploratory method [0.3] but unfortunately for a 
stats package [0.3] exploration works best on columns [1.3] so you may wish to 
do that exploration in Excel [1.9] or [1.2] you will find Genstat helps [1.3] 
so a strategy for the analysis [4.4] nothing changes you've changed the problem 
but you haven't changed the strategy which is [0.2] please start by looking 
critically at your data [1.3] so start with data exploration and that's usually 
graphs [0.5] so you look at all the data [0.2] are there any odd 
observations [0.7] you could do those Excel or in Genstat [0.6] and you could 
get one graph for each plot [0.5] so there'd be fifteen graphs [0.4] be one way 
of exploring the data [1.2] let's have a look at that first [5.4] i'll come 
back to that for the second part [0.3] so there's a way [2.4] in Genstat [1.3] 
of exploring the data [4.1] and you will find that Genstat has a little menu [2.
5] which [1.6] puts this out automatically [1.0] so here's the graphs [0.3] for 
[0.2] er one treatment [0.3] the second treatment [0.2] and the third treatment 
[2.8] and here is the graph [0.3] for [0.4] the means [0.4] for the three 
treatments [0.3] so there's a bit of analysis [0.3] but here we have all our 
data [0.8] this is the ninety observations [0.7] there are fifteen lines [0.2] 
because we have fifteen plots [0.8] and each line has six points [0.2] because 
we have six time points [0.4] so we actually have all our data [0.4] here [0.8] 
and we could see if there was some odd observations [0.7] i don't see anything 
particularly odd [0.4] 
sm0678: when you say odd [0.7] what exactly do you mean [8.7] 
nm0658: the full set of numbers [2.0] consists [0.2] here's all our data [0.9] 
these are all the numbers [1.4] and what we have [0.3] is we have 
fifteen plots so we're actually using all the numbers [0.3] in those graphs you 
can see every number [0.4] somewhere there so we're using [0.2] all our data [0.
8] in producing those plots 
sm0678: my question is by looking at the graphs [0.2] what [0.9] are we looking 
for [0.7] 
nm0658: er okay [10.0] does anybody have anything they've found without knowing 
seriously what they're looking for [3.3] 
sm0679: there is a 
nm0658: any impressions 
sm0679: increase er [0.9] to an X [0.2] -axis [0.5] and 
nm0658: there seems to be an increase [1.4] any er re-, remember your chick 
experiment [0.4] and the increases sort of straight lines or curves [1.4] 
sm0680: curves 
sm0681: curves [0.5] 
nm0658: curves [1.7] we we [0.4] sort of like that 
sm0682: yes 
nm0658: or [0.2] wiggeldy [1.2] 
sf0683: some of them [0.5] 
sm0682: mm some of them were 
sf0683: and some were straight mm [2.1] 
nm0658: i don't notice any that go sort of right like that [1.0] sort of [0.2] 
starting going up to the top and coming down [0.4] remember [0.4] we're 
worrying about statistics there's variation so [0.2] th-, everything you can't 
have things that are exactly straight 'cause we're just connecting the points 
[1.5] do you do you think [0.3] that [0.3] for some of these [0.6] it would be 
sensible to have a straight line model [0.3] would that be a rough reasonable 
summary [0.6] 
sm0684: [0.4] 
nm0658: for all the plots [1.7] are there any plots where you think a straight 
line model isn't going to be sensible [2.1] i don't actually see many [0.4] 
which is surprising usually you find [0.4] maybe one treatment is curved and 
the other treatments are straight [1.2] as you found with the barley and the 
wheat one was more curved than another you might [0.3] want to [0.2] recognize 
that [1.4] does anybody see any very surprising observations [2.4] i don't [0.
8] i don't i don't sort of see a sudden spike like this [0.4] which might be a 
recording error [1.3] so this is exploration [0.6] and [0.4] exploration can be 
positive or negative [0.7] usually i find [0.2] you notice one or two very odd 
observations [0.2] here i don't see anything very odd [0.4] and things seem to 
be increasing [0.7] so that here [0.7] where this [0.2] is the mean [0.3] of 
that one [2.2] i feel [0.2] reasonable confidence that drawing a straight line 
[0.4] which is the average for those points [0.5] is probably a reasonable 
summary [1.1] and i notice now [0.2] that this straight line so this is 
analysis and this 
is just presentation of the raw data [1.9] and i now notice in this summary [0.
