f1-l lU'-hE..W Q'-.

lt:'S,1C)~j
Exercises
5.1 Parameter or estimate? Say wheth er each of the boldface nurnhcrs is either a parameter or an
estimate. !H int: Befo re prov iding an opinion, clearl y state the implied research question.]
(A) A study of surv ival in 1225 newly diagnosed b1·easl cancer cases lind5 that survival varies greatly by £5Tt/t'\A<f. S
stage of di agnosis. The average seven-year survi va l rates for Stage l breast cancer was 92%; the Stage II
survi val rate was 71 %; the stage 11 I survival rate was 39%; and the stage IV survival rate was 11 %.
(B) A review o f' divo rce reco rd s for a cou nty in Connecti c ut indicates that the marriages that end in divorce
last an average ol'72 months. 1)£P840 S Orl \JS c.; cc.
IF \J'S 61) T() (i8-\£~LI l i 1sE.'iOl'4 D (or-1t.J Tl [Link]~ E STI N>~lih
(C) For U.S. men, the average life expecta ncy is 76. For women, it' s 81. [Life expectancy is based on data ·
from t he Nat io na l Death Index; deaths a re unif<:irmly recorded in the U.S.] Pt-..l<..J'.i18cR3

~ ampling experiment. Let us conduct a small. artificial sampling e.'(periment that will demonstrate
~ is meant by the sampling di stributio n for a mean (SOM). A tiny fi nite population consists of the
fo llowing va lues: ( 1, 3, 5. 7, 9) .This population has ~t ~· 5 and c; cc 2 .828. ll
Db i-.1 E 1"1 (.Lf>-"::, S 3/ 2- ~ 6 -
(A) Construct a ste mplo t of the populatio n. S 'Et 61,lL-' 1'1£ tL.£',/ Fo\Z F<JL.L. ~~PoNSE
(B) List of all possible unique samp les each of 11 = 2 frorn th is population. For exa mple, the first sample can
be { 1, 3). The second such sample can be {I, 5 J. T here 10 such unique sampl es.
(C) Calculate the mean of each of the 10 sampl es .
(D) Construct a stcmplot o rthe IO sample means. 1 his is an experimental samp ling di stribution of the
mean (SDM).
(E) ls the SDM more or less norma l shaped than th<: population'? This is a demonstration of the central
limit theorem, which slates that the SDM tend to be mo1·e normal than the populatio n. Thi s is particularly
true with when the sample size is large. Although the sample size in this experi ment is small , you should
still a trend toward the nor mali zati on or the SDM.
(F) Calcu late the mean of the SDM (i.e .. the mean of th e means). Is the mean of the SOM equal to, Jess
than, or greater than th e mean oflhe po pu lation? This is a demonstration of the unbiasedness of the x-bar as
a reflectio n of ~l.
(G) Use yo ur calculator to calculate the standard deviation of the SDM O.e., the standard deviation of the
means is analogous to the SEA{). Report both the sand cr derived by your ca lculator. Are these standard
deviat ions equa l to, less than, or greater than a in population (which is 2.828)'1 This demonstrates that the
SEM is smaller than the CT.

5.3 Health survey. A survey takes a simple random sample of 500 people fro m a town of 55,000. On the
average, it find s 2.30 health probl ems pe r person (standard deviation c= i .65 ). Say whether each of the
following statements is true or false. Explain your reasoning in rach instance.
(A) The standard dev iati on of the mean ( i.e .. the SEAf) is O.C738. ~IJE ,
(B) T he 95% confidence interval for the average number of health problems in th sampl is (2.16, 2.44)._A\L.S C
(C) The 95% confidence in terv a l fo r the ave rage number of health r,roblcms in th o wn s (2.16, 2.44 ). flv(
(D) While th e num ber or health problem s in the population iS [Link] normal ly distributed, according to the Tfl_vE.
central limit theorem (Exercise 5.2E) it reaso nabl e to assume that the sampling distribution of the mean -
(SDM) will be normal.

