Download Exam 2 Cheat Sheet - Business Statistics | BMGT 230B and more Study notes Business Statistics in PDF only on Docsity! Chapter 8: Random Variables and Probability Models a
Mean: 1 =T.) z + Binomial coefficient ia *'*
oh Grech ett mu (fermen) © “Tinamial Distribntion Formnta
© Pepulation meen a
© Variance: of - Var0
‘oo tle sigma (For standard devia
© Population vance
© The weighted average of the squared deviations (x u)'2 of the vanabl
the variance. the mest scatiered the vrhnes of X ar on average
© Fea disicte manda vata le.
2 Diy = SA pagan
ry of x euecesses in x trials
successes in sample
fr
fis meaa y.
sent on akon im —aeitor 6 Buin gone! Va) = EK y= ROS
Eos uern the expected squared distance Beam .
1 Rinles for means and variances + Varizace~ aga, standard devistion ~ cqprvasiance)
© AddsuaSublaticn, ‘© The binomial sampling distribution for counts can be used when the
ef the population, because the » observations
ample 1 seller than 10%.
ill be nearly independent wien the size of the por
ation isch lager than the sire ofthe sample
6 Miitiptienion Normal Distributions
GX) —akx) ‘Bell chaped. symmetrical, mean = median = mode
1 Yat “eva + Location determined ty the mean, spreed determin
__ of SDaX)=IaresDoQ, © Probability is mzacured by the ares under the curve
© Its erandom variable and a and D are fed muscber, then Standard Normal Distribution, theZ distribution
* . is defined to be 0, Standard deviation is 1
va independent random variables then
REQ) !EC¥)
8 Va Va VartY)
© Thestandarel deviation af a random variable ie SDEX}= SQRTVartx))
© Empurcal ole
© PC ieZ=1)~ 68
5 Perera 9
© Pi3<Z<3)- 997
Binomial Distribution
+ General Characteristics
‘Atal has only two possible outcomes — “success” 9°”
co There isa fixed sumber. 2 of :denticz! trials na
8 Theta of teexporiment me pendent feeb ther + Tran om mrs bain: 2 |
© Theprolclay of «swum. ecco to Ua
(© Tfproprescats ths probability of a sucezss, thea (I-p) = q is the probatilty of a failure Chapter 10; Testing Hypotheses about Proportions
> Satsment1)= nul iypethens ithe esa 2)
> Slam 2)—alenatve hypothesis hen fx)
‘Two Mahods of Hypothes Testing
Chapier 9: Sampling Distributions smd Confidence Titervisls for Proportions 1. Using p-vatues:
9. Roles for Sample Proportion © Cale tet
{Ce tat Fal pvr ec ae el
. mpare to 1 or a/2 (one sided or two sided)
2 te proportion » of sue 2. Using cntical veluss:
6 Scket nope
* Conte the cafidence intra
confidence Interval ofa Sample Proportion fps yw the concentra FIR Hic poste the eonlence mere jet He
; Tete foc popultion pesparon
vy? = 9 Npandng>9 . aN
2 95% confidence > 95% of samples of this size will produce confidence intervals thet cap 2 -
‘ture the true proportion of the population (and we =xpect 5% of sur samples to produce on as
itera st fio apr the tv roporson :
2 Mavs af emor 22 SE(B) Orble te it ofthe CL wich teeta the © Coxfidence tri to et ypotes
Me Fou Sipe tr ype . Te
+ Assumptions nd Condiinns > Define the Ispoteses to ext, andthe reqied
2 Independence Amun: Ae imple stern dependent of each eter? and he required egifieance lee!
> ‘Was the sample randomly generated? 2. Caleolate the value of the test statistic.
9 1084Condion Trampling dae without eplceent then th ample si. mus Be 3 Ti thers beod onthe ebro datz
tvs IO the opin : > Tins the pale bs
» SuetesFalic Conon: Tiesaup sem telaneenrietia tae a (hs? yao ton leo the th th vee, Tee earth
alae the renge te evidence again! the mulliypothesisand the oe cofient you cas Be
+ Sane sz fora dies matin of wor - N\a.5-pIR) = n={=") ped-v9) about your ntespetation
, © The magnitude or size of an effect relates to the real-life relevance of the phenomenon uncovered.
© shenpicanbnonn,wsep= |b reer cert ‘The p-value does NOT asseee the relevance of the effect, aor ite magnitude,
© cae Zable For means use when cis known || «Type Lerror eget and thal Hypothesis w ect tev ncouctly ees ts
Vas (So corso Ponca .
: annie ee iaea | 2 The pebablty of aang x Type ane ih gic level
|] «A Type MH error fail to rejact and the ull hypothesis is false (:ncomectly keep 3 false He)
@ @ Conf Tat al
ig Lev Conf Ley _42 for 2
7 0%
2 The probability of making a Type II exroris labeled 8.
The power ala lest is T=
9 Reducing a seduces the power of 2 test and thus increases 6,
Chapter 11: Confidence Intervals and Hypothesis Tests for Means
© Standard deviation= s/\/n, and is called sampling enraged js puller tha * se if
the population by a factor of vn. =O Xtr%*o / vn
«The mean=. Mean of all ¥ 1s pretty much equal to populatién mean ~~
© Central Limit Theorem: When randomly sampling from any population with mean and standard
deviation s, when 7 is large enough, the sampling distribution is approximately normal: ~ N(u, s/n).
Usually 25, 40 to overcome extreme skewness/outliers.
©) X-bar: sample mean
© Standard Exror of (X-bar) = SEC-bar) = sin
Confidence Intervals
©) X-bar= t".1« SE(Y-bar) statistic —parameter
© same thing as: X-bar +/- ME Standardized lest statistic: a deviation of statistic
Confidence interval: statistic + (critical value) + (standard deviation of statistic)
*® Fora sample of size n, the sample standard deviat
Single-Sample
n—1=degrees of freedom (SO much easier to ph.
Put into L1. Go back into STAT > Cale > 1-Var S Statistic Danaea Deven
© SEM=s/», standard error of the mean Saenen 7
T Distributions vn
ul Sample Proportion [pa=p)
An
afin
«® One-sample f statistic
© [When s is known, the sampling distribution is Nm, s/V/n)]
«® But whens is estimated from the sample standard deviation s, the sampling distribution follows at
distribution Au, s/n) with degrees of freedom n —1
©) When w is very large. s is avery good estimate of s, and the corresponding ¢ distributions are very
close to the normal distribution
® The / distributions become wider for smaller sample sizes, reflecting the lack of precision in estimat-
ing s from s.
One-sample T-Confidence Intervals
©) Margin of error tt =t* sin, leading to xbar +/- the ME
One-sample T-Test
® = Stating the null and alternative hypotheses (Hp versus H,)
© Deciding on a one-sided or two-sided test
Tf X has a binomial distribution with paramcters m and p, then: la
of freedom
P(X =b= (tea — py
\e arve with Table T
By = "Pp
o, = Vipa=P)
Mp =P
fa
_ [pd=py
Vn
If Z is the mean of a candom sample of size n from an infinite
population with mean pt end standard deviation o, then: