Modern Marketing Research Concepts Methods and Cases 2Nd Edition Feinberg Solutions Manual Full Chapter PDF

Modern Marketing Research Concepts
Methods and Cases 2nd Edition

Feinberg Solutions Manual
Visit to download the full and correct content document: https://testbankdeal.com/dow
nload/modern-marketing-research-concepts-methods-and-cases-2nd-edition-feinberg
-solutions-manual/
CHAPTER 8
DATA ANALYSIS AND STATISTICAL METHODS: UNIVARIATE AND BIVARIATE
ANALYSES
CHAPTER OUTLINE
The Use of Statistical Methods in Marketing Research pg 390
Statistical models and methods help us understand what is going on under all the inevitable noise
in the real world. All empirical human knowledge is firmly based on statistical evidence.
The most sophisticated data analysis cannot make up for shaky problem definition, bad study
design, improper sampling, poor measurement, sloppy field work, or careless data processing:
“Garbage in, garbage out.”
Overview of Data Analysis Procedures pg 392
To help identify which data analysis technique to use, determine:

• the number of variables to be analyzed together
• whether the data is to be described or used to make inferences
• what level of measurement is available in the variable(s) of interest
Number of Variables to Analyze

How many variables do we wish to interrelate?
• focus on a single variable is univariate analysis
• considering two variables is bivariate analysis
• considering many variables at once is multivariate analysis
Description versus Inference

• Descriptive statistics provides summary measures for the sample data.
• Inferential statistics uses probability theory to make statements about the population
based upon sample results.
Level of Measurement
Statistical techniques must be chosen to match the scale level(s) inherent in the variable(s) being
analyzed or the results will be meaningless.
Overview of Univariate Data Analysis Procedures

A first step in almost all marketing research projects is to perform univariate analyses on some,
or most, available variables. Figure 8.1 presents an overview of statistical techniques appropriate
for univariate data analysis, including the relevant descriptive statistics for “central tendency”
and for “dispersion” and the appropriate tests of statistical inference for each data type.
Figure 8.1 Overview of univariate data analysis procedures (page 394 in text)
© 2013 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly
accessible website, in whole or in part.
Illustration Data Set: Weight, Height, Gender, Age, Likestat for Marketing Research
Students
This section (page 395) continues an ongoing example of data from a Marketing Research class
to evaluate scale types and possible relationships to analyze.
Descriptive Statistics pg 396
Descriptive statistics provide summary measures of the data contained in all the elements of a
sample, particularly measures of central tendency, which describes where the bulk of the data are
located, and dispersion, which describes how ‘spread out’ the data values are around the central
measure.
Measures of Central Tendency

• mean—sum of the values divided by the sample size, used for interval data
• median—the middle value of the data, above and below which half the values fall; used
for ordinal data
• mode—the category of a variable that occurs most often, used for nominal data (can be
bimodal)
A robust measure is resistant to large changes from a few outliers. The median is robust; the
mean is not.
Measures of Dispersion
Measures of dispersion provide a sense of the ‘spread’ of the distribution of the variable.
• standard deviation (used for interval data)
n
(X i − X )2
s= i −1
n −1
• absolute frequencies—number of items in the sample in each category of the variable
(used for nominal data)
• relative frequencies—proportion of items in the sample in each category of the variable
(used for nominal data)
Frequencies are typically presented in a histogram, a bar graph of the relevant cell counts.
Hypothesis Testing pg 400
The Concept of a Null Hypothesis

Hypothesis testing begins with the statement of a null hypothesis (H0) that a population
parameter takes on a particular value. The alternative hypothesis is what we are attempting to
verify. With enough evidence, the null hypothesis can be rejected in favor of the alternative
hypothesis; if not, we conclude that there is insufficient evidence to reject it.
The null hypothesis is never accepted as valid. Examining the sampling distribution of a statistic
reveals whether the sample value is different enough from the proposed null hypothesis value to
have occurred just by sampling error alone. When the alternative hypothesis is directional, a
“one-tailed” hypothesis test is used.
Possible Errors
Type I error (α) means that H0 is true and is rejected (“convicting the innocent”). Type II error
(β) means that H0 is false and is not rejected (“freeing the guilty”). In practice we will have to
balance ‘tolerable’ degrees of the two types of error, α and β.
Table 8.1 Summary of Hypothesis Testing Errors
True Condition
Sample Conclusion H0 is true H0 is false
Do not reject H0 Correct decision Type II error
Confidence level Probability = β
Probability = 1 – α
Reject H0 Type I error Correct decision

Significance level Power of the test
Probability = α Probability = 1– β
Type I Error (α)

The usual starting point in hypothesis testing is to specify the maximum level of Type I error (α)
considered tolerable, called the significance level. When H0 is true, we will occasionally obtain
unlikely values of the sample statistic, from the tails of the distribution; that persuade us that H0
is not true. The likelihood of being correct (not rejecting H0 when it is true) is 1 – α, referred to
as the confidence level of the test.
Type II Error (β)

The probability of rejecting a false null hypothesis is called the power of the test, 1 – β. Type II
error occurs when the value was actually drawn from a different sampling distribution entirely (a
different value of the parameter than the one suggested by H0), and it happens to be fairly likely
under H0, so we don’t reject it. For a given sample size, β increases as α decreases.
The cost to an organization of a Type II error can be much greater than that of a Type I error.
Steps in Hypothesis Testing

1. Formulate a null (H0) and an alternative (H1) hypothesis.
2. Select the appropriate statistical test for the data type.
3. Specify the significance level, α.
4. Determine value of the test statistic for the given α (from table or computer program).
5. Perform the statistical test, yielding a value of the statistic.
6. Compare the value of the statistic calculated in item 5 with the value obtained in step 4. If
the computed test statistic is greater than the tabulated value, the null hypothesis is
rejected at significance level α.
The information in the hypothesis test is summarized in a single quantity called the p-value.
Inferential Statistics pg 404
Choosing the appropriate inferential statistical test also depends on the scale level of the data to
be analyzed.
Interval Data
The z-test and the t-test, are appropriate tests for the population mean based on interval data. If
neither the population standard deviation nor whether the distribution is normal is known, a t-test
must be used.
z Test
The z test allows us to decide whether the sample mean ( X ) is likely to have come from a
population with the hypothesized mean value (H0: μ = μ0). It is appropriate when:
• σ is known
• σ is unknown but the sample size is large enough for the Central Limit Theorem to apply
(generally n>30)
If the population standard deviation () is known, then:

X −
z=
/ n
If σ is not known, we use the sample standard deviation, s, instead:
X − X −
z= =
sX s/ n
If the calculated z value is large enough to occur by sampling error alone less than α of the time,
the null hypothesis is rejected in favor of the alternate hypothesis with 1- α confidence.
For proportions, the appropriate test statistic is:

p −
z=
p(1 − p)/n
where p = sample proportion and π = hypothesized proportion. To use this test, np(1-p) ≥ 5 must
hold.
t-Test
If n < 30 and σ is unknown, the t test should be used, as long as and the population is itself
known to be normal. The t test also holds for any sample size. The population standard deviation,
σ, is estimated by the sample standard deviation, s. The number of degrees-of-freedom (df) must
be known in order to apply the t-test. For the testing of means, the df is always n-1.
(Critical values of the t statistic are presented in Table A.2 on page 659.)
Nominal Data
The chi-square test is a procedure for comparing a hypothesized or expected population
distribution across nominal categories (“E”) against an observed distribution (“O”):
k
(O − Ei ) 2
2 =  i
i =1 Ei
where k = number of categories of the variable, Oi = observed number of respondents and Ei =
hypothesized number of respondents.
Critical values of the chi-square distribution are used as cutoffs against which to compare the
computed value. For a univariate chi-square test, df = k – 1.
Bivariate Procedures pg 410
Bivariate analysis is needed to determine whether the values of one variable offer useful
information about the values of another. The scale level of the variables involved determines the
appropriate analytic method and associated statistics. (For interrelationships among variables of
different scale levels, see Chapter 9.)
Descriptive Statistics for Bivariate Analysis pg 411
The most important descriptive statistical procedure appropriate for use with two interval
variables is simple linear regression. The main measure of association is called the linear
correlation coefficient or “the correlation.”
Linear Correlation Coefficient
Linear correlation is a measure of the degree to which two interval variables are associated. The
linear correlation coefficient, rXY, is a scale-independent measure of the linear relationship
between X and Y:
n
Cov ( X , Y ) (X i − X )(Yi − Y )

rXY = = i =1
Cov( X , X )Cov(Y , Y )  n 2

n
2
 i ( X − X )    (Yi − Y ) 
 i =1   i =1 
n
(X i − X )(Yi − Y )
1 n
 ( X i − X )   (Yi − Y ) 
= i =1
(n − 1) sx s y
=  
(n − 1) i =1  sx
 
  s y 
If r = 0, there is no linear relationship between the variables, although there may be another,
more complex relationship.
The coefficient of determination (r2) is the exact percentage of variation shared by two variables.
Simple Regression
Regression helps us understand how one or more independent variables are related to a
dependent variable of interest, and to make predictions based on this understanding.
Partitioning the Sum of Squares

Regression compares the total deviations across all observations:
n n n
 (Yi − Y ) 2 =  (Yî − Y ) 2 +  (Yi − Yî ) 2
i =1 i =1 i =1
↓ ↓ ↓
SS Total = SS Explained + SS Unexplaine d
by Regression by Regression
↓ ↓ ↓
Total = Explained + Unexplained
Variation Variation Variation
(SST) (SSR) (SSE)
where
Yi = the ith observation on the dependent variable (Y)
Y = the mean of the {Yi}
Yî = the predicted value of the ith observation of the dependent variable
Figure 8.7 Partitioning Total Deviation in Simple Regression (page 417)
The variation explained by the regression is always positive and so reduces the total variation;
the question will be whether this degree of reduction is significant, meaning greater than what
would be expected by chance alone.
Fitting the Regression Line

To predict a value for the dependent variable ( Yî ) based on knowledge of the independent
variable (Xi), estimate a regression line Yˆ = a + bX that minimizes the sum of squared errors
i i
(“least squares”):
n
SS XY (X i − X )(Yi − Y )
b = = i =1
n
(X
SS XX
i − X )2
i =1
a = Y − bX
The Meaning and Significance of the Coefficients

The slope (b) always tells us how many units the dependent variable (Y) changes when we
change the independent variable (X) by one unit. The intercept (a) tells us the predicted value of
Y when X is zero; this only has meaning if an X value of zero is meaningful. Estimated quantities
are accompanied by measures of how much uncertainty the estimate entails—the number of
standard errors the estimates are from zero, listed as t-statistics in the output. Large values
indicate that estimates that far from zero are unlikely to have come about purely by chance.
Measures of Regression Fit and Quality

