Chapter 15Frequency Distribution, Cross-Tabulation, and
Hypothesis TestingTrue/False Questions1. In a frequency
distribution, two variables are considered at a time.(False,
moderate, page 4272. !"e valid percentage column in a
frequencydistribution represents percentagescalculated by e#cluding
t"e cases wit" missing values.(!rue, moderate, page 42$%. Frequency
data may be used to construct a "istogram, or a vertical bar
c"art.(!rue, easy, page 42$4. !"e most commonlyusedstatistics
associatedwit"frequencies aremeasures oflocation, measures of
variability, and measures of occasion. (False, moderate, page
4%&'. (ean, median, and mode are measures of
variability.(False, moderate, page 4%&). If t"e entire sample
is c"anged by adding a fi#ed constant to eac" observation, t"ent"e
mean, variance, and median c"ange by t"e same fi#ed amount.(False,
moderate, page 4%&7. (easures of location or central tendency
include t"e mean, median, and mode.(!rue, easy, page 4%&*. !"e
mode represents t"e "ig"est pea+ of t"e distribution. (!rue,
difficult, page 4%1$. If a variable is measured on a nominal scale,
t"e median s"ould be used.(False, moderate, page 4%11&. !"e
mean is sensitive to e#tremely small or e#tremely large values
(outliers.(!rue, moderate, page 4%111. If all t"e values in t"e
data are multiplied by a constant, t"e range is multiplied by
t"esame constant.(!rue, difficult, page 4%112. If all t"e data
points are multiplied by a constant, t"e interquartile range is
multipliedby t"e same constant.(!rue, difficult, page 4%11$$1%. !"e
difference betweent"e meanandanobservedvalue is calledt"e
standarddeviation from t"e mean.(!rue, difficult, page 4%114. !"e
variance can never be negative.(!rue, easy, page 4%11'. ,"en t"e
data points are scattered, variance is small.(False, difficult,
pages 4%1-4%21). ,"en calculating t"e standard deviation, we divide
by n because t"e sample is drawnfrom a population and we are trying
to determine "ow muc" t"e responses vary fromt"e mean of t"e entire
population.(False, difficult, page 4%217. In a symmetric
distribution, t"e values on eit"er side of t"e center of t"e
distributionare t"e same, and t"e mean mode, and median are
equal.(!rue, easy, page 4%21*. !"e tendency of t"e deviations from
t"e mean to be larger in one direction t"an in t"eot"er is called
s+ewness. (!rue, difficult, page 4%2 1$. If t"e +urtosis is
positive, t"e distribution is more pea+ed t"an a normal
distribution.(!rue, moderate, page 4%22&. !"e first step in
"ypot"esis testing is to select an appropriate statistical
tec"nique andt"e corresponding test statistic.(False, moderate,
page 4%%21. !"e null "ypot"esis is always t"e "ypot"esis t"at is
tested.(!rue, easy, page 4%422. !"e alternative "ypot"esis is a
statement t"at some difference or effect is e#pected.(!rue, easy,
page 4%42%. !"enull
"ypot"esisreferstoaspecifiedvalueoft"epopulationparameter,not
asample statistic.(!rue, moderate, page 4%424. In commercial
mar+eting researc", t"e two-tailed test is used more often t"an a
one-tailed test. (False, moderate, page 4%'2'. !"e one-tailed test
is more powerful t"an t"e two-tailed test. (!rue, easy, page
4%'2&&2). !"e test statistic measures "ow close t"e sample
"as come to t"e null "ypot"esis andoften follows a well-+nown
distribution, suc" as t"e normal,t, or
c"i-squaredistribution.(!rue, moderate, page 4%'27. !ype II error
occurs w"ent"e sample results lead tot"e re.ection of t"e
null"ypot"esis w"en it is in fact true.(False, moderate, page
4%'2*. !"e probability of !ype I error ( is also called t"e level
of significance.(!rue, moderate, page 4%'2$. For a given level of ,
increasing t"e sample si/e will decrease , t"ereby increasingt"e
power of t"e test.(!rue, moderate, page 4%)%&. In determining
t"e critical value of t"e test statistic, t"e area to t"e rig"t of
t"e criticalvalue is eit"er or 02. It is for a one-tail test and 02
for a two-tail test.(!rue, moderate, page 4%7%1. If t"e probability
associated wit" t"e calculated or observed value of t"e test
statistic(!1 234 is less t"an t"e level of significance (, t"e null
"ypot"esis is re.ected.(!rue, moderate, page 4%7%2. If t"e
calculated value of t"e test statistic is greater t"an t"e critical
value of t"e teststatistic (!1 25, t"e null "ypot"esis is
re.ected.(!rue, moderate, page 4%7%%. In tests of associations, t"e
null "ypot"esis is t"at t"ere is no association between
t"evariables (6&78.is 9:! related to8..(!rue, difficult, page
4%*%4. Intests of differences, t"enull "ypot"esis is t"at t"ereis
adifference(6&78.isdifferent t"an8..(False, difficult, page 4%*
%'. 3cross-tabulationis t"e mergingof t"e frequencydistributionof
twoor morevariables in a single table to "elp us to understand "ow
one variable relates to anot"ervariable.(!rue, moderate, page
4%*%). !"eintroductionofat"irdvariableclarifiest"einitial
association(orlac+ofitobserved between two variables and can result
in four possibilities.(!rue, difficult, page 44&%7.
2ross-tabulation e#amines associations between variables, not
causation.(!rue, moderate, page 44%2&1%*. !"e null
"ypot"esis,6&, w"enusingt"ec"i-squarestatistic, is t"at t"ere
is noassociation between t"e variables. (!rue, easy, page 444%$.
,"en using t"e c"i-square statistic in cross-tabulation, t"e null
"ypot"esis (6&
willbere.ectedonlyw"ent"ecalculatedvalueoft"etest
statisticisgreatert"ant"ecritical value of t"e c"i-square
distribution wit" t"e appropriate degrees of freedom.(!rue,
difficult, page 444 4&. !"e c"i-square statistic can be
estimated on counts of data or on percentages.(False, difficult,
page 44'41. ,"en variables are perfectly associated, t"e p"i
coefficient assumes t"e value of 1.(!rue, easy, page 44'42.
,"enmeasuringt"estrengt"ofassociationwit"t"econtingencycoefficient,
t"ema#imum value of 1 is ac"ieved w"en t"e variables are perfectly
associated.(False, difficult, page 44)4%. 4ambda assumes t"at t"e
variables are measured on an ordinal scale.(False, difficult, page
44)44. 3nasymmetrical valueof1"appensw"eneac"independent
variablecategoryisassociated wit" a single category of t"e
dependent variable.(False, difficult, page 44)4'. ,"en conducting
cross-tabulation analysis in practice, if 6& is re.ected, t"e
pattern oft"e relations"ips"ouldbe
interpretedbycomputinganappropriate statistic (p"icoefficient,
contingency, 2ramer;s