AP Statistics Confidence Intervals Summary Sheet

Name Conditions Calculations & Interpretation

Finding Sample Size 𝑝̂ (1 − 𝑝̂ )

𝑝̂ ± 𝑧 ∗ √
𝑛
1) Random sample taken? 𝑝̂(1−𝑝̂)
MOE = 𝑧 ∗ √ I am ___ % confident that my interval from ____ to
2) Independence within the group? 𝑛
1 sample
 Check the 10% condition: 10(𝑛) ≤ 𝑁 ____ captures the population proportion of all
z-interval
3) Approximately normal sampling distribution? (𝑧 ∗ )2 (𝑝̂ )(1 − 𝑝̂ ) (context).
for 𝑛=
 Check that large counts condition is met. (𝑀𝑂𝐸)2
proportions
o 𝑛𝑝̂ = # ≥ 10 The confidence interval gives a set of plausible
o 𝑛(1 − 𝑝̂ ) = # ≥ 10 If 𝑝̂ is unknown let 𝑝̂ = 0.5 values for the population proportion of (context).
Always round up Because # is/is not contained in the interval, it
is/is not a plausible value. We have good reason to
believe/doubt the claim (context).

For an observational study:

1) Random samples taken?
2) Independence within each group? 𝑝̂1 (1 − 𝑝
̂)
1 𝑝̂2 (1 − 𝑝
̂)
2
(𝑝̂1 − 𝑝̂2 ) ± 𝑧 ∗ √ +
 Check the 10% condition: 10(𝑛1 ) ≤ 𝑁1 and 10(𝑛2 ) ≤ 𝑁2 𝑛1 𝑛2
3) Independence between the groups?
4) Approximately normal sampling distribution? I am ___% confident that my interval from ___ to
Check the large counts condition for both groups. ___ captures the population difference in the
2 sample  𝑛1 𝑝
̂1 = # ≥ 10 𝑛2 𝑝
̂2 = # ≥ 10 proportions of all (context for both groups).
z-interval  𝑛1 (1 − 𝑝̂) = # ≥ 10 𝑛 (1 − 𝑝
̂)
1 2 2 = # ≥ 10
for
Because the interval does not contain 0 there is
proportions
For an experiment: convincing evidence that the true proportion of
1) Treatments randomly assigned? (context) is different for (context).
2) Approximately normal sampling distribution?
Check the large counts condition for both groups. Because the interval does contain 0 there is not
 𝑛1 𝑝
̂1 = # ≥ 10 𝑛2 𝑝
̂2 = # ≥ 10 convincing evidence that the true proportion of
 𝑛1 (1 − 𝑝 ̂) = # ≥ 10 𝑛 (1 − 𝑝
̂) (context) is different for (context).
1 2 2 = # ≥ 10

D. Champion Spring 2018

 AP Statistics Confidence Intervals Summary Sheet
Name Conditions Calculations & Interpretation
1) Random sample taken? 𝑠𝑥̅
𝑥̅ ± 𝑡 ∗ 𝑑. 𝑓. = 𝑛 − 1
2) Independence within the group? Finding Sample Size √𝑛
 Check the 10% condition: 10(𝑛) ≤ 𝑁
𝑠 I am ___ % confident that my interval from
3) Approximately normal sampling distribution? MOE = 𝑧 ∗ 𝑥̅𝑛
√ ____(units) to ____(units) captures the population
 Population is normal so the sampling distribution is
1 sample normal OR mean (context).
(𝑧 ∗ )(𝑠𝑥̅ ) 2
t-interval  𝑛 ≥ 30 so CLT applies so the sampling distribution is 𝑛=( )
for means approximately normal OR 𝑀𝑂𝐸 The confidence interval gives a set of plausible
 Given a histogram, stem plot, or dotplot: Graph values for the population mean of (context).
degrees of We use z* to approx. t* Because # is/is not contained in the interval, it
should be symmetric or slightly skewed with no since we don’t know the is/is not a plausible value. We have good reason to
freedom is apparent outliers.
n−1 degrees of freedom. believe/doubt the claim (context).
 Given ONLY raw data:
o Sketch boxplot: Graph should be symmetric -OR-
Always round up There is convincing evidence that (context). All of
or slightly skewed with no outliers.
o Sketch a Normal Probability Plot – Graph the plausible values in the interval for the true
should appear linear. mean (context) were greater than/less than
(# units).
For an observational study:
1) Random samples taken? 𝑠12 𝑠22
2) Independence within each group? (𝑥̅1 − 𝑥̅2 ) ± 𝑡 ∗ √ +
𝑛1 𝑛2
 Check the 10% condition: 10(𝑛1 ) ≤ 𝑁1 and 10(𝑛2 ) ≤ 𝑁2
3) Independence between the groups? 𝑑. 𝑓. = (𝑛1 − 1) or (𝑛2 − 1) whichever is smaller.
2 sample
t-interval 4) Approximately normal sampling distribution?
 BOTH Populations are normal so the sampling distribution is normal OR I am ___ % confident that my interval from
for means
 BOTH 𝑛 ≥ 30 so CLT applies so the sampling distribution is approximately normal ____(units) to ____(units) captures the population
OR difference in the means (context for both groups).
degrees of  Given BOTH histograms, stem plots, or dot plots: Graphs should be symmetric or
slightly skewed with no apparent outliers. Because 0 is not in this interval, this interval does
freedom is
 Given ONLY raw data: support (context) have different (context).
(𝑛1 − 1)
or o Sketch BOTH boxplots: Graphs should be symmetric or slightly skewed with no
outliers. Because 0 is in this interval, this interval does not
( 𝑛2 − 1)
o Sketch BOTH Normal Probability Plots – Graphs should appear linear. support (context) have different (context).
whichever
is smaller For an experiment:
Because all the plausible values in this interval are
1) Treatments randomly assigned?
positive/negative, this interval supports the belief
2) Approximately normal sampling distribution? Same conditions as #4 above.
that ….

