Name: Hour: Date:

Can you taste the rainbow?

Many students claim that they can taste the different colors of Skittles. Today we will conduct
an experiment and perform a significance test to see if students really can “taste the rainbow”.

Collect data: 97
How many correct? ______ 150
How many total? ______ Results will vary

true proportion of correct

Parameter: p
D= Fo
-

STATE: Statistic: =
0.647
guesses for all skittles
Hypotheses: Significance level: 5% ( a = 0.05 )
Ho : p = 0.20

Ha :p > 0.20

PLAN: Name of procedure: one

sample z test for p
Check conditions:
Random : random sample of skittles generalize to all skittles
→
so we can

not I lol 150 E. To (all skittles)

.
: so sampling without replacement is 0K
→

sampling distribution of Fis

g.
" T
Large counts 15010.201=30210 15010.801=120210
us : so the -
,

DO: General Formula: standardized statistic parameter approximately Normal

-

-
- -
-

test statistic SD
Specific Formula:
z =
PI Use p not p
fTtF Picture (of the Normal curve):
,

Work: n

⇐
0.647--0.20
JE ' so
Noo
MMMM
ago
-0.647
.

Test statistic:
2- = 13.68 -2=13.680
P-value:
P-value
CONCLUDE: Based on the P-value, what conclusion do you make?
ret
probability
pilot {
Assuming Ho is true ( p -0.20) there is approximately O
were
an
,

greater purely by Because

'

of getting chance
a
p of 0.647 or .

qgigsig.es?qf0so.05-TwerejectHdeis'annd
we do have convincing evidence
that this class can taste the rainbow ( p 0.28 >
.
 Name: Hour: Date:

Significance Tests: The 4 Step Process

Important ideas: standardized statistic parameter
-

STATE : Parameter, statistic ,

hypotheses o
= -

test statistic
-

,
SD
PLAN : Name procedure check conditions ,
n

Do : General + specific formula work

Test statistic ,
P -
value
,
Z=fFgg two-SIDED
Ha :p # null value

CONCLUDE : Interpret P value -

.
Decision -1 conclusion(context) Multiply P value +2
-

Check Your Understanding

1. Sometimes parents and grandparents like to recount how difficult life was when they were kids, such as
having to walk 10+ miles to school (in the snow, uphill both ways). A random sample of 180 teenagers
were selected and 40% had heard stories from their parents or grandparents about how difficult life was
when they were kids. Do these data provide convincing evidence at the α = 0.05 significance level that
the proportion of all teenagers who have heard stories from their parents or grandparents about how
difficult life was when they were kids differs from 0.50?
p true proportion of all teens
-

STATE: Parameter: who have heard stories Statistic: 0.40 D=

Hypotheses: to :p 0.50 Significance level: 0=0.05
-

Hai p# 0.50
PLAN: Name of procedure:
one sample z test for p
Check conditions: "
of 180 teenagers generalize to all teens
"
Random random :
sample -
so we can

180C Ho ( millions of teenagers) - so

10% :
sampling without replacement is 0k
Large counts 18010.507=90210 18010.507=90210 so the
sampling distribution
: a →

DO: General Formula:

standardized off is approximately Normal
test statistic
.
= statistical
Specific Formula: SD

Picture: 2- = Work:
2=0.40-0.502 Test statistic:
2=-2.680
NC so
.

.
. 037 ) n

pE 180
P-value:

P-value 0.0036×2
-

0.60
2- 2.68 p.ua/ue--O.OO72T
-

=
CONCLUDE:

Assuming Ho is true (p -
0.50 , there is a 0.0072 probability of getting a B of 040
lower OR 0.60
higher purely by
or or
Because 0.0072<0.05 we reject Ho
chance .

and do have
we
convincing evidence that the true proportion of teenagers who
have heard stories differs from 0.50
2. A 95% confidence interval for the proportion of all teenagers who have heard stories from their parents or
.

grandparents about how difficult life was when they were kids is (0.328, 0.472). Explain how the
confidence interval is consistent with, but gives more information than, the test.
The null hypothesis value ( p -0.50) is not contained in the confidence
interval ,
so we would reject Ho The interval also gives other .

plausible values for the true

proportion .