ANOVA ADVICE
come to lectures importantinfo that i maynot email you about
Do NOT use course notes
Only for examples
Do tutorials but do NOT trust solutions
Study using past papers
focus on latest ones some content is diff
don't make rewrite notes
pretty notes
DO YOUR QUIZ's
course notes
WHAT WILL BE COVERED
T tests revision introstat
Chp 9
F distr f tests
CRD t one way ANOVA t fiteasDsigdifference
RCBD two ANOVA
way
Factorial Design two way ANOVA w intera
Tests for
normality
, T TESTS
introstat Chp9 Pg 208 211
DATA
We have e groups with the following info on each
Grt Gra PSEMIYESTICS
ni E S na ta S2
Ni na is It necessary
HYPOTHESE POPULATION MEAN
Ho µ MI
Hi Mi two sided
Ma
or
Mi plz one sided
c
or
pl plz
KEY ASSUMPTION
We assume that there is a common
variance 02 between groups I 32
if sit si it is just chance
by
there is still a to
perform a e test
way
if this assumption does not hold but
it has a complicated formula for y is not
covered in this course
,TEST STATISTIC
E lot Tal Ini Ma
spool Vint
now sit.net rEEiii's
Well if Ho is true we'd expect the
two sample means to be quite similar
15C 0521 TO
3 He Mal 0
how likely are we to
It g observe the data we did
assumingHo is true
txt
Beg probably reject Ho
Ild tt
probably don'treject Ho
How do we decide what is big
look at table
for this we need d of
, DEGREES OF FREEDOM
df of It is n t
df In 1 na 1
n t Mz 2
txt and If we can use our tables to
get a
palm
using
NZ p value f significance level
D value
Probability of observing your test static t
eg
assuming that Ho is true
p value is informed
your by your
observations
significance level d
a pre defined cutoff value chosen by the experimenter
Q is not determined observations
by your
A I error i.e the probability that
Type you
reject Ho when Ho is actually true
ALWAYS report your actual p value if it is available
come to lectures importantinfo that i maynot email you about
Do NOT use course notes
Only for examples
Do tutorials but do NOT trust solutions
Study using past papers
focus on latest ones some content is diff
don't make rewrite notes
pretty notes
DO YOUR QUIZ's
course notes
WHAT WILL BE COVERED
T tests revision introstat
Chp 9
F distr f tests
CRD t one way ANOVA t fiteasDsigdifference
RCBD two ANOVA
way
Factorial Design two way ANOVA w intera
Tests for
normality
, T TESTS
introstat Chp9 Pg 208 211
DATA
We have e groups with the following info on each
Grt Gra PSEMIYESTICS
ni E S na ta S2
Ni na is It necessary
HYPOTHESE POPULATION MEAN
Ho µ MI
Hi Mi two sided
Ma
or
Mi plz one sided
c
or
pl plz
KEY ASSUMPTION
We assume that there is a common
variance 02 between groups I 32
if sit si it is just chance
by
there is still a to
perform a e test
way
if this assumption does not hold but
it has a complicated formula for y is not
covered in this course
,TEST STATISTIC
E lot Tal Ini Ma
spool Vint
now sit.net rEEiii's
Well if Ho is true we'd expect the
two sample means to be quite similar
15C 0521 TO
3 He Mal 0
how likely are we to
It g observe the data we did
assumingHo is true
txt
Beg probably reject Ho
Ild tt
probably don'treject Ho
How do we decide what is big
look at table
for this we need d of
, DEGREES OF FREEDOM
df of It is n t
df In 1 na 1
n t Mz 2
txt and If we can use our tables to
get a
palm
using
NZ p value f significance level
D value
Probability of observing your test static t
eg
assuming that Ho is true
p value is informed
your by your
observations
significance level d
a pre defined cutoff value chosen by the experimenter
Q is not determined observations
by your
A I error i.e the probability that
Type you
reject Ho when Ho is actually true
ALWAYS report your actual p value if it is available