“r2-adjusted” alters r2 so that it will not increase when statistically unrelated or meaningless
variables are included. The standard error of the regression, SE(Reg) or se relates to the total
n
error left unexplained by the regression, SSe =  (Yi − Yî )2 , and quantifies the degree of ‘spread’
i =1
of the points around the regression line.
Using the Regression Model to Predict

To calculate a “mean” of some population quantity, divide by the degrees-of-freedom.
Regression output contains an Analysis of Variance or “ANOVA Table” that lists sums-of-
squares and associated degrees-of-freedom and mean-squares. Mean-squares measure variation;
in regression models, the important mean-square is associated with the Error, which measures
“how much variance in each data point can be attributed to error.”
The F-test is the ratio of the amount of variance attributable to the regression to the amount of
variance attributable to error; it refers to all the independent variables in your model taken
together whereas the t-test refers to each of the independent variables separately.
Inferential Statistics for Bivariate Analysis pg 424
Basic Inferential Test for Regression Coefficients

To test whether a regression coefficient has a specific value, calculate how many standard errors
the sample (b) value is from the hypothesized value (β):
b−
t =
sb
Inferential Test for Difference in Population Group Means

To test whether there is a difference in population group means, run a simple regression where
the independent (X) variable is the group indicator.
Differences in Proportions
To test differences in proportions, (e.g., “Did a greater proportion of people like the new ad,
compared with the old one?”), one should most properly use binary logistic regression, covered
in Chapter 9. However, for fairly large samples when the proportions in question aren’t close to 0
or 1, ordinary regression can sometimes be used on the binary data as an approximation to test
whether two population proportions are equal. Because this test is only approximate, it can in
some cases—such as when the proportions involved are very near 0 or 1—yield misleading
answers.
Nominal Association: The Chi-Square Test
When evaluating cross-tabulation tables, first check the chi-square for significance, then closely
examine those tables with significant chi-square values to get a sense of how the variables are
related. The chi-square test determines whether a relationship between two nominal variables
observed in a cross-tabulation is more than can be attributed to sampling error; if so, we may
claim they are related.
If A and B are independent, the expected number of items in the cell (joint category) AiBj is
n Ai n Bj R C (O − E ) 2
Eij = . The chi-square statistics is given by  = 
2 ij ij
, where Oij is the
n i =1 j =1 Eij
number actually observed in cell AiBj. When the calculated value of chi-square is more than the
cut-off value (based on its degrees-of-freedom) listed in the appropriate statistical table, we may
reject the null hypothesis of independence.
KEY TERMS pg 431
absolute frequency
The total number of times an event occurred during a study or experiment, disregarding the
number of opportunities it had to occur.
alternative hypothesis
A specific statement, opposed to the “null hypothesis,” subjected to statistical test. The
alternative hypothesis is typically what the researcher anticipates will actually happen, or the
hoped-for outcome (for example, in a clinical trial of a new drug or an ad test).
analysis of variance (ANOVA)

A statistical method that allows the simultaneous comparison of the means of many groups; in
this way, it generalizes the two-sample t-test. If the groups differ on a single variable (e.g., year
in college), the analysis is a one-way ANOVA; if on two variables (e.g., year in college and
gender), a two-way ANOVA; and so on. ANOVA is a special case of ordinary multiple linear
regression.
bimodal
Describes a variable with two distinct, relatively high rates of incidence (modes, or “humps”).
chi-square test
The main test done in categorical data analysis, often to determine whether two nominal
variables are independent of one another; can also be used to test whether a sample variance is
significantly different from a hypothesized value.
coefficient of determination
A measure of how much of the variation in one variable is explained by a simple linear
regression on a second variable; r2.
confidence level
1 -α, that is, the stated significance level (α) subtracted from 1.
covariance
A measure of dispersion for some quantity, equal to the square of the standard deviation.
decompose
In statistical models, the variance in the dependent variable is decomposed into that attributable
to the model (covariates) and that to random error. ANOVA models often feature an explicit
decomposition into various sources of variation.
discrete
A quantity or variable for which a list of distinct values can be made, each of which occurs with
some positive probability. Which day of the week an event will occur on is discrete; how many
times one flips a coin before getting “tails” is also discrete, even though it is limitless; the set of
all possible numbers between zero and one is not discrete, but rather continuous.
F-test
An exceptionally common test throughout statistics, and ubiquitous in various forms of
regression. Indicates whether the “entire model”—that is, all independent variables taken
together—explain a significant proportion of the variation in the dependent variable.
inferential statistics
A form of statistical analysis that relies on a sample to reach conclusions about a larger group
(e.g., the population from which the sample was drawn).
least squares
A general principle used throughout statistical modeling, suggesting that one minimize the sum
of squared deviations between observed values and those predicted by the model.
linear correlation coefficient

A measure of how much of the variance in one (interval-scaled) variable is reduced by knowing
the value of another (interval-scaled) variable; obtained by running a simple linear regression of
one variable on the other; r.
mean-squares
In statistical models, an estimate of an average squared distance, calculated by dividing a sum-
of-squares by the associated degrees-of-freedom.
median
A common measure of central tendency, the middle value of a variable in the sample. Exactly
half the values fall above and half below the median.
multiplication rule
A fundamental rule of probability stating that the probability of two independent events
occurring is the product of their individual probabilities.
null hypothesis
A specific statement, opposed to the “alternative hypothesis,” subjected to statistical test. The
null hypothesis typically states that there is no effect (i.e., the effect equals zero), or that two
quantities are equal to one another (i.e., their difference is zero), and this is what the researcher
wishes to cast doubt on through data collection and statistical testing.
one-tailed test
A statistical test of a hypothesis that specifies a direction (e.g., that some parameter’s value is
specifically greater than zero).
power of the test

Formally, the probability of rejecting the null hypothesis when it is false; usually denoted by (1 –
β) (where β is the probability of Type II error, not a regression coefficient, which sometimes uses
the same symbol).
p-value
A crucial quantity in any statistical test, distilling how strong the evidence is in favor of some
quantity not being zero (in the population, based on one’s sample data). All regression programs
report p-values for each coefficient and for the entire model.
regression line
A straight line that represents the idealized relationship between two variables, given that they
are interval scaled and related linearly to one another.
relative frequency
The number of times an event occurred during a study or experiment, divided by the total
number of opportunities for occurrence (or, by the total number of events that did occur).
robustness
Any statistic whose value resists dramatic change in the presence of outliers.
scale independent
When a statistic’s value does not change when the scale of measurement does (e.g., from pounds
to kilograms).
significance level
Formally, the greatest probability a researcher sets of rejecting the null hypothesis, even though
it happens to be true; usually denoted by a and frequently set in applied work at 0.05 or 0.01;
denoted by the greek letter α, it can also sometimes refer to the p-value of a statistical test.
simple linear regression

A model that posits a linear relationship between a dependent variable, Y, and just a single
predictor variable, X.
t-test
A statistical test performed on sample means (or differences of them), when the population
variance is unknown; commonly applied in the form of a “two sample” t-test, in ANOVA, and
for regression coefficients.
two-tailed test
A statistical test of a hypothesis that does not specify a direction (e.g., that some parameter’s
value is not equal to zero, where values both greater than zero or less than zero can invalidate the
hypothesis).
type I error (α)

The probability that the null hypothesis is (incorrectly) rejected when it is true; also known as
alpha (α) error; analogous to “convicting the innocent.”
type II error (β)

The probability that one (incorrectly) fails to reject the null hypothesis when it is not true; also
known as beta (β) error; analogous to “failing to convict the guilty.”
z-test
If a variable has a standard normal (z) distribution, one can test the null hypothesis of whether it
is equal to zero by checking values on a standard normal table or by computer.
DISCUSSION QUESTIONS
1 Distinguish between descriptive and inferential statistics.
Descriptive statistics summarize sample data without making any inferences about the
population. Inferential statistics use statistical analyses to make inferences about the population
based on data from a sample, with associated statements of uncertainty based on sampling error.
2 A sales manager had promised the entire sales force a special trip if average daily sales per
salesperson were $8,000 or more. A sample of ten salespersons yielded the following results:
average daily sales per salesperson = $7,800; standard deviation = $400. Can the sales
manager conclude that the entire sales force (i.e., the population) had reached the goal?
Explain your reasoning.
This problem can be viewed two ways. In the first, the salespeople must show their average daily
sales were greater than $8,000. If this is the case, no statistical tests are needed: the sample mean
of $7,800 can never demonstrate that the true mean for all salespeople was actually greater than
$8,000.
A second interpretation is that the manager can only deny the salespeople their trip if the true
mean is shown to be less than $8,000. Under this interpretation, we simply form the sample test
statistic:
(8,000 − 7,800 ) /( 400 / 10 = 1.581
This would be a one-tailed test. Assuming that sales in the population were normally distributed
would allow us to use a t-test, even with such a small sample, and in this case we would have
n-1 = 9 degrees of freedom. According to Table A.2 (page 659), the 0.10 and 0.05 cutoffs for t(9)
are 1.3830 and 1.8331, respectively. Thus, we would conclude that we had weak evidence to
deny the salespeople their trip. The exact p-value is 0.0742.
3 A political research firm undertook a sample of registered voters in a small community to see
whether a particular candidate would win the election. The sample size was 500, and the
result was that 51 percent of the sample favored this candidate. How would you set up the
null and alternative hypotheses in a manner relevant to settling the question at hand? If
people vote as they say they will, do the results indicate that this candidate will win the
election? What assumptions must you make in order to say anything meaningful on the basis
of these data?
The null hypothesis, as always, is one of equality. In this case, it would say that the true
probability, π, of the candidate getting a randomly-selected vote was exactly ½. So, our null
hypothesis is π = ½, and our (one-tailed) alternative hypothesis is π > ½:
H0: π = ½
H1: π > ½
n = 500
p = 0.51
Let us conduct the relevant statistical test based on this sample. The test statistic is
z = ( p −  ) / ( p)( p - q)/n
z = (0.51 − ½) / [(½)(1 - ½)/500] = 0.447
The one-sided p-value turns out to be 0.133. Thus, this is unacceptably weak evidence in favor of
the candidate’s winning, and we could not conclude the probability of doing so was significantly
different from ½.
4 MINICASE: The Lake City chief of police wanted to discover how fast the average car
traveled on a particular stretch of highway. To obtain this information, he placed a hidden
radar device beside the highway and clocked speeds, in miles per hour, for one hour. The
following data were recorded by the device:
73 49 70 63 83 61
55 61 60 68 62 64
52 56 69 60 55 71
65 66 59 62 59 58
Calculate the appropriate statistics for central tendency and dispersion. What are the
problems with this design?
The mean speed = 1501/24 = 62.54 mph.
n
2
 ( Xi − X )
The sample variance is s 2 = i =1 = 54.87, and the sample standard deviation is therefore
n −1
the square root of that, 7.41.
The problem with this design is that it includes a disproportionately large number of fast cars in
the sample, because more of the faster cars pass by the radar detector in a one hour stretch. (A
“fast” car increases its probability of selection if it is traveling faster than the others, simply
because it covers more highway per hour than a slower car.) To correctly determine the speed of
an “average” car, the officer would have to “freeze” a stretch of highway and then record the
speeds of all cars in this sample to determine the speed of an “average” car. Also, with only one
device doing the testing, the reliability and validity are compromised. A better method would be
to have several radar devices along the stretch of road.
5 What questions must researchers answer in order to select the appropriate bivariate
statistical procedure for a given data set? Are there some data types for which no bivariate
analysis is possible?
To select the appropriate statistical procedure, one needs to know the data type of the relevant
variables and whether one intends to make inferences about the population. Bivariate analyses
can be performed on all data types, although not all combinations of data types are covered in
this chapter.
6 When can simple linear regression be used? Specifically, what sort of data must one have,
and what assumptions are made about the relationship between the variables in question?
When should it not be used?
Simple linear regression can only be used on one intervally-scaled dependent variable and one
intervally-scaled independent variable to assess how the independent variable is related to the
dependent variable. It assumes a linear relationship between the two variables. It should not be
used on ordinal or nominal data, and it is not appropriate if the relationship between the two
variables is more complex than a simple linear correspondence.
7 A marketing manager was given the following table of frequency counts to assess the nature
of the relationship between age and attendance at NFL games. What conclusion can be
drawn?
Age
Attend NFL Games Younger Than 40 Older Than 40 Total
Yes 466 231 697
No 224 323 547
Total 690 554 1,244
The expected values are calculated with this formula:

