, Lecture 1 notes
Chap a :
Hypothesis Testing
S
j -
Mo
-
2 =
normal
Remembe r. . . Assume
6/5n distribution
Inference
Inferenceabout known/
with O ↓
a
one pop
unknown (s)
& n(30 + t
=
parameters 02
i
OR
Parametric us Nonparametric tests n(30 >
z 3 Mo
-
- -
=
↓ &
S/ Un
small samples
·
big sample
dist dist
·
nor m ·
no
assumption
3 ①
·
quantitative data ·
categorical or quantitative data There are
Sign Test
↳ ↓ non-parametric Mann-Whitney -
tests ②
we have parameters we do not have parameters ; Wilcox on
04) and distribution
(n ,
i no
③ Kruskal Wallis
Test
normal distribution
"
distribution-free statistics"
Two
categories
~
↓
(
-
~
less than more than
median median
# -
both have fixed
prob of 0 . S
focus on mean
focus on
median
/Post
Ho
to:median
: M =
q
Ha :
u +
q
=
observation) hyp value
↓
- sign
① The
Sign Test +
·
compare observation to hypothesized value of population median-D observation <hyp value
↓
of recorded
Test stat
signs
↳
·
= no .
+
sign
-
observation =
hyp value
Example of ↓
test
sign significance
:
S %
eliminate observation and proceed
with smaller
sample
NB : P-value must be XI
in a two tail test
OR
& probabilities P-value xp(x), 7)
=
2
-
x = 7
-S
Excel
= 2x(1-Binom dist (6 .
, 10 0 . 5, TRUE)]
,
= 0 3435 OR
p .
= 1 - p(x <6)
Chap a :
Hypothesis Testing
S
j -
Mo
-
2 =
normal
Remembe r. . . Assume
6/5n distribution
Inference
Inferenceabout known/
with O ↓
a
one pop
unknown (s)
& n(30 + t
=
parameters 02
i
OR
Parametric us Nonparametric tests n(30 >
z 3 Mo
-
- -
=
↓ &
S/ Un
small samples
·
big sample
dist dist
·
nor m ·
no
assumption
3 ①
·
quantitative data ·
categorical or quantitative data There are
Sign Test
↳ ↓ non-parametric Mann-Whitney -
tests ②
we have parameters we do not have parameters ; Wilcox on
04) and distribution
(n ,
i no
③ Kruskal Wallis
Test
normal distribution
"
distribution-free statistics"
Two
categories
~
↓
(
-
~
less than more than
median median
# -
both have fixed
prob of 0 . S
focus on mean
focus on
median
/Post
Ho
to:median
: M =
q
Ha :
u +
q
=
observation) hyp value
↓
- sign
① The
Sign Test +
·
compare observation to hypothesized value of population median-D observation <hyp value
↓
of recorded
Test stat
signs
↳
·
= no .
+
sign
-
observation =
hyp value
Example of ↓
test
sign significance
:
S %
eliminate observation and proceed
with smaller
sample
NB : P-value must be XI
in a two tail test
OR
& probabilities P-value xp(x), 7)
=
2
-
x = 7
-S
Excel
= 2x(1-Binom dist (6 .
, 10 0 . 5, TRUE)]
,
= 0 3435 OR
p .
= 1 - p(x <6)