5.4 lab instrument. Meas urements from a particular iab instrument has known standard deviation CJ = I 0.
A lab ass istant takes 4 measure me nts wit h this instrument an.J calculates the mean of the four samples .
(A) Explai n the advantage using th e average of th e four measurements rather th an each individual
measurement as the true value for a sample. /YJf:I\.J.5 All-£ /VloR..€ P1utc. 1S€' KEFL£C7torJS oF A ~tktJ. 1N01VtDl1AL
l'YIFA 5uR.cP1 &175.
(B) Calculate the standard error o f the mean when r. = 4. SUr).::: ID =- ~
/R
(C) How many times mu st the assistant repeat the measurement to redw:c the standard error of the mean to
2.5? [Hint: Rearrange the formula SEM c= a I ...Jn to get 11 = (cr / SliM)2 to determine the required sample
s ize]. fi ) ,._
n l/
:=. o / 2 . 5 == I fo

C :\ LJ scrs\B. Burl Ci erst man \Dropbo x\S tatPr imer\estirn atio n..,;:xcrcises .docx Page 2 of4
 so~,:~ 1-
..,30
== 1. 2. 1'B'
5.5 Pharmacy survey. A survey of30 pharmacies found that the average cost ofa month's supply ofa ~
particular drug was $33 . The standard deviation of the cost of the prescription is assumed to be $7. / 33± \. ~t · 1- 2 7
(A) Based on the data, determine the mean price µ with 90% confidence for the Rx from all pharmacie:f" : 'c3:t J. · fO ~
(B) N ow determme. µ wit. h 95°/ nfid
, o co 1 ence. = 3~;!:" \-'ib · (. 11-15'::: 33-:t. 2. -')o =(7>cl~O 35.50
) - '3o. 'r<>I 3S. {()
(C) What is the margin of error for 90% confidence? What is the margin of error for 95% confidence? f, l-10 f -'.2-S"c)
(D) A pharmacist reads that a 95% confidence interval for the average price is $30.50 to $35.50. Asked to
explain the meaning of this, the pharmacist states "95% of all pharmacies sell the drug for between $30.50
and $35.50." ls the pharmacist correct? Explain your response. 'wrzo.:>€, . \1-'< C.1 (',PP1.,,~ T.:i ~ / t,10< nJc,vrl)Vf<.L. \/AUJF 'S .