n Ai n Bj
Eij =
n
The chi-square test calculations, if done by hand, would look like this:
Cell Oij Eij Oij – Eij (Oij – Eij)2 (Oij – Eij)2/Eij

Number
1,1 466 386.6 79.4 6,304.36 16.31
1,2 231 310.4 -79.4 6,304.36 20.31
2,1 224 303.4 -79.4 6,304.36 20.78
2,2 323 243.6 79.4 6,304.36 25.88
Total 1,244 1,244.0 83.28
= χ2
R C (Oij − Eij ) 2
The chi-square value, using  =  2
, is 83.28; with 1 df, the p-value is
i =1 j =1 Eij
-20
7.13 × 10 , meaning there is exceptionally strong evidence that age and NFL attendance are
related. For (R - l)(C - 1) = 1 df, the critical value of χ2 at α = 0.05 is 3.84. On this basis, the null
hypothesis of no relationship between age and attendance at NFL games is rejected.
8 The same marketing manager also had a table of frequency counts between age and
attendance at college football games. Given these new data, what conclusion can be drawn?
Age
Attend College Younger Than 40 Older Than Total
Football Games 40
Yes 242 271 513
No 251 265 516
Total 493 536 1,029
The chi square test calculations:

Cell Oij Eij Oij – Eij (Oij – Eij)2 (Oij – Eij)2/Eij
Number
1,1 242 245.78 -3.78 14.29 0.058
1,2 271 267.22 3.78 14.29 0.053
2,1 251 247.22 3.78 14.29 0.058
2,2 265 268.78 -3.78 14.29 0.053
Total 1,029 1,029.00 0.22
= χ2
, is 0.22; with 1 df, the cutoff value is again
Eij
i =1 j =1
3.84, but the p-value is 0.637, which is definitely not significant.
9 In a study of advertising effects, two waves of consumers were interviewed. Wave 1 took
place before a new campaign was introduced, and Wave 2 a few months after the new
campaign had started. As part of the analysis of the data, a comparison was made between
the demographic characteristics of the consumers in Wave 1 and Wave 2, under the
hypothesis that they had not changed. Typical of the reported results is the following:
Gender
Male Female
Wave 1 52% 47%
Wave 2 48% 53%
The researchers used knowledge of the sample size to calculate a chi-square of 4.16. What
conclusion can be drawn from this result?
The calculated chi-square value of 4.16 entails one degree of freedom; therefore, the associated
p-value is 0.0414. The critical chi-square for significance at 1 df and α = 0.05 is 3.84. This
evidence is therefore strong enough to conclude that the consumers’ demographics had changed
between the two waves; the waves and gender distributions are not independent of each other.
We then look back into the table to interpret this relationship. However, we cannot do this given
the way information is presented in the table. The percentages in the table are in the wrong
direction. What we are interested in is a comparison of the male/female percentages in wave 1
versus wave 2. This would be directly observable if we had row percentages. What we have
instead, for example, is the percentage of all males in both waves that were in wave 1 (52%).
This is not a useful piece of information for comparing the gender composition of the waves. It
would be even more of a problem as the difference in size of the waves gets larger.
An anecdote from the field: a large set of tables like this one got a marketing research supplier
fired from a big project. Casting percentages the proper way is considered an important part of
marketing research competence.
10 Mark Schwinn and Leopoldine Grier were resident advisors (RAs) assigned to the first coed
hall at Bindlesmock College. A welcome party was scheduled to take place the weekend after
classes started, and the RAs had the responsibility of ordering the soft drinks. They could not,
however, agree on kinds of soft drinks they should order (diet cola versus regular cola versus
other assorted flavors). Thirty to 40 cases were to be ordered, and any unopened cans could
be returned. Mark wanted to order 16 cases of regular cola, 6 diet cola, 6 regular non-cola,
and 2 diet non-cola; Leopoldine wanted to order, respectively, 11, 6, 9, and 4 cases. To avoid
running out of anything, they decided to order the larger estimate of each kind, making a 35-
case order (16 cases regular cola, 6 diet cola, 9 regular non-cola, and 4 diet non-cola). The
actual consumption at the party was 12 cases of regular cola, 4 diet cola, 8 regular non-cola,
and 1 diet non-cola. Is there a difference in preferences between regular versus diet soft
drinks? Cola versus non-cola? Is there reason to believe the two variables are in fact
independent?
All that matters is what was actually consumed. Note that, if ALL of a particular type was
consumed, we could make no inferences, as preference for that type could have been stronger
than the stock at the party accommodated. But that was not the case here, and we can make the
following table:
Regular Diet
Cola 12 4
Non-Cola 8 1
We must be careful when applying the chi-square tests, as some of the cell counts are quite low.
Let us first test whether “Cola” and “Regularness” are independent. The expected cell counts are:
Regular Diet
Cola 12.8 3.2

Non-Cola 7.2 1.8
, is 0.694; with 1 df, the p-value is 0.40,
Eij
i =1 j =1
meaning we have no evidence at all that the variables are related. We can carry out similar tests
for just “Regular vs. Diet” (20 vs. 5) and for “Cola vs. Non-Cola” (16 vs. 9). Both tests have 1 df.
The test statistics and p-values are 9.0 and 0.0027 for “Regular vs. Diet” and 1.96 and 0.16 for
“Cola vs. Non-Cola.” We therefore conclude there is evidence of a difference only for the
former.
11 In a recent study on American travel habits, the following data were obtained.
Case Gender Children at Respondent Vacations

Home? Age per Year
1 M Y 25 1
2 M N 52 16
3 F N 34 8
4 F Y 33 1
5 F Y 51 5
6 F Y 29 0
7 M Y 35 2
8 F N 27 8
9 M Y 46 4
10 M N 30 10
11 F N 45 14
12 M Y 38 3
a Is there a relationship between the presence of children at home and the gender of the
respondent?
We can run a binary logistic regression to predict “Children at Home” based on “Gender” (or, in
fact, vice-versa, as they are both binary; but the way we will do it makes more sense in terms of
interpretation). The regression output looks like the following:
Wald p-Value Two Odds-Ratio

Model b Std. Error Statistic df Tailed Exp(b)
Constant 0.0000 0.8165 0.0000 1 1.0000 1.0000
Gender 0.6931 1.1902 0.5823 1 0.5603 1.9999
Predicted
Actual No Yes % Correct
No 3 2 60.0%
Yes 3 4 57.1%
Average % Correct 58.3%
As we can see from the output, the p-value on “Gender” is 0.5603, which is unacceptably large.
We therefore conclude there is no evidence of a relationship between Gender and Children at
Home.
Note that it is also possible to approach this question using a contingency table and chi-square
analysis. However, that approach would involve some cells with very few observations and
expected values, calling it into question.
b What is the relationship between age and the presence of children at home? How might
you assess this via regression?
We can run another logistic regression analysis, this time using Age to predict Children in the
Home (we could also run an ordinary regression for the reverse relationship if we wished to).
The results of this regression are as follows:
Wald p-Value Two- Odds-Ratio

Model b Std. Error Statistic df Tailed Exp(b)
Constant 0.7501 2.5125 0.2986 1 0.7653 2.1172
Age -0.0111 0.0656 -0.1696 1 0.8653 0.9889
Predicted
Actual 0 1 % Correct
0 0 5 0.0%
1 0 7 100.0%
Average % Correct 58.3%
Based on these results, we see the p-value on “Age” is 0.8653, which is unacceptably large. We
therefore conclude there is no evidence of a relationship between Age and Children at Home.
Again, note that a Chi-square analysis would not be useful here, due to very small expected cell
sizes.
c What is the relationship between age and number of vacations taken per year? Calculate
r2. (Hint: Plot the data points before doing a regression analysis. Look for the effects of
another, or “extraneous,” variable; and calculate your least-squares regression line or
lines accordingly. Is one equation appropriate? Would two explain the data better? If so,
how? If not, why not?)
The regression analysis output is:
r r2 Aj. r2 SE Observ.
0.485 0.235 0.159 4.822 12
p-
Source of Sum-of- Mean F Test Value
Variation Squares df Squares Statistic One -
Tailed
Regression 71.487 1 71.487 3.075 0.1101

