Transcriptions
Note: this content has been automatically generated.
00:00:00
good morning and managers who i'm head of things good laboratory animals signs at the university of zurich
00:00:07
and um the other lot lighter of uh of the courses i'm
00:00:13
lazy and a continuing education courses for like animal experiment teachers
00:00:18
it's actually are in a crime a location to
00:00:22
change behaviour of scientists uh in and experimentation
00:00:28
uh_huh and today i want to talk to you about um a little and
00:00:35
investigation purely mathematical that i've started two
00:00:39
years ago letters and friends
00:00:43
after having heard a or a year in and it goes on so
00:00:49
on the meeting of the space so slightly off and experimentation
00:00:55
and michael fasting a statistician from great britain and
00:00:59
presented a or can he claimed that
00:01:03
inclusion of both sexes in animal studies that could be done and no statistical cost
00:01:11
i couldn't believe that i needed it takes that uh mine mathematics fund enough to
00:01:15
matter can that be too of course we can take this answer project um
00:01:21
and and that all start end at you know we're on it
00:01:25
now for two years that's how difficult it turned out
00:01:29
the dish the difficulty stems of on one hand uh from being apologises all the just
00:01:35
in in being a um a mathematician or so uh to find common ground
00:01:40
and of uh in of of doing such as that it was very difficult it is to very
00:01:47
difficult when i say i want to know this can you calculate that uses but that's critical
00:01:52
and so one of the main uh topics here is how can include
00:01:58
both sexes in animal studies and i've read
00:02:02
now uh papers discussing that it's always
00:02:07
refer to as if you have to double the group sizes so so in this case you have a a study
00:02:13
was twelve uh males as is done in many fields and
00:02:18
but you know by including the second sex emails
00:02:23
had a man a double the size of animals this is
00:02:28
also important as we are all under pressure into strains
00:02:32
of and and and the three r.s we do supplied fine
00:02:37
and to use as few animals if possible the cities the impact to them
00:02:43
so we have to think hard to use not too many and so would it
00:02:47
help us to include both sexes into you really have to double the size
00:02:53
of the groups um as i said already fasting has the opposite opinion he uh things you
00:02:58
can do this and have no it just the cost and you don't lose power well
00:03:08
we we start then um some model in and um found that if you just include
00:03:15
six as a compound or a second factor that you don't and lines
00:03:19
in a so called multi factorial design um the increase in
00:03:24
group sizes it actually marginal it's small and and
00:03:28
typical animal study when you add two in a in a in
00:03:32
a study that hasn't for a groups what at one or
00:03:36
two animals to each group and you would be fine so that the outcome as you see in that are not well
00:03:44
doing so is uh the same as a you have heard before
00:03:48
you include six as a component but you don't analysts
00:03:52
if you want to enlist sex you have to model it in and you have to then analysts
00:03:56
whether it has a influence on and the primary better that you and i and rising
00:04:04
to do so you have to um of course correct mike testing and you have
00:04:08
to a model in interaction effects between and sex and the other very
00:04:15
if you do that what happened is actually that you see down here in the lower graph
00:04:22
um when you do model with interaction you increase the power of your experiment
00:04:28
instead of losing power so knowing that you have both sexes and
00:04:36
including this internal analysis and make to study at or you can continue to uh with
00:04:42
our and modelling and in the morning streams
00:04:47
and scenario that the uh use
00:04:50
very analysed the data for sex and um for for that and alive and you
00:04:56
have to include twenty percent more animals in the study
00:05:00
to really get in a a meaningful result
00:05:05
that said there is no reason to not include both sexes
00:05:09
in an experimental studies and that uh come from me
00:05:12
yeah i'm in in all the just like training um