5.6 Laboratory scale. A manufacturer of a laboratory scale with a digital readout claims the scale is
accurate to 0.0015 ofa gram. You read the fine print in the documentation that accompanies the scale and
fmd that the manufacturer means to say that the weights will have standard deviation er= 0.0015 grams.
You are willing to assume that the measurement error varies according to a normal pdf with a mean equal
to the true weight µ of an object. Duplicate measurements on an object produce weights of 24.31 and 24.34
g_!:.ams. Estimate the true weight of the objects with 95% confidence. q S-°i'. a Fol;«-:=. ).t-{. ;25:1: 1. q(. . . 0010 <:, -:z.)
'1< ~ 21 . 315 ('l,t>.. Q..L.<:.....i,...;,,"oe.') S£ffl.::: ~ : . ~oro6 l -= '.lt.f.32.51: . 002.1 =(°:'4.,z.3,1-f. '!>r!J
5.7 Hemoglobin survey, sample size. Hemoglobin levels in 11-year old boys have a normal distribution
with unknown mean µ and er = 1.209 g/dl. How large a sample is needed to estimate µ with 95%
confidence and a margin of error of0.5? n =( .q<o .-; ·20 q)'1-:::
2 2 . 'i lo -r-4> 3 z.
5.8 Sugar consumption survey, sample size. A nutritionist is willing to assume that the standard deviation
of the weekly sugar consumption in children is 100 grams. How large a sample is needed to calculate a
95% confidence interval for ~ so that its margin of error is no greater than IO grams?
n-=(::q~;~ S2.-)'- ::
0 3'3'-/. r(o--..:r 38S
5.9 Birth weights. Assume the birth weights of newborns in a population of full-term infants varies
normally with er= 2 pounds Random samples of size n are selected from this population of birth weights of
full term infants. Calculate 95% confidence intervals forµ based on each of the following samples. State
the margin of error in each instance. .1.f_ffl rYl (LC.L, V c L)
(A) n = 81, sample mean= 6.2 pounds . ?.. 22.2. o . t.f~(,, ( S. 151 '=,. 6)
(B) n: 36, sample mean_= 7.0 pounds
(C)n-9, samplemean - 5.8pounds
. ?, 3 3 3 o. e,,5~ ((,. 5-; 1--1-)
, ~ c;.(, 7 /. 307 ('-f.,; 1-1)
1
5.10 AIDS risk factor study. A national survey of AIDS risk factors based on a random sample of2673
heterosexual adults found the 170 individuals reported two or more sexual partners in the past 12 months.
(Assume the responses are truthful.)
(A) Describe the population to which inferences will be made. Describe the parameter to be estimated. vs: l·h-n:1Z.,S€"f.->~LS @
(B) Calculate the point estimate for the prevalence of this risk factor. f :-
~ ;: . 0 '=, 3 '=. ''""' of- 5 \J~" lc."'f
(C) What is the parameter of interest?~ sample method be used to calculate confidence . ., ,,
intervalsfortheprevalenceofthisriskfactor? r Ylft:: ;J.1•73 · . Ole,3.b · C.1- . 01:,3!.. ""(5 -r Ok.
(D) Calculate a 95% CI for the prevalence of multiple sexual partners rn this population,(-::> SEP ;: /: of.3 fo • (I - . 01,1 '-)
(E) Calculate a 90% CI for the prevalence. oo
!-. 0 1o 3 c.1: \. q c., . . q 't 1- '{ :,. lo. 'f 3
~. o{:;~i;.-t: I. t..'f .. o ot.J1-1-= .o~ ~kt. . oo "f =!- =(o~, · o-1) ::- . C>s..~b -t. . ooq 3 :::.. ( .o5"1ro 13 ::. . oo 41- 2-0

~1 5.11 Patient preference. A doctor finds that 7 out of 8 of her patients preferred to have a "fmger stick"
blood sample compared to a venous blood draw. Can you use the large sample method to determine the
proportion of patients preferring the finger-stick method with 95% confidence using this data? Justify your
"pf :;: .
response with the "npq rule."

5.12. Sample size for a proportion. You are planning a study that wants to estimate a population--
1i · .'815 · Cl - .'i1S) :: o.g-'7'5' ---7 tJ <J7 OK.

proportion with 95% confidence. How many individuals do you need to study to achieve a margin of error

is begun so yot ume

ofno greater than .06? A reasonable estimate for the population proportion is not available before the study
p* = 0.50. How large a sample is needed to cut the margin of error to . ?
1
f7::: c,...fv'k"' - /.C/b.,__ · ~5" ·. ,.;- .::- 2 &,<a .1f;--? .).fo1 (l.::: r.H, -z.. _ ·:2.· s
rY1'2.-- • Ob .b

= {Ofo Tri -'1' /ofo8'

C:\Users\B. Burt Gerstman\Dropbox\StatPrimer\[Link] Page 3 of4

 DD£ ·'f/t
K£'/

5: Introduction to Estimation
Review Questions
I 1. Define "statistical inference." Gfl-1 ERh t...t e.1t-l G, r-~ 91\Th TD
2.
Li"R()ER v~~<::>ff\51()"'1 w I (i\.L ('Jt..ATE\) C~Tt.;f~j
Name the two primary forms of statistical inference. \'.:.S11Mr-.11(),-.J i f'.ll-1:1:5
3. Name the two forms of statistical estimation. Po , r-.1, £'=>,tMA. "Tlov-l) <.or,1f,c;>t1--l (. c ,r-i,8!'[Link]
4. Although parameters and estimators are related , they are not the same. List ways they differ.
5. What does it mean when we say that the sample mean is an unbiased estimator ofµ? E.51 r<'l"'-°'E'S FR.c,"' <JP..,~ f "'Ve
1