Residual 232.513 10 23.251
Total 304 11
p-  
Stand. Value Lower Upper
Std. Coeff. Test Two- 95% 95%
Model b Error Beta Statistic Tailed CL CL
Constant -4.157 5.958 -0.698 0.5012 -17.43 9.12
Age 0.274 0.156 0.485 1.753 0.1101 -0.07 0.62
The p-value on Age is 0.1101, which is too large to conclude that Age is related to Number of
Vacations. However, a quick look at the data shows that the households taking many vacations
are those without children! We should include number of children in the regression as well
(Chapter 9 provides details on how to interpret such a multiple regression). Doing so yields the
following results:
r r2 Aj. r2 SE Observ.
0.979 0.958 0.949 1.186 12
p-Value
Source of Sum-of- Mean F Test One
Variation Squares df Squares Statistic Tailed
Regression 291.343 2 145.671 103.582 0.0000
Residual 12.657 9 1.406
Total 304 11
 
Stand. p-Value Lower Upper
Std. Coeff. Test Two 95% 95%
Model b Error Beta Statistic Tailed CL CL
Constant 1.788 1.540 1.160 0.2757 -1.697 5.272
Age 0.250 0.038 0.443 6.508 0.0001 0.163 0.337
Children at
Home -8.693 0.695 -0.851 -12.503 0.0000 -10.265 -7.120
Note that the p-value on Age is now 0.0001! And the adjusted r2 is now 0.949, from a prior value
of 0.159. These are enormous differences, especially in the significance of Age, all due to the
inclusion of the important variable “Number of Children.”
Obviously, two “clumps” of data points exist here. They are distinguished by the variable
“children present at home.” r-square for the whole data set is only 0.235. However, when we add
in “number of children”, it skyrockets to 0.958, which is enormous by comparison. This
illustrates the importance of “rough-sketching” the data before performing analyses.
d Is the relationship obtained to answer (c.) significant (i.e., was a significant proportion of
the variance in the dependent variable explained by the regression)?
When Age was the only independent variable, it was not significant, but it became strongly
significant when an additional independent variable (“Number of Children” was included).
e Using the equation or equations you computed in the previous problem, complete the
following data set by using your regression model(s) to predict “vacations per year” for
the new cases, 13–16:
Case Gender Children at Respondent Age Vacations

Home? per Year
13 M Y 40
14 F N 40
15 M Y 65
16 F N 18
Using the two different regressions above, we get the following predictions; the first column is
for the model excluding “Number of Children”, using “Age” as the only predictor:
Children at Vacations Vacations

Case Gender home? Respondent Age per Year per Year
13 M Y 40 6.8 3.1
14 F N 40 6.8 11.8
15 M Y 65 13.6 9.4
16 F N 18 0.8 6.3
Note that their predictions are very different! The rightmost column’s predictions are presumably
more accurate, given the relative fit of the two models.
Two of these answers should be approached with extreme caution: the ages 18 and 65 are outside
the range of data from which the regression was calculated. It is usually considered invalid to
extrapolate any regression to data points beyond those input, in this case, the range of 25-52
years. It would be reasonable to simply say that we cannot use the regression for these values,
even though we can easily use the regression equation to calculate them.
REVIEW QUESTIONS
1 When describing a variable, why might a researcher want to use measures of both central
tendency and dispersion?
a Two samples with the same central tendency can have very different levels of dispersion,
which might lead to different conclusions about them.
b Two samples with the same level of dispersion might have very different central
tendencies, which might lead to different conclusions about them.
c They are not both necessary—measures of central tendency suffice to describe a variable.
d both a and b
Ans: d
Rationale: Although measures of central tendency tend to give us some idea about a variable,
understanding the dispersion of that variable—how well the central tendency describes the
location of the “bulk” of individual observations—is essential to our understanding, both
informally and in a formal, statistical sense, of how accurate our conclusions about that variable
might be.
2 Distinguish between significance level and confidence level.

a The significance level gives the probability (α) of committing a Type I error, whereas the
confidence level gives the probability (β) of committing a Type II error.
b The significance level gives the probability (α) of committing a Type I error, whereas the
confidence level gives 1 – the probability (α) of committing a Type I error.
c The significance level gives the probability (β) of committing a Type II error, whereas
the confidence level gives 1 – the probability (β) of committing a Type II error.
d One is simply the negative of the other.
Ans: b
Rationale: A Type I error is committed when, on the basis of a sample, the null hypothesis is
rejected when it is, in fact, correct. A Type II error is committed when, on the basis of a sample,
we fail to reject the null hypothesis when, in fact, it is incorrect. The significance level is defined
as the probability of committing a Type I error, which is denoted by α. The confidence level is
defined as the probability of not committing a Type I error, and thus equals 1– α. The probability
of committing a Type II error is denoted by β; it is typically unrelated to either the confidence
(1 – α) or significance (α) levels.
3 What is the power of the test?

a the likelihood of correctly accepting a true null hypothesis
b the likelihood of correctly rejecting a false null hypothesis
c 1–α
d β
Ans: b
Rationale: The power of the test is defined as the probability of correctly rejecting a false null
hypothesis. Because the probability of failing to reject the null hypothesis when it is, in fact,
false is given by β, the power of the test is equal numerically to 1 – β. Note that power can be
calculated only when a specific alternative hypothesis is stated.
4 From a sample of Nintendo users, the following frequency count was generated for the
categories of the variable age in a sample of size n = 745.
Age Frequency
18–24 315
25–39 205
40–54 115
55 and older 110
Does the sample data suggest that Nintendo usage spread evenly across these four population age
groups?
a No
b Yes
c We cannot determine if usage is spread evenly across these four population age groups.
Ans: a
Rationale: This question offers us a chance to use the chi-square test. Recall that the chi-square
test is used to check a hypothesized distribution across categories (in this case, an equal or
“uniform” distribution of Nintendo usage across the age categories indicated) against an
observed distribution (the results of the sample). We wish to test whether or not Nintendo usage
is spread evenly across the four age categories. If it were, we would expect equal numbers of
persons in each category. Given that the sample was of size 745, we would expect 186.25
persons in each age group from the sample if there was an even distribution of usage over age.
The first step in calculating the chi-square test statistic is to find the difference between the
observed number of persons in each category and the number we expect there; we then square
this difference, divide by the number expected, and sum over all categories. This summed value
is the chi-square test statistic, and we compare this value to the chi-square distribution. If the
calculated value is sufficiently large, then we say that there is sufficient evidence to reject the
hypothesized distribution. The calculations are straightforward:
4
(Oi − Ei ) 2 (315 − 186.25) 2 (110 − 186.25) 2
 =
2
= + ... + = 149.36
i =1 Ei 186.25 186.25
We must compare the calculated value of 149.36 to a critical value from a chi-square table or
have a statistical program calculate the exact p-value. Even at α = 0.001, a chi-square value (with
df = 3) is 16.27; the computer tells us the exact value p = 3.6 × 10–32. By either comparison, we
have very strong reason to reject the null hypothesis that usage is evenly distributed over age, in
favor of the alternative hypothesis that it is not.
5 The manager of a movie theater hypothesized that twice as many of the theater’s patrons were
younger than 30 as were 30 and older. A sample of n = 690 patrons showed 450 were younger
than 30, and 240 were 30 and older. What can you conclude about the theater manager’s
hypothesis regarding the population of patrons?
a The theater manager’s hypothesis is correct; we cannot reject the hypothesis of the
distribution specified at even the 5 percent level.
b The theater manager’s hypothesis is incorrect; we can reject the hypothesis of the
distribution specified at the 5 percent level.
c We cannot tell whether the theater manager’s hypothesis is correct or incorrect.
d none of the above
Ans: a
Rationale: Again, we can use the chi-square test to evaluate the null hypothesis that the
distribution of patrons has twice as many persons in the “younger than 30” category as in the “30
and older” category. (See the previous question and its solution for more detail on calculations.)
When we compute the chi-square statistic in this condition (with the expected number of patrons
in the “younger than 30” and “30 and older” categories as 460 and 230, respectively), we
generate a test statistic of 0.652; this is not significant at even the 5 percent level, for which the
critical chi-square value with df = 1 is 3.84. In fact, the computer tells us that the p-value
associated with the calculated 0.652 test statistic is p = 0.42, which is far short of significance.
Note that, as always in hypothesis testing, we set up the null hypothesis as an equality statement
(“the size of the younger group EQUALS twice that of the older one”), and here we find that we
cannot reject the manager’s claim.
6 A company had adopted the following decision rule with respect to introducing a new product:
“If average monthly consumption is 300 ounces or more, we will enter into test market.” An in-
home placement test of n = 80 yielded the following results: average monthly consumption = 290
ounces; standard deviation = 45. Given their stated decision rule, what decision should they
make?
a The company should enter the test market.
b The company should not enter the test market.
c If entering the test market is relatively cheap, the company should do it; otherwise, the
company should not enter the test market.
d We cannot determine, on the basis of this sample, whether or not the company should
enter the test market or not.
Ans: b
Rationale: Based on this sample, it is more likely that average monthly consumption is less than
300 ounces than that it is 300 ounces or more; thus, the company should not enter the test
market, according to the decision rule stated. Note that no statistical tests are needed here,
because the test statistic is in a wrong direction; monthly consumption in the sample would need
to be more than the stated (population) cutoff of 300 for the company to even consider entering
the market.
7 How do you test to see whether a simple regression has explained a significant portion of the
variation in the dependent variable?
a A significant portion of the variation is explained if the coefficient on the intercept is
significant.
b A significant portion of the variation is explained if the coefficient on the dependent
variable is significant.
c A significant portion of the variation is explained if the computed value of the F-test is at
least 2.
d A significant portion of the variation is explained if the r-squared value is greater than
some threshold, given by the specific application.
Ans: d
Rationale: The r-squared value gives us the proportion of variation in the dependent variable that
is explained by variation in the independent variable. Obviously, as r-squared increases in size, a
larger proportion of variation in the dependent variable is being explained by independent
variable variation; however, what constitutes a “significant proportion” of the variation being
explained differs based on the context. None of the other answers has any relation to
significance, except (c), for the F-test; however, the appropriate cutoff needs to be looked-up,
and is likely to be quite a bit greater than 2.
8 For which of the following uses is the chi-square test valid?

a The univariate chi-square test is used to test a hypothesized distribution over nominal
categories against the results of a sample.
b The bivariate chi-square test is used to test the hypothesized independence of two
nominal variables.
c a and b
d none of the above
Ans: c
Rationale: Answers a and b give the definitions of the chi-square test for univariate and bivariate
procedures, respectively. Note that there are other uses of the chi-square test in more complex
statistical settings as well.
9 A study of collegiate basketball attendance related the number of home games attended in a
year and the number of years the attendee has lived in the city:
Case Number of Home Years Lived in City

Games Attended
1 8 28
2 2 6
3 1 3
4 3 12
5 8 20
6 4 23
What is the relationship, if any, between the two variables?

