STATISTICS
•
Population -
the complete set of things we could
observe
Parameter the
thing want to know
• -
we
•
Sample -
group of observations Over:# n from the
population .
Statistic the
thing we measure in sample
•
-
our
Sampling inherent
property of research sampling error is the difference between
•
error -
an
-
that takes samples from larger the observed sample mean
populations .
and the population parameter
.
Sampling error two samples will be
•
means that no
the same and that the sample statistics will
exactly ,
bounce around even
though the population parameter
the same
stays .
size to
Sampling is
higher in small samples than larger sampling G- sample
•
error error as
samples
wider
•
The
sampling distribution will be for small
samples
•
This means that larger samples provide more accurate
estimates of the population parameter than smaller
samples .
Random samples allow to make
meaningful
-
us
inferences about the population
If we do not
sample randomly our sample may
-
,
not be representative of the population
-
Unrepresentative samples lead to biased statistics .
-
Hypothesis are to do with the population + parameter
Statistical related to the
question
hypothesis is directly
.
-
The alternative 1-11 less
hypothesis
: : → more or
y
HO
jmutuauy
exclusive
hypothesis likely
-
The nun
equally
.
: →
, under Ho calculate the
probability of
possible outcome of
•
can our
we
every
experiment .
'
These always follow known distributions .
1. start the assumption that Ho is true ( even though we
probably don't believe that )
2 out observations t compare the observed data with what would expect
.
Carry our we
if Ho were true .
3 If data ( they wouldn't
very often under Ho ) then
our are
sufficiently surprising e.
g. occur
.
,
we
reject
the null
hypothesis .
The P value :
the data the p value
• we
quantify the
suprisingness of observed
by calculating .
the value is the
probability of observing data at least extreme ours if
•
p as as
Ho were true .
•
If that
probability ( p) is below a certain threshold ,
we
reject Ho
• In
psychology ,
we
usually adopt the threshold of 0.05 ( 5%)
If our
p value less than 0-05
statistically significant reject null
hypothesis
-
:
-
is +
fail
significant reject hypothesis
value null
if 0-05
statistically to
i
-
p
-
is
greater than not ,
+
1 is Ho is
Type I error
:
false positive true true
✓
I
Reject Type
1-to error
false positive (false +ve)
Type
:
2 error
✓
2
Do not Type
er ror
reject Ho ( false ve
-
)
Research statements about that
Hypotheses general the world
:
we
believe to be true
derived observation
from
theory or
°
0
known population +
parameter
about the population from
statistical
Hypothesis formal statements
:
are
which we are drawing our sample .
formal statement that the effect
Null
Hypothesis : we are
looking for does not exist in the
population
formal statement that the effect
Alternative
Hypothesis :
does exist in
we are
looking for the
population
•
Population -
the complete set of things we could
observe
Parameter the
thing want to know
• -
we
•
Sample -
group of observations Over:# n from the
population .
Statistic the
thing we measure in sample
•
-
our
Sampling inherent
property of research sampling error is the difference between
•
error -
an
-
that takes samples from larger the observed sample mean
populations .
and the population parameter
.
Sampling error two samples will be
•
means that no
the same and that the sample statistics will
exactly ,
bounce around even
though the population parameter
the same
stays .
size to
Sampling is
higher in small samples than larger sampling G- sample
•
error error as
samples
wider
•
The
sampling distribution will be for small
samples
•
This means that larger samples provide more accurate
estimates of the population parameter than smaller
samples .
Random samples allow to make
meaningful
-
us
inferences about the population
If we do not
sample randomly our sample may
-
,
not be representative of the population
-
Unrepresentative samples lead to biased statistics .
-
Hypothesis are to do with the population + parameter
Statistical related to the
question
hypothesis is directly
.
-
The alternative 1-11 less
hypothesis
: : → more or
y
HO
jmutuauy
exclusive
hypothesis likely
-
The nun
equally
.
: →
, under Ho calculate the
probability of
possible outcome of
•
can our
we
every
experiment .
'
These always follow known distributions .
1. start the assumption that Ho is true ( even though we
probably don't believe that )
2 out observations t compare the observed data with what would expect
.
Carry our we
if Ho were true .
3 If data ( they wouldn't
very often under Ho ) then
our are
sufficiently surprising e.
g. occur
.
,
we
reject
the null
hypothesis .
The P value :
the data the p value
• we
quantify the
suprisingness of observed
by calculating .
the value is the
probability of observing data at least extreme ours if
•
p as as
Ho were true .
•
If that
probability ( p) is below a certain threshold ,
we
reject Ho
• In
psychology ,
we
usually adopt the threshold of 0.05 ( 5%)
If our
p value less than 0-05
statistically significant reject null
hypothesis
-
:
-
is +
fail
significant reject hypothesis
value null
if 0-05
statistically to
i
-
p
-
is
greater than not ,
+
1 is Ho is
Type I error
:
false positive true true
✓
I
Reject Type
1-to error
false positive (false +ve)
Type
:
2 error
✓
2
Do not Type
er ror
reject Ho ( false ve
-
)
Research statements about that
Hypotheses general the world
:
we
believe to be true
derived observation
from
theory or
°
0
known population +
parameter
about the population from
statistical
Hypothesis formal statements
:
are
which we are drawing our sample .
formal statement that the effect
Null
Hypothesis : we are
looking for does not exist in the
population
formal statement that the effect
Alternative
Hypothesis :
does exist in
we are
looking for the
population