!LJ',•,l\)OM "'"R.1 fl.. ~t..ES • ""\t"'l<'-ofl-0.S

6. A particular random sample of n observations can be used to calculate a sample mean. We can J t'J(}(,
determine the characteristics of the distribution of means derived by other samples of the same size ·
taken from the same populations under identical conditions without actually taking additional samples.
The theorem that postulates that the sampling distribution of the mean (SDM) will tend to be normal
(Gaussian) is called the central limit theorem . This SDM will have a mean ofµ and standard deviation
that is equal to SE" I"\ ::: cT/ {o [formula].
7. The standard deviation of the SDM is called the-= $~[~ ({"\
_ _ _ _ _ _ _ __
8. When the SDM is normal (Gaussian), 95% of the sample means will fall within 1.96 Sf fl)$
_ _ _ _ _ ofµ.
9. ~ or false? The square root law says that the SEM is inversely proportional to the square root of
the of the sample size.
10. Match the definitions below with each of the following terms: alpha , statistical inference, parameter,
confidence interval, sampling distribution of the mean, standard error of the mean
(a) A numeric constant that describes something about a pmf, pdf, or statistical population. F'l'sR.l'[Link]..
(b) The act of generalizing from the data to a natural phenomenon or population with calculated degree
of certainty $,;,. ,1.'.:,,1C.1>-L lNfEJ2-£N CE.
(c) The hypothetical frequency distribution of all possible sample means based on samples of size
n taking from the same population under identical conditions . ..S,',.ff\PL1rJ6 Dt":;T~t~VT,a i-.1 cF .f'\8'"1 Cs t>n)
(d) The chance researcher is willing to take in not capturing the parameter with an estimate. al..
(e) The standard deviation of the sampling distribution of the mean. J"Tf'sl'H),t,-12..() E~([Link]:.. of' 11'\tt-,,..J (S El""\)
(f) An interval that is calculated so that it has known probability of capturing the parameter. Co,-1f"1 '.>o-1 CE:
11. What percentage of 95% confidence intervals forµ will fail to captureµ? 5 1.,
12. What percentage of (l-alpha)l00% confidence intervals forµ will fail to captureµ? cl.· 100 %
13. The CI for the mean seeks to capture the Pov0u,..-c-1oi-l mean. [M/C: (a) sample (b) population].
i 4.A confidence interval is 3 ± 1.2. The 3 in this statement is the \7<>1.i, C:S11 'a~d1he 1.2 is the
M~KQIN 0~ ~ ( ) ~

.is. ~ the point estimator of ~ -

J 6. '-fZ- is the point estimator of p.
17. j is the point estimator of _j[_. Ol~?'ifR.
18. The formula for calculating confidence intervals for p presented in this;.should only be used in large
sam ples as tested by this " rule." np4-- ?,,- 5" ~ r~,s,<::>i--.1 or"p p,.5 ts,,1Y1,..,E. OF p p~~
~~~~~1 - - - 19___ W_h....,a'-t-d~ EP stand for? What does it measure? How do we estimate it (i.e., what's the formula)?~S[P~ I"\
f'f2<0{7@"1t<:1c-1 20. The sample size requirements for estimating a mean with confidence depends on these things. rn;, (J. l -d-..
21. The sample size requirements for estimating a proportion with confidence depend on these things. ~ p;,t:: )-c(.
I
22. Increasing the sample size will I t-1 C::: 'j:....Ep,._ S£ the precision of the estimate by i)R-R..t----P..S · · - - . ing J
the confidence interval length .
23. Requiring increased confidence will IN C.f...~S E the margin of error.

Macintosh HD:Users:buddygerstman:Dropbox:StatPrimer:[Link] Page 1 of 1