a There does not appear to be any relationship between the two variables; even though
higher number of years living in the city appears to correlate with more home games
attended in a year, this relationship is not significant at the 5 percent level.
b There appears to be a positive linear relationship; when a simple regression is estimated
on the data, it demonstrates that living in the city for more years causes a person to attend
more home games, and this relationship is significant at the 5 percent level.
c We cannot determine anything about the relationship between the two variables on the
basis of this sample.
d There appears to be a positive linear relationship; as the number of years lived in the city
increases, the number of home-games attended increases, and this relationship is
significant at the 5 percent level.
Ans: d
Rationale: Let us consider the regression output of Y = games on X = Years:
Coefficients b Std. Error Std. Beta t-Test p-Value Two-Tailed

Statistic
Intercept 0.328 1.380 0.238 0.824
X = Height 0.261 0.078 0.860 3.369 0.028
r r2 Adj. r2 SE(Reg) n
0.860 0.739 0.674 1.719 6
Source of Sum-of- df Mean F-Test p-Value One-Tailed
Variation Squares Squares Statistic
Regression 33.52 1 33.52 11.35 0.028
Error 11.81 4 2.95
Total 45.33 5
This regression indicates that there is a significant positive linear relationship between the two
variables, with a coefficient significant at the 5 percent level; more precisely, p = 0.028.
However, we must always stop short of making any statements about “causality,” for which
regression is not suited; this is why answer b is not correct. For example, it might be the case that
people who live in the city longer come to like the sports team more, and thus they attend more
home games; alternatively, it might be the case that people who like the sports team in question
more end up living in the city longer, to pursue their interest in the team.
10 What does the F-statistic in a regression with a single independent variable tell us?
a The F-statistic conveys the likelihood that the coefficient on the intercept is zero.
b The F-statistic conveys the likelihood that the coefficient on the single independent
variable is zero.
c The F-statistic tells us no more than the t-statistic for that variable’s coefficient.
d both b and c
Ans: d
Rationale: The F test is used to test the hypothesis that all of the independent variables’
regression coefficients are equal to zero. When a regression has only one independent variable,
this is equivalent to testing the hypothesis that the variable’s coefficient is equal to zero (in fact,
the F-statistic is equal to the square of the t-statistic for that variable’s coefficient). Note that this
is true only in the special case of a simple linear regression, with but a single predictor variable;
for a multiple regression, this relationship does not hold.
TOPICS FOR FURTHER DISCUSSION
A. A number of local citizens groups are fighting the introduction of a Walmart superstore in
their area. Among their claims are that the introduction of Walmart puts local businesses out of
business and that Walmart employees put a greater burden on taxpayers through public
assistance programs for food, rent and health care than employees of companies that pre-existed
the introduction of Walmart. Discuss with the students
• what data must be provided in order to support or reject these hypotheses
• the scale level(s) of the relevant variables
• the kind of statistical analyses should be done on this data to support or reject these
hypotheses
B. Discuss with the students the balance of Type I and Type II errors that might be desired
in hypothesis-testing in various types of situations, such as:
• Agent X, commonly occurring in most grain products, is associated with allergic
reactions in 75% of the population, including labored breathing, stomach upset and skin
rashes.
• Agent Y, which is only created as an industrial waste by certain factories in a specific
industry, is associated with quick death 57% of the time and symptoms requiring
hospitalization 95% of the time.
• The release of an IPOD-like device by a stereo manufacturer will increase its sales by
65%.
• Lengthening hours of operation of the “diamond lane,” during which commuters may not
use the third freeway lane unless they are in a vehicle containing more than one person,
by one half hour in the morning will increase commute time between 8:30 and 9:00 a.m.
by one hour.
Although there are no absolute correct answers in these cases, ask students to explain their
reasoning and the various factors and possible consequences they considered.
FURTHER READING
Aczel, Amir D., and Jayavel Sounderpandian. Complete Business Statistics, 7th ed. New York:
McGraw-Hill/Irwin, 2008.
Hyperstat Online: http://davidmlane.com/hyperstat.
Introductory Statistics: Concepts, Models, and Applications,

http://www.psychstat.missouristate.edu.
Lane, David. Hyperstat, 2nd ed. Cincinnati, OH: Atomic Dog/Cengage Learning, 2002.
The Little Handbook of Statistical Practice, http://www.tufts.edu/~gdallal/LHSP.HTM.
Multivariate Statistics: Concepts, Models, and Applications,