5] that [0.7] these [0.6] all three seem to be going up [0.3] but this maybe is 
going up more gradually A it's lower [0.2] and it's going up more gradually [1.
1] so from these [0.6] i feel that this is probably a fair summary of the data 
i don't see any reason [0.2] to say oh gosh it's not fair because of [0.2] this 
[0.9] and in here i'm starting my analysis [1.1] and i've done that in a very 
simple way and visually [2.3] does that help to answer your question [9.6] 
[cough] [3.4] okay so we have our data [1.9] and [0.4] i was looking for the 
strategy [0.5] okay [2.6] so [1.7] i suggest for all analyses you start with a 
simple summary [0.2] and then you go on to simple analyses [2.1] what could the 
simple analyses be [4.0] well here we have the data [4.0] one simple analysis 
[0.2] could be to analyse the data on day three [2.4] just take one time point 
[0.9] another one [0.6] day eight [0.8] so there's a very simple analysis [0.3] 
you could analyse [0.2] each of your observations [0.2] separately [1.8] the 
next simple analysis [0.4] could be [0.6] to take a useful summary [2.5] one 
summary might be the difference between 
day eight and day three [0.2] has the change [0.2] been the same for each 
treatment and each replicate [1.7] so that's what we suggest [4.7] i keep 
losing [0.5] the er [17.0] so i've suggested the first simple analysis could be 
the data at each time point [1.4] then we could have simple function [0.9] like 
the final minus the initial [0.4] or [0.6] we could have the slope [0.2] of 
each line [1.0] now you don't want the slope [0.2] if too many lines could be 
very curved [0.4] but here i think getting the slope of each line [0.2] might 
be a sensible summary [8.8] as a strategy what sort of strategy is this [1.6] 
well i would claim that the repeated measures [0.5] are like observations at a 
lower level [0.5] they're not exactly the same [0.7] but they're a little like 
a split plot analysis they're like taking [0.2] day [0.4] as [0.2] a level 
within a petri dish [2.1] and then [0.5] in this example we have six 
observations within each petri dish [0.2] for the six days [1.4] or we have ten 
weights within each animal [1.1] so it's a little like a split plot [0.2] 
experiment [0.3] where the factor time [0.7] is within [0.4] the treatment [1.
1] just as [0.3] an 
ordinary split plot [0.8] but it's not quite a split plot [0.2] because we 
don't randomize the times [1.0] we can't [2.4] like [0.7] anything like that [0.
2] the analysis will be simpler if we first get a summary value at the plot 
level [1.2] so our analysis is going to be simple if we summarize up [0.6] to 
the petri dish [1.5] whatever we do [0.2] we've got fifteen petri dishes [0.2] 
whatever summary we'd like to get [0.5] that would be a simple [0.2] analysis 
[0.2] if we can go [0.2] you were asking about split plots and i was saying [1.
0] simple experiments are at one level [0.9] well repeated measures [0.2] bring 
in a second level [0.9] there are many methods for analysing repeated measures 
[0.2] bringing in the two levels [0.7] but the simplest is not to have two 
levels [0.3] but is to summarize the data [0.3] from the repeated measures up 
to the one level [2.1] because that's where we apply the treatment [1.6] which 
summary is appropriate [0.5] depends on the data and the objectives [0.9] and 
the booklet on analysis that i gave out in week eleven [0.3] gives you some 
more details [3.9] so [0.7] the graphical display [0.3] indicates [0.3] that 
a useful summary might be the slope of the regression line for each petri dish 
[3.1] so we then have a problem how do we get the fifteen slopes [2.0] in Excel 
we could use the data as they stand [1.0] in Genstat [0.2] be better to stack 
the data [1.2] so now to do those slopes [0.5] because Genstat works with 
columns [0.2] be better to stack [0.2] the data [1.9] and that was shown [0.3] 
earlier [3.5] and once you analyse [0.2] with the stack data [0.9] you will 
find [1.1] and the way we'd go through that [0.3] and get the regression [0.2] 
you will find [0.3] that here are the fifteen [0.8] observations [0.3] and here 
are the fifteen slopes [1.4] and here is the analysis [0.3] where we're 
actually analysing the slopes [1.0] we're getting the individual slopes [0.4] 
and there's fourteen degrees of freedom here [0.4] because there's fifteen 