D. Champion Spring 2018

 AP Statistics Confidence Intervals Summary Sheet
Name Conditions Calculations & Interpretation
For an observational study: 𝑠𝑑
1) Random sample taken? 𝑥̅𝑑 ± 𝑡 ∗
√𝑛
2) Groups are not independent – However, the data pairs should be independent of one 𝑑. 𝑓. = 𝑛 − 1 𝑤ℎ𝑒𝑟𝑒 𝑛 𝑖𝑠 𝑡ℎ𝑒 # 𝑜𝑓 𝑝𝑎𝑖𝑟𝑠.
another.
Paired 3) Approximately normal sampling distribution? I am ___ % confident that my interval from
t-interval  Population of differences is normal so the sampling distribution is normal OR ____(units) to ____(units) captures the population
 𝑛 ≥ 30 so CLT applies so the sampling distribution for the differences is mean difference (context).
degrees of approximately normal OR
freedom is  Given a histogram, stem plot, or dot plots of differences: Graph should be Because 0 is not in this interval, this interval does
n–1 symmetric or slightly skewed with no apparent outliers. support (context) have different (context).
where n  Given ONLY raw data:
is the o Sketch boxplot of differences: Graph should be symmetric or slightly skewed Because 0 is in this interval, this interval does not
# of pairs with no outliers. support (context) have different (context).
o Sketch Normal Probability Plots of differences – Graph should appear linear.
For an experiment: Because all the plausible values in this interval are
1) Treatments (or treatment order) randomly assigned? positive/negative, this interval supports the belief
2) Approximately normal sampling distribution? Same conditions as #3 above. that ….
Interpreting Confidence Level Percent Statement (i.e. What does it mean to be __% confident?)

If we take many samples of the same size from (context about population) and construct a __% confidence interval for each sample, about
__% of the intervals will capture the actual (population proportion of …. or population mean … of …).

NOTE: The focus is on the method – not the interval values … you can write statement before finding your confidence interval.
Finding z* Finding t*
InvNorm(area,0,1) 𝐶%
Critical ± 0.50 = 𝑒𝑑𝑔𝑒 𝑣𝑎𝑙𝑢𝑒𝑠 invT(area,d.f.)
Values 𝐶% Area: edge value (decimal) 2
± 0.50 = 𝑒𝑑𝑔𝑒 𝑣𝑎𝑙𝑢𝑒s Area: edge value (decimal)
2 Mean: 0
df: _____
Standard Deviation: 1 Degrees of freedom = 𝑛 – 1
Normality When n is sufficiently large (𝑛 ≥ 30), the sampling distribution of 𝑥̅ is approximately normal, even if the population distribution is NOT
Notes normal. When the population is known to be Normal, the sampling distribution of 𝑥̅ is also normal for ANY sample size.

D. Champion Spring 2018