http://www.psychstat.missouristate.edu.
The Research Methods Knowledge Base, at http://www.socialresearchmethods.net.
Rice University Virtual Stats Lab, http://onlinestatbook.com/rvls.html.
Statsoft Electronic Statistics Textbook, http://www.statsoft.com/textbook/stathome.html.
Stockburger, David. Introductory Statistics: Concepts, Models and Applications. Cincinnati, OH:
Atomic Dog/Cengage Learning, 2001.
Trochim, William, and James P. Donnelly. The Research Methods Knowledge Base, 3rd ed.
Cincinnati, OH: Atomic Dog/Cengage Learning, 2007.
Another random document with
no related content on Scribd:
America was opened after the feudal mischief was spent, and so
the people made a good start. We began well. No inquisition here,
no kings, no nobles, no dominant church. Here heresy has lost its
terrors. We have eight or ten religions in every large town, and the
most that comes of it is a degree or two on the thermometer of
fashion; a pew in a particular church gives an easier entrance to the
subscription ball.
We began with freedom, and are defended from shocks now for a
century by the facility with which through popular assemblies every
necessary measure of reform can instantly be carried. A congress is
a standing insurrection, and escapes the violence of accumulated
grievance. As the globe keeps its identity by perpetual change, so
our civil system, by perpetual appeal to the people and acceptance
of its reforms.
The government is acquainted with the opinions of all classes,
knows the leading men in the middle class, knows the leaders of the
humblest class. The President comes near enough to these; if he
does not, the caucus does, the primary ward and town-meeting, and
what is important does reach him.
The men, the women, all over this land shrill their exclamations of
impatience and indignation at what is short-coming or is unbecoming
in the government,—at the want of humanity, of morality,—ever on
broad grounds of general justice, and not on the class-feeling which
narrows the perception of English, French, German people at home.
In this fact, that we are a nation of individuals, that we have a
highly intellectual organization, that we can see and feel moral
distinctions, and that on such an organization sooner or later the
moral laws must tell, to such ears must speak,—in this is our hope.
For if the prosperity of this country has been merely the obedience of
man to the guiding of Nature,—of great rivers and prairies,—yet is
there fate above fate, if we choose to spread this language; or if
there is fate in corn and cotton, so is there fate in thought,—this,
namely, that the largest thought and the widest love are born to
victory, and must prevail.
The revolution is the work of no man, but the eternal
effervescence of Nature. It never did not work. And we say that
revolutions beat all the insurgents, be they never so determined and
politic; that the great interests of mankind, being at every moment
through ages in favor of justice and the largest liberty, will always,
from time to time, gain on the adversary and at last win the day.
Never country had such a fortune, as men call fortune, as this, in its
geography, its history, and in its majestic possibilities.
We have much to learn, much to correct,—a great deal of lying
vanity. The spread eagle must fold his foolish wings and be less of a
peacock; must keep his wings to carry the thunderbolt when he is
commanded. We must realize our rhetoric and our rituals. Our
national flag is not affecting, as it should be, because it does not
represent the population of the United States, but some Baltimore or
Chicago or Cincinnati or Philadelphia caucus; not union or justice,
but selfishness and cunning. If we never put on the liberty-cap until
we were freemen by love and self-denial, the liberty-cap would mean
something. I wish to see America not like the old powers of the earth,
grasping, exclusive and narrow, but a benefactor such as no country
ever was, hospitable to all nations, legislating for all nationalities.
Nations were made to help each other as much as families were;
and all advancement is by ideas, and not by brute force or mechanic
force.
In this country, with our practical understanding, there is, at
present, a great sensualism, a headlong devotion to trade and to the
conquest of the continent,—to each man as large a share of the
same as he can carve for himself,—an extravagant confidence in our
talent and activity, which becomes, whilst successful, a scornful
materialism,—but with the fault, of course, that it has no depth, no
reserved force whereon to fall back when a reverse comes.
That repose which is the ornament and ripeness of man is not
American. That repose which indicates a faith in the laws of the
universe,—a faith that they will fulfil themselves, and are not to be
impeded, transgressed or accelerated. Our people are too slight and
vain. They are easily elated and easily depressed. See how fast they
extend the fleeting fabric of their trade,—not at all considering the
remote reaction and bankruptcy, but with the same abandonment to
the moment and the facts of the hour as the Esquimau who sells his
bed in the morning. Our people act on the moment, and from
external impulse. They all lean on some other, and this
superstitiously, and not from insight of his merit. They follow a fact;
they follow success, and not skill. Therefore, as soon as the success
stops and the admirable man blunders, they quit him; already they
remember that they long ago suspected his judgment, and they
transfer the repute of judgment to the next prosperous person who
has not yet blundered. Of course this levity makes them as easily
despond. It seems as if history gave no account of any society in
which despondency came so readily to heart as we see it and feel it
in ours. Young men at thirty and even earlier lose all spring and
vivacity, and if they fail in their first enterprise throw up the game.
The source of mischief is the extreme difficulty with which men are
roused from the torpor of every day. Blessed is all that agitates the
mass, breaks up this torpor, and begins motion. Corpora non agunt
nisi soluta; the chemical rule is true in mind. Contrast, change,
interruption, are necessary to new activity and new combinations.
If a temperate wise man should look over our American society, I
think the first danger that would excite his alarm would be the
European influences on this country. We buy much of Europe that
does not make us better men; and mainly the expensiveness which
is ruining that country. We import trifles, dancers, singers, laces,
books of patterns, modes, gloves and cologne, manuals of Gothic
architecture, steam-made ornaments. America is provincial. It is an
immense Halifax. See the secondariness and aping of foreign and
English life, that runs through this country, in building, in dress, in
eating, in books. Every village, every city, has its architecture, its
costume, its hotel, its private house, its church, from England.
Our politics threaten her. Her manners threaten us. Life is grown
and growing so costly that it threatens to kill us. A man is coming,
here as there, to value himself on what he can buy. Worst of all, his
expense is not his own, but a far-off copy of Osborne House or the
Elysée. The tendency of this is to make all men alike; to extinguish
individualism and choke up all the channels of inspiration from God
in man. We lose our invention and descend into imitation. A man no
longer conducts his own life. It is manufactured for him. The tailor
makes your dress; the baker your bread; the upholsterer, from an
imported book of patterns, your furniture; the Bishop of London your
faith.
In the planters of this country, in the seventeenth century, the
conditions of the country, combined with the impatience of arbitrary
power which they brought from England, forced them to a wonderful
personal independence and to a certain heroic planting and trading.
Later this strength appeared in the solitudes of the West, where a
man is made a hero by the varied emergencies of his lonely farm,
and neighborhoods must combine against the Indians, or the horse-
thieves, or the river rowdies, by organizing themselves into
committees of vigilance. Thus the land and sea educate the people,
and bring out presence of mind, self-reliance, and hundred-handed
activity. These are the people for an emergency. They are not to be
surprised, and can find a way out of any peril. This rough and ready
force becomes them, and makes them fit citizens and civilizers. But if
we found them clinging to English traditions, which are graceful
enough at home, as the English Church, and entailed estates, and
distrust of popular election, we should feel this reactionary, and
absurdly out of place.
Let the passion for America cast out the passion for Europe. Here
let there be what the earth waits for,—exalted manhood. What this
country longs for is personalities, grand persons, to counteract its
materialities. For it is the rule of the universe that corn shall serve
man, and not man corn.
They who find America insipid—they for whom London and Paris
have spoiled their own homes—can be spared to return to those
cities. I not only see a career at home for more genius than we have,
but for more than there is in the world.
The class of which I speak make themselves merry without duties.
They sit in decorated club-houses in the cities, and burn tobacco and
play whist; in the country they sit idle in stores and bar-rooms, and
burn tobacco, and gossip and sleep. They complain of the flatness of
American life; “America has no illusions, no romance.” They have no
perception of its destiny. They are not Americans.
The felon is the logical extreme of the epicure and coxcomb.
Selfish luxury is the end of both, though in one it is decorated with
refinements, and in the other brutal. But my point now is, that this
spirit is not American.
Our young men lack idealism. A man for success must not be pure
idealist, then he will practically fail; but he must have ideas, must
obey ideas, or he might as well be the horse he rides on. A man
does not want to be sun-dazzled, sun-blind; but every man must
have glimmer enough to keep him from knocking his head against
the walls. And it is in the interest of civilization and good society and
friendship, that I dread to hear of well-born, gifted and amiable men,
that they have this indifference, disposing them to this despair.
Of no use are the men who study to do exactly as was done
before, who can never understand that to-day is a new day. There
never was such a combination as this of ours, and the rules to meet
it are not set down in any history. We want men of original perception
and original action, who can open their eyes wider than to a
nationality,—namely, to considerations of benefit to the human race,
—can act in the interest of civilization; men of elastic, men of moral
mind, who can live in the moment and take a step forward.
Columbus was no backward-creeping crab, nor was Martin Luther,
nor John Adams, nor Patrick Henry, nor Thomas Jefferson; and the
Genius or Destiny of America is no log or sluggard, but a man
incessantly advancing, as the shadow on the dial’s face, or the
heavenly body by whose light it is marked.
The flowering of civilization is the finished man, the man of sense,
of grace, of accomplishment, of social power,—the gentleman. What
hinders that he be born here? The new times need a new man, the
complemental man, whom plainly this country must furnish. Freer
swing his arms; farther pierce his eyes; more forward and forthright
his whole build and rig than the Englishman’s, who, we see, is much
imprisoned in his backbone.
’Tis certain that our civilization is yet incomplete, it has not ended
nor given sign of ending in a hero. ’Tis a wild democracy; the riot of
mediocrities and dishonesties and fudges. Ours is the age of the
omnibus, of the third person plural, of Tammany Hall. Is it that Nature
has only so much vital force, and must dilute it if it is to be multiplied
into millions? The beautiful is never plentiful. Then Illinois and
Indiana, with their spawning loins, must needs be ordinary.
It is not a question whether we shall be a multitude of people. No,
that has been conspicuously decided already; but whether we shall
be the new nation, the guide and lawgiver of all nations, as having
clearly chosen and firmly held the simplest and best rule of political
society.
Now, if the spirit which years ago armed this country against
rebellion, and put forth such gigantic energy in the charity of the
Sanitary Commission, could be waked to the conserving and
creating duty of making the laws just and humane, it were to enroll a
great constituency of religious, self-respecting, brave, tender, faithful
obeyers of duty, lovers of men, filled with loyalty to each other, and
with the simple and sublime purpose of carrying out in private and in
public action the desire and need of mankind.
Here is the post where the patriot should plant himself; here the
altar where virtuous young men, those to whom friendship is the
dearest covenant, should bind each other to loyalty; where genius
should kindle its fires and bring forgotten truth to the eyes of men.
It is not possible to extricate yourself from the questions in which
your age is involved. Let the good citizen perform the duties put on
him here and now. It is not by heads reverted to the dying
Demosthenes, or to Luther, or to Wallace, or to George Fox, or to
George Washington, that you can combat the dangers and dragons
that beset the United States at this time. I believe this cannot be
accomplished by dunces or idlers, but requires docility, sympathy,
and religious receiving from higher principles; for liberty, like religion,
is a short and hasty fruit, and like all power subsists only by new
rallyings on the source of inspiration.
Power can be generous. The very grandeur of the means which
offer themselves to us should suggest grandeur in the direction of
our expenditure. If our mechanic arts are unsurpassed in usefulness,
if we have taught the river to make shoes and nails and carpets, and
the bolt of heaven to write our letters like a Gillot pen, let these
wonders work for honest humanity, for the poor, for justice, genius
and the public good.[237] Let us realize that this country, the last
found, is the great charity of God to the human race.
America should affirm and establish that in no instance shall the
guns go in advance of the present right. We shall not make coups
d’état and afterwards explain and pay, but shall proceed like William
Penn, or whatever other Christian or humane person who treats with
the Indian or the foreigner, on principles of honest trade and mutual
advantage. We can see that the Constitution and the law in America
must be written on ethical principles, so that the entire power of the
spiritual world shall hold the citizen loyal, and repel the enemy as by
force of nature. It should be mankind’s bill of rights, or Royal
Proclamation of the Intellect ascending the throne, announcing its
good pleasure that now, once for all, the world shall be governed by
common sense and law of morals.
The end of all political struggle is to establish morality as the basis
of all legislation. ’Tis not free institutions, ’tis not a democracy that is
the end,—no, but only the means. Morality is the object of
government. We want a state of things in which crime will not pay; a
state of things which allows every man the largest liberty compatible
with the liberty of every other man.
Humanity asks that government shall not be ashamed to be tender
and paternal, but that democratic institutions shall be more
thoughtful for the interests of women, for the training of children, and
for the welfare of sick and unable persons, and serious care of
criminals, than was ever any the best government of the Old World.
The genius of the country has marked out our true policy,—
opportunity. Opportunity of civil rights, of education, of personal
power, and not less of wealth; doors wide open. If I could have it,—
free trade with all the world without toll or custom-houses, invitation
as we now make to every nation, to every race and skin, white men,
red men, yellow men, black men; hospitality of fair field and equal
laws to all.[238] Let them compete, and success to the strongest, the
wisest and the best. The land is wide enough, the soil has bread for
all.
I hope America will come to have its pride in being a nation of
servants, and not of the served. How can men have any other
ambition where the reason has not suffered a disastrous eclipse?
Whilst every man can say I serve,—to the whole extent of my being I
apply my faculty to the service of mankind in my especial place,—he
therein sees and shows a reason for his being in the world, and is
not a moth or incumbrance in it.
The distinction and end of a soundly constituted man is his labor.
Use is inscribed on all his faculties. Use is the end to which he
exists. As the tree exists for its fruit, so a man for his work. A fruitless
plant, an idle animal, does not stand in the universe. They are all
toiling, however secretly or slowly, in the province assigned them,
and to a use in the economy of the world; the higher and more
complex organizations to higher and more catholic service. And man
seems to play, by his instincts and activity, a certain part that even
tells on the general face of the planet, drains swamps, leads rivers
into dry countries for their irrigation, perforates forests and stony
mountain chains with roads, hinders the inroads of the sea on the
continent, as if dressing the globe for happier races.
On the whole, I know that the cosmic results will be the same,
whatever the daily events may be. Happily we are under better
guidance than of statesmen. Pennsylvania coal-mines and New York
shipping and free labor, though not idealists, gravitate in the ideal
direction. Nothing less large than justice can keep them in good
temper. Justice satisfies everybody, and justice alone. No monopoly
must be foisted in, no weak party or nationality sacrificed, no coward
compromise conceded to a strong partner. Every one of these is the
seed of vice, war and national disorganization. It is our part to carry
out to the last the ends of liberty and justice. We shall stand, then,
for vast interests; north and south, east and west will be present to
our minds, and our vote will be as if they voted, and we shall know
that our vote secures the foundations of the state, good will, liberty
and security of traffic and of production, and mutual increase of good
will in the great interests.
Our helm is given up to a better guidance than our own; the
course of events is quite too strong for any helmsman, and our little
wherry is taken in tow by the ship of the great Admiral which knows
the way, and has the force to draw men and states and planets to
their good.
Such and so potent is this high method by which the Divine
Providence sends the chiefest benefits under the mask of calamities,
that I do not think we shall by any perverse ingenuity prevent the
blessing.
In seeing this guidance of events, in seeing this felicity without
example that has rested on the Union thus far, I find new confidence
for the future.
I could heartily wish that our will and endeavor were more active
parties to the work. But I see in all directions the light breaking.
Trade and government will not alone be the favored aims of
mankind, but every useful, every elegant art, every exercise of the
imagination, the height of reason, the noblest affection, the purest
religion will find their home in our institutions, and write our laws for
the benefit of men.[239]
NOTES
THE LORD’S SUPPER