petri dishes [1.2] and [0.3] we find [0.4] that er the effect of slope is 
statistically significant [0.2] and these are [0.2] the three slopes [1.6] 
which are the lin-, the slopes of those three lines [0.3] and we find [0.2] 
that these first two [0.4] treatments are about the same [0.2] but this slope 
[0.3] is rather [1.2] flatter [9.1] there 
are many other methods of analysis of repeated measures [0.5] and Genstat [0.4] 
has a whole set of dialogues specially for that [1.3] they all try and get more 
levels more information [0.3] by leaving the data at the two levels [0.3] 
rather than summarizing up [0.5] to one level [1.6] they're often attractive in 
principle [2.9] they're needed that should be an if [0.3] they're needed if 
there is not enough data [4.8] but [0.9] to me they have a major problem [0.4] 
that they are much more complicated [0.4] and often [0.3] their real problem is 
you can't tailor the analysis to the precise objectives of your research [0.7] 
so they're like many complicated analyses [0.4] that they're wonderful [0.3] in 
principle [0.3] but in practice they don't help [1.1] that they are playing 
with data [0.2] often [1.1] and [0.2] my conclusion is [0.2] use the simple 
methods wherever possible [1.4] and if you do use the more complex methods 
which you can because Genstat provides them from menus [0.9] then don't just 
use them [0.2] make sure they add constructively [0.4] to what you were able to 
do quite simply [0.3] with the simple methods [6.1] okay [0.5] practical work 
[1.7] 
the practical follows the topics covered in this session so it's useful for you 
to review those [0.3] so we're doing some on designing some on managing data 
and some on repeated measures [1.5] in each case [0.4] i've deliberately used 
examples from the Genstat guides [0.5] so [0.5] you don't have to finish [0.3] 
just concentrate on those parts you find most interesting [0.3] and that will 
help you [0.4] get more experience [0.2] in using Genstat [5.2] and [1.4] two 
final slides [3.4] this is now the end of the five sessions [0.3] which are 
specifically for the analysis of experimental data [1.5] you now should have 
two things [0.8] the first is [0.2] the broad picture of the role of statistics 
in research projects [1.0] which has come from [0.3] sessions one to five [1.4] 
that was last term [0.8] so you should have an idea of how you use statistics 
in design in data management [0.8] you should have 
reviewed [0.3] basic statistical techniques [1.0] statistical inference ANOVA 
simple regression [0.2] that was the second part last term [1.1] now you should 
be familiar with some of the special methods for analysing experimental data [2.
2] and hopefully [0.9] you are therefore ready [0.3] for a brief introduction 
[0.4] to [0.5] the role of modern statistical methods [0.9] to help you [0.4] 
in processing your research data [7.8] the lecture room next week [1.0] is [0.
2] the plant sciences [0.3] lecture room so we're not here next week [0.4] 
we're together with the other group [0.3] in plant sciences [0.2] 
ss: 
nm0658: in 
ss: [0.3] 
nm0658: sorry 
ss: 
nm0658: the ground floor i pres-, the ground floor lecture room in plant 
sciences [0.5] 
sm0685: you mean the small lecture theatre [0.9] 
sm0686: there's one [0.3] 
nm0658: i hope not it's w-, 
sf0687: there's one lecture theatre anyway isn't it 
nm0658: sorry 
sf0687: just one there isn't it 
nm0658: there's just one one lecture 
theatre i've been told [0.4] it's got to take seventy of us 
sf0687: no it's large 
sm0688: 
nm0658: so it's the large lecture theatre 
ss: 
nm0658: and next week the practical is again for you in the Met department [0.
5] 
sm0689: is it just for us or will it be everybody else here as well [0.3] 
nm0658: everybody is in the lecture and the practical is still split [0.3] into 
the two groups [0.7] as you go [2.3] can you please [2.3] we want your critical 
review of [0.3] these five lectures [0.8] and so we have another of the 
evaluations [0.2] this is on this session [0.3] any comments [0.2] remember 
we've changed the course a lot this year [0.3] so any comments [0.2] they don't 
have to be polite [0.4] and i'll collect this in the practical [0.7] so can you 
take one as you go out [0.3] and then i'll collect these in the practical