Mr. Emerson did not wish to have his sermons published. All that
was worth saving in them, he said, would be found in the Essays.
Yet it seemed best, to Mr. Cabot and to Mr. Emerson’s family, that
this one sermon should be preserved. A record of a turning-point in
his life, it showed at once his thought and his character; for he not
only gives the reasons why he believes the rite not authoritatively
enjoined, and hence recommends its modification or discontinuance,
but with serenity and sweetness renders back his trust into his
people’s hands, since he cannot see his way longer to exercise it as
most of them desire.
In the month of June, 1832, Mr. Emerson had proposed to the
church, apparently with hope of their approval, that the Communion
be observed only as a festival of commemoration, without the use of
the elements. The committee to whom the proposal was referred
made a report expressing confidence in him, but declining to advise
the change, as the matter was one which they could not properly be
called upon to decide.
The question then came back to the pastor, whether he was willing
to remain in his place and administer the rite in the usual form.
He went alone to the White Mountains, then seldom visited, to
consider the grave question whether he was prepared, rather than to
continue the performance of a part of his priestly office from which
his instincts and beliefs recoiled, to sacrifice a position of advantage
for usefulness to his people to whom he was bound by many ties,
and in preparation for which he had spent long years. He wrote, at
Conway, New Hampshire: “Here among the mountains the pinions of
thought should be strong, and one should see the errors of men from
a calmer height of love and wisdom.” His diary at Ethan Allan
Crawford’s contains his doubts and questionings, which Mr. Cabot
has given in his Memoir. Yet there was but one answer for him, and
after a fortnight, he came back clear in his mind to give his decision,
embodied in this sermon, to his people. On the same day that it was
preached, he formally resigned his pastorate. The church was loth to
part with him. It was hoped that some other arrangement might be
made. Mr. Cabot learned that “several meetings were held and the
proprietors of pews were called in, as having ‘an undoubted right to
retain Mr. Emerson as their pastor, without reference to the
opposition of the church.’ At length, after two adjournments and
much discussion, it was decided by thirty votes against twenty-four
to accept his resignation. It was voted at the same time to continue
his salary for the present.”
Thus Mr. Emerson and his people parted in all kindness, but, as
Mr. Cabot truly said, their difference of views on this rite “was in truth
only the symptom of a deeper difference which would in any case
sooner or later have made it impossible for him to retain his office; a
disagreement not so much about particular doctrines or observances
as about their sanction, the authority on which all doctrines and
observances rest.”
In the farewell letter which Mr. Emerson wrote to the people of his
church, he said:—
“I rejoice to believe that my ceasing to exercise the pastoral office
among you does not make any real change in our spiritual relation to
each other. Whatever is most desirable and excellent therein
remains to us. For, truly speaking, whoever provokes me to a good
act or thought has given me a pledge of his fidelity to virtue,—he has
come under bonds to adhere to that cause to which we are jointly
attached. And so I say to all you who have been my counsellors and
coöperators in our Christian walk, that I am wont to see in your faces
the seals and certificates of our mutual obligations. If we have
conspired from week to week in the sympathy and expression of
devout sentiments; if we have received together the unspeakable gift
of God’s truth; if we have studied together the sense of any divine
word; or striven together in any charity; or conferred together for the
relief or instruction of any brother; if together we have laid down the
dead in a pious hope; or held up the babe into the baptism of
Christianity; above all, if we have shared in any habitual
acknowledgment of the benignant God, whose omnipresence raises
and glorifies the meanest offices and the lowest ability, and opens
heaven in every heart that worships him,—then indeed we are
united, we are mutually debtors to each other of faith and hope,
engaged to persist and confirm each other’s hearts in obedience to
the Gospel. We shall not feel that the nominal changes and little
separations of this world can release us from the strong cordage of
this spiritual bond. And I entreat you to consider how truly blessed
will have been our connection if, in this manner, the memory of it
shall serve to bind each one of us more strictly to the practice of our
several duties.”
Page 18, note 1. The doctrine of the offices of Jesus, even in the
Unitarianism of Dr. Channing, was never congenial to Mr. Emerson’s
mind. He notes the same with regard to his father, and even to his
Aunt Mary, in spite of her Calvinism. Any interposed personality
between the Creator and the created was repugnant to him. Even in
March, 1831, he is considering in his journal that his hearers will say,
“To what purpose is this attempt to explain away so safe and holy a
doctrine as that of the Holy Spirit? Why unsettle or disturb a faith
which presents to many minds a helpful medium by which they
approach the idea of God?” and he answers, “And this question I will
meet. It is because I think the popular views of this principle are
pernicious, because it does put a medium, because it removes the
idea of God from the mind. It leaves some events, some things,
some thoughts, out of the power of Him who causes every event,
every flower, every thought. The tremendous idea, as I may well call
it, of God is screened from the soul.... And least of all can we believe
—Reason will not let us—that the presiding Creator commands all
matter and never descends into the secret chambers of the Soul.
There he is most present. The Soul rules over matter. Matter may
pass away like a mote in the sunbeam, may be absorbed into the
immensity of God, as a mist is absorbed into the heat of the Sun—
but the soul is the kingdom of God, the abode of love, of truth, of
virtue.”
Page 19, note 1. In the hope of satisfying those of his people who
held to the letter of the Scriptural Law, Mr. Emerson made the
foregoing clear statement with regard to the authority for the rite,
from the professional point of view. It seems quite unlike his usual
method, and there is little doubt that in it appears the influence of his
elder brother, William, whose honest doubts had led him to abandon
even earlier the profession of his fathers. In the introductory note to
the chapter on Goethe, in Representative Men, is given an account
of his unsuccessful pilgrimage to Weimar, in hopes that the great
mind of Germany could solve these doubts. There is a letter still
preserved, written by William, soon after his return, to his venerable
kinsman at Concord, Dr. Ripley, in which he explains with great
clearness his own reasons for not believing that the Communion rite
was enjoined by Jesus for perpetual observance. The argument on
scriptural grounds there clearly stated is substantially the same as
that which his younger brother makes use of in the beginning of this
sermon. Thus far he has spoken of outward authority; from this point
onward he speaks from within—the way native to him.
Page 25, note 1. Mr. Emerson left the struggles of the Past behind,
and did not care to recall them. Thus, writing of Lucretia Mott, whom
he met when giving a course of lectures in Philadelphia, in January,
1843, he said:—
“Me she taxed with living out of the world, and I was not much
flattered that her interest in me respected my rejection of an
ordinance, sometime, somewhere. Also yesterday—for
Philadelphian ideas, like love, do creep where they cannot go—I was
challenged on the subject of the Lord’s Supper, and with great
slowness and pain was forced to recollect the grounds of my dissent
in that particular. You may be sure I was very tardy with my texts.”
Mr. Emerson’s journal during the period of trial and decision, in the
mountains, shows that he was reading with great interest the life of
George Fox. The simplicity of the Society of Friends, their aversion
to forms and trust in the inward light, always appealed to him.
In his essay on The Preacher he says:—
“The supposed embarrassments to young clergymen exist only to
feeble wills.... That gray deacon, or respectable matron with
Calvinistic antecedents, you can readily see, would not have
presented any obstacle to the march of St. Bernard or of George
Fox, of Luther or of Theodore Parker.” This hints at the help he had
found in the Quaker’s history in his time of need.
HISTORICAL DISCOURSE AT CONCORD

Mr. Emerson’s Discourse was printed soon after its delivery, and
with it, in an Appendix, the following notice of the celebration of the
second centennial anniversary of the incorporation of the town, sent
to him by “a friend who thought it desirable to preserve the
remembrance of some particulars of this historical festival.”
“At a meeting of the town of Concord, in April last, it was

voted to celebrate the Second Centennial Anniversary of the
settlement of the town, on the 12th September following. A
committee of fifteen were chosen to make the arrangements.
This committee appointed Ralph Waldo Emerson, Orator, and
Rev. Dr. Ripley and Rev. Mr. Wilder, Chaplains of the Day.
Hon. John Keyes was chosen President of the Day.
“On the morning of the 12th September, at half past 10
o’clock, the children of the town, to the number of about 500,
moved in procession to the Common in front of the old church
and court-house and there opened to the right and left,
awaiting the procession of citizens. At 11 o’clock, the Concord
Light Infantry, under Captain Moore, and the Artillery under
Captain Buttrick, escorted the civic procession, under the
direction of Moses Prichard as Chief Marshal, from
Shepherd’s hotel through the lines of children to the Meeting-
house. The South gallery had been reserved for ladies, and
the North gallery for the children; but (it was a good omen) the
children overran the space assigned for their accommodation,
and were sprinkled throughout the house, and ranged on
seats along the aisles. The old Meeting-house, which was
propped to sustain the unwonted weight of the multitude
within its walls, was built in 1712, thus having stood for more
than half the period to which our history goes back. Prayers
were offered and the Scriptures read by the aged minister of
the town, Rev. Ezra Ripley, now in the 85th year of his age;—
another interesting feature in this scene of reminiscences. A
very pleasant and impressive part of the services in the
church was the singing of the 107th psalm, from the New
England version of the psalms made by Eliot, Mather, and
others, in 1639, and used in the church in this town in the
days of Peter Bulkeley. The psalm was read a line at a time,
after the ancient fashion, from the Deacons’ seat, and so
sung to the tune of St. Martin’s by the whole congregation
standing.
“Ten of the surviving veterans who were in arms at the
Bridge, on the 19th April, 1775, honored the festival with their
presence. Their names are Abel Davis, Thaddeus Blood, Tilly
Buttrick, John Hosmer, of Concord; Thomas Thorp, Solomon
Smith, John Oliver, Aaron Jones, of Acton; David Lane, of
Bedford; Amos Baker, of Lincoln.
“On leaving the church, the procession again formed, and
moved to a large tent nearly opposite Shepherd’s hotel, under
which dinner was prepared, and the company sat down to the
tables, to the number of four hundred. We were honored with
the presence of distinguished guests, among whom were
Lieutenant-Governor Armstrong, Judge Davis, Alden Bradford
(descended from the 2d governor of Plymouth Colony), Hon.
Edward Everett, Hon. Stephen C. Phillips of Salem, Philip
Hone, Esq., of New York, General Dearborn, and Lieutenant-
Colonel R. C. Winthrop (descended from the 1st governor of
Massachusetts). Letters were read from several gentlemen
expressing their regret at being deprived of the pleasure of
being present on the occasion. The character of the speeches
and sentiments at the dinner were manly and affectionate, in
keeping with the whole temper of the day.
“On leaving the dinner-table, the invited guests, with many
of the citizens, repaired to the court-house to pay their
respects to the ladies of Concord, who had there, with their
friends, partaken of an elegant collation, and now politely
offered coffee to the gentlemen. The hall, in which the
collation was spread, had been decorated by fair hands with
festoons of flowers, and wreaths of evergreen, and hung with
pictures of the Fathers of the Town. Crowded as it was with
graceful forms and happy faces, and resounding with the hum
of animated conversation, it was itself a beautiful living
picture. Compared with the poverty and savageness of the
scene which the same spot presented two hundred years
ago, it was a brilliant reverse of the medal; and could scarcely
fail, like all the parts of the holiday, to lead the reflecting mind
to thoughts of that Divine Providence, which, in every
generation, has been our tower of defence and horn of
blessing.
“At sunset the company separated and retired to their
homes; and the evening of this day of excitement was as
quiet as a Sabbath throughout the village.”
Within the year, Mr. Emerson had come to make his home for life
in the ancestral town, and had become a householder. Two days
after the festival, he drove to Plymouth in a chaise, and was there
married to Lidian Jackson, and immediately brought his bride to her
Concord home.
His aged step-grandfather was the senior chaplain at the
Celebration, and his brother Charles, who was to live with him in the
new home, was one of the marshals.
In preparation for this address Mr. Emerson made diligent
examination of the old town records, and spent a fortnight in
Cambridge consulting the works on early New England in the
College Library. I reproduce most of his references to his authorities
exactly, although there are, no doubt, newer editions of some of the
works.
Page 30, note 1. This story is from Bede’s Ecclesiastical History
(chapter xiii., Bohn’s Antiquarian Library). Mr. Emerson used it in full
as the exordium of his essay on Immortality, in Letters and Social
Aims.
Page 30, note 2. The poem “Hamatreya,” wherein appear the
names of many of these first settlers, might well be read in
connection with the opening passages of this address.
Mr. Emerson’s right of descent to speak as representative of Peter
Bulkeley, who was the spiritual arm of the settlement, as Simon
Willard was its sword-arm, may here be shown: Rev. Joseph
Emerson of Mendon (son of Thomas of Ipswich, the first of the name
in this country) married Elizabeth, daughter of Rev. Edward Bulkeley,
who succeeded his father, the Rev. Peter Bulkeley, as minister of
Concord. Edward, the son of Joseph of Mendon and Elizabeth
Bulkeley, was father of Rev. Joseph Emerson of Malden, who was
father of Rev. William Emerson of Concord, who was father of Rev.
William Emerson of Harvard and Boston, the father of Ralph Waldo
Emerson.
Page 31, note 1. Neal’s History of New England, vol. i., p. 132.
Page 31, note 2. Neal, vol. i., p. 321.
Page 31, note 3. Shattuck’s History of Concord, p. 158.
Page 32, note 1. On September 2, 1635, the General Court
passed this order:—
“It is ordered that there shalbe a plantac̃on att Musketequid & that
there shalbe 6 myles of land square to belong to it, & that the
inhabitants thereof shall have three yeares im̃ unities from all publ[ic]
charges except traineings; Further, that when any that plant there
shall have occac̃on of carryeing of goods thither, they shall repaire to
two of the nexte magistrates where the teames are, whoe shall have
the power for a yeare to presse draughts, att reasonable rates, to be
payed by the owners of the goods, to transport their goods thither att
seasonable tymes: & the name of the place is changed & here after
to be called Concord.”
Page 32, note 2. Shattuck, p. 5.
Page 33, note 1. In his lecture on Boston (published in the volume
Natural History of Intellect) Mr. Emerson gives an amusing
enumeration of some troubles which seemed so great to the
newcomers from the Old World: he mentions their fear of lions, the
accident to John Smith from “the most poisonous tail of a fish called
a sting-ray,” the circumstance of the overpowering effect of the sweet
fern upon the Concord party, and the intoxicating effect of wild
grapes eaten by the Norse explorers, and adds: “Nature has never
again indulged in these exasperations. It seems to have been the
last outrage ever committed by the sting-rays, or by the sweet fern,
or by the fox-grapes. They have been of peaceable behavior ever
since.”
Page 34, note 1. Johnson’s Wonder-Working Providence, chap.
xxxv. Mr. Emerson abridged and slightly altered some sentences.
Page 35, note 1. Mourt, Beginning of Plymouth, 1621, p. 60.
Page 35, note 2. Johnson, p. 56. Josselyn, in his New England’s
Rarities Discovered, speaks with respect of “Squashes, but more
truly squontersquashes; a kind of mellon, or rather gourd; ... some of
these are green; some yellow; some longish like a gourd; others
round, like an apple: all of them pleasant food, boyled and buttered,
and seasoned with spice. But the yellow squash—called an apple-
squash (because like an apple) and about the bigness of a pome-
water is the best kind.” Wood, in his New England Prospect, says:
“In summer, when their corn is spent, isquotersquashes is their best
bread, a fruit much like a pumpion.”
Page 36, note 1. Nashawtuck, a small and shapely hill between
the Musketaquid and the Assabet streams, at their point of union,
was a pleasant and convenient headquarters for a sagamore of a
race whose best roadway for travel and transportation was a deep,
quiet stream, the fish of which they ate, and also used for manure for
their cornfields along the bluffs. Indian graves have been found on
this hill.
Page 36, note 2. Josselyn’s Voyages to New England, 1638.
Page 36, note 3. Hutchinson’s History of Massachusetts, vol. i.,
chap. 6.
Page 36, note 4. Thomas Morton, New England Canaan, p. 47.
Page 37, note 1. Shattuck, p. 6.
The old Middlesex Hotel, which stood during the greater part of the
nineteenth century on the southwest side of the Common, opposite
the court- and town-houses, had fallen into decay in 1900, and was
bought and taken down by the town as an improvement to the public
square to commemorate the one hundred and twenty-fifth
anniversary of Concord Fight. It is probable that Jethro’s Oak, under
which the treaty was made, stood a little nearer the house of Rev.
Peter Bulkeley, the site of which, about one hundred paces distant
on the Lowell road, is now marked by a stone and bronze tablet.
Page 38, note 1. Depositions taken in 1684, and copied in the first
volume of the Town Records.
Page 39, note 1. Johnson’s Wonder-Working Providence.
Page 39, note 2. New England’s Plantation.
Page 39, note 3. E. W.’s Letter in Mourt, 1621.
Page 40, note 1. Peter Bulkeley’s Gospel Covenant; preached at
Concord in New England. 2d edition, London, 1651, p. 432.
Page 41, note 1. See petition in Shattuck’s History, p. 14.
Page 41, note 2. Shattuck, p. 14. This was the meadow and
upland on the Lowell road, one mile north of Concord, just beyond
the river. On the farm stands the unpainted “lean-to” house, now
owned by the daughters of the late Edmund Hosmer.
Page 42, note 1. Concord Town Records.
Page 43, note 1. Bancroft, History of the United States, vol. i., p.
389.
Page 44, note 1. Savage’s Winthrop, vol. i., p. 114.
Page 44, note 2. Colony Records, vol. i.
Page 44, note 3. See Hutchinson’s Collection, p. 287.
Page 46, note 1. Winthrop’s Journal, vol. i., pp. 128, 129, and the
editor’s note.
Page 46, note 2. Winthrop’s Journal, vol. ii., p. 160.
Page 48, note 1. Town Records.
With the exception of the anecdotes in this and the following
sentence, almost the whole of this account of the theory and practice
of the New England town-meeting was used by Mr. Emerson in his
oration, given in December, 1870, before the New England Society
in New York. The greater part of the matter used in that address is
included in the lecture on Boston, in the volume Natural History of
Intellect.
The New England Society of New York recently published the
Orations delivered before it previous to 1871, including Mr.
Emerson’s, as far as it could be recovered from the scattered
manuscript, and the newspaper reports of the time.
Page 50, note 1. Hutchinson’s Collection, p. 27.
Page 51, note 1. Shattuck, p. 20. “The Government, 13 Nov.,
1644, ordered the county courts to take care of the Indians residing
within their several shires, to have them civilized, and to take order,
from time to time, to have them instructed in the knowledge of God.”
Page 52, note 1. Shepard’s Clear Sunshine of the Gospel,
London, 1648.
Page 52, note 2. These rules are given in Shattuck’s History, pp.
22-24, and were called “Conclusions and orders made and agreed
upon by divers Sachems and other principal men amongst the
Indians at Concord in the end of the eleventh Month (called January)
An. 1646.”
The following are interesting specimens of these:—
Rule 2. “That there shall be no more Powwawing amongst the
Indians. And if any shall hereafter powwaw, both he that shall
powwaw, and he that shall procure him to powwaw, shall pay twenty
shillings apiece.”

Modern Marketing Research Concepts Methods and Cases 2Nd Edition Feinberg Solutions Manual Full Chapter PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Modern Marketing Research Concepts Methods and Cases 2Nd Edition Feinberg Solutions Manual Full Chapter PDF

Uploaded by

Copyright:

Available Formats

Modern Marketing Research Concepts

Methods and Cases 2nd Edition

The Use of Statistical Methods in Marketing Research pg 390

Overview of Data Analysis Procedures pg 392

To help identify which data analysis technique to use, determine:

Number of Variables to Analyze

Description versus Inference

Overview of Univariate Data Analysis Procedures

Descriptive Statistics pg 396

Measures of Central Tendency

Hypothesis Testing pg 400

The Concept of a Null Hypothesis

Table 8.1 Summary of Hypothesis Testing Errors

Reject H0 Type I error Correct decision

Type I Error (α)

Type II Error (β)

Steps in Hypothesis Testing

Inferential Statistics pg 404

If the population standard deviation () is known, then:

For proportions, the appropriate test statistic is:

Bivariate Procedures pg 410

Descriptive Statistics for Bivariate Analysis pg 411

Cov ( X , Y ) (X i − X )(Yi − Y )

Cov( X , X )Cov(Y , Y )  n 2

Partitioning the Sum of Squares

Figure 8.7 Partitioning Total Deviation in Simple Regression (page 417)

Fitting the Regression Line

The Meaning and Significance of the Coefficients

Measures of Regression Fit and Quality

Using the Regression Model to Predict

Inferential Statistics for Bivariate Analysis pg 424

Basic Inferential Test for Regression Coefficients

Inferential Test for Difference in Population Group Means

analysis of variance (ANOVA)

linear correlation coefficient

power of the test

simple linear regression

type I error (α)

type II error (β)

1 Distinguish between descriptive and inferential statistics.

z = (0.51 − ½) / [(½)(1 - ½)/500] = 0.447

The mean speed = 1501/24 = 62.54 mph.

The expected values are calculated with this formula:

Cell Oij Eij Oij – Eij (Oij – Eij)2 (Oij – Eij)2/Eij

The chi square test calculations:

3.84, but the p-value is 0.637, which is definitely not significant.

Cola 12.8 3.2

Case Gender Children at Respondent Vacations

Wald p-Value Two Odds-Ratio

Wald p-Value Two- Odds-Ratio

The regression analysis output is:

Regression 71.487 1 71.487 3.075 0.1101

Case Gender Children at Respondent Age Vacations

Children at Vacations Vacations

2 Distinguish between significance level and confidence level.

3 What is the power of the test?

8 For which of the following uses is the chi-square test valid?

Case Number of Home Years Lived in City

What is the relationship, if any, between the two variables?

Coefficients b Std. Error Std. Beta t-Test p-Value Two-Tailed

Hyperstat Online: http://davidmlane.com/hyperstat.

Introductory Statistics: Concepts, Models, and Applications,

The Little Handbook of Statistical Practice, http://www.tufts.edu/~gdallal/LHSP.HTM.