Marketing Research Study Note
Marketing Research Study Note
Frequency Distribution
; Frequency distribution
; Measure of variability
; Statistics that indicate the distribution's dispersion
; range and variance or standard deviation .
; A mathematical distribution with the objective of obtaining a count of the number of responses
associated with different values of one variable and to express these counts in percentage terms.
; frequency distribution for a variable produces a table of frequency counts, percentages, and cumulative
percentages for all the for all the values associated with that variable.
; Conducting Frequency Analysis
; Calculating the frequency for each value of the variable
Range
; Range
; The difference between the smallest and largest values of a distribution
;
; Calculate the percentage and cumulative percentage for each value, adjusting for any missing values
; Plot the frequency histogram
; Calculate the descriptive statistics, measures of location, and variability
; A frequency distribution helps determine the extent of illegitimate responses. ( 0 8
; Standard deviation
; n-1 sample .
Mean
; Mean ()
; The average; that value obtained by summing all elements in a set and dividing by the number
; The average number of stores shopped for groceries is 3.0 per household
of elements.
; The department store is being patronized by more than 10% of the households
; The heavy and light users of a brand differ in terms of psychographic characteristics
; One hotel has a more upscale image than its close competitor.
; Familiarity with a restaurant results in greater preference for that restaurant.
Mode
; Mode ()
; A measure of central tendency given as the value the occurs the most in a sample distribution
; Mode is the value that occurs most frequently.
Median
(valid) . , Ho
; Median ()
; A measure of central tendency given as the value above which half of the values fall and
below which half of the values fall.
; The middle value when the data are arranged in ascending or descending rank order.
; Null hypothesis []
; A statement suggesting no expected difference or effect. If the null hypothesis is not rejected, no
changes will be made.
; Alternative hypothesis []
; A statement suggesting some difference or effect is expected. Accepting the alternative hypothesis
; If the variable is measured on an interval / ratio scale => mean is best, median is alternative.
. two-tailed test .
; MR one-tailed test .
; One-tailed test
; A test of the null hypothesis where the alternative hypothesis is expressed directionally
; )
a) Determine probability associated with test statistic (TScal) : Compare with level of significance, a
if probability of TScal < significance level(a), then reject Ho
b) Determine critical value of test statistic (TScr) : Determine if TScr falls into (Non) Rejection Region
if TScal > TScr, then reject Ho
; Critical value
; The value of the test statistic that divides the rejection and nonrejection regions. If the calculated
value of the test statistic is greater than the critical value of the test statistics, the null hypothesis is
rejected.
; Also known as alpha error, it occurs when the sample results lead to the rejection of a null
hypothesis that is in fact true.
; 95% ,
5% .
; Level of significance
Cross-Tabulations
; Cross-tabulation
; A statistical technique that describes two or more variables simultaneously and results in tables that
reflect the joint distribution of two or more variables that have a limited number of categories or distinct
values.
; Cross-tab categorize .
Type-II error
; Type-II error
; Also known as beta error, occurs when the sample results lead to nonrejection of a null
; Contingency tables
; Cross-tabulation tables; contain a cell for every combination of categories of the two variables.
Power of a Test
; Power of a test
; Cross-tab MR :
; 1) Cross-tabulation analysis and results can be easily interpreted and understood by managers who are
not statically oriented.
; 2) The clarity of interpretation provides a stronger link between research results and managerial action
; 3) Cross-tabulation analysis is simple to conduct and more appealing to less-sophisticated researcher.
; Bivariate cross-tabulation
; Cross-tabulation with two variables
; The probability of rejecting the null hypothesis when it is in fact false and should be rejected.
; reject ( )
; 1-b. b a . a (0.001)
b . 2 . a=0.05 .
a b , power of the test
. ( )
; 0~1 , c=0 . 1 . C
Chi-Square
table size .
; .
; Chi-square statistic Ho=There is no association between the variables
Cramer's V
; Chi-square statistic
;
The
statistic
used
to
test
the
statistical
significance
of
the
observed
association
in
cross-tabulation. It assists in determining whether a systematic association exists between the two
; Cramer's V
; A measure of the strength of association used in tables larger than 2 2
variables.
; Chi-square distribution
; A skewed distribution whose shape depends solely on the number of degrees of freedom. As the
; phi . V 0~1 .
number of degrees of freedom increases, the chi-square distribution becomes more symmetrical.
V , .
Cross-tabulation in practice
Construct the cross-tabulation table
Test the null hypothesis that there is no association between the variables using the chi-square statistics
If Ho is rejected, determine the strength of the association using an appropriate statistic (phi coefficient,
contingency coefficient, or Cramer's V)
Phi Coefficient
; Phi coefficient
; A measure of the strength of association in the special case of a table with two rows and two
columns
Contingency coefficient
; Contingency coefficient
; A measure of the strength of association in a table of any size.
If Ho is rejected, interpret the pattern of the relationship by computing the percentages in the direction of
the independent variable, across the dependent variable. Draw marketing conclusions.
[Discussion Problems]
[SPSS ]
1. In each of the following situations, indicate the statistical analysis you would conduct and the appropriate test
N
Valid
Missing
10
0
a. Respondents in a survey of 1,000 household were classified as heavy, medium, light, or nonusers of ice
Mean
64.0000
cream. They were also classified as being in high-, medium-, or low-income categories. Is the consumption of
Median
75.0000
Mode
80.00
Std. Deviation
33.73096
Variance
Minimum
1137.778
.00
Maximum
110.00
; Data , ,
test .
a = 0.05
df = (r-1)(c-1) = 6
2. The current advertising campaign for a major soft drink brand would be changed if less than 30 percent of the
consumers like it.
a. Formulate the null and alternative hypotheses.
b. Discuss the type-I and type-II errors that could occur in hypothesis testing.
Type I error : Error occurred by bad measurement. Error created by chance.
Type II error : Not rejecting Ho which has be rejected. It can occur by small sample size and bad
measuring procedure. You will miss your opportunity because you didn't do anything.
3. A major department store chain is having an end-of-season sale on refrigerators. The number of refrigerators
sold during this sale at a sample of 10 stores was
80
110 0
40
70
80
100 50
80
30.
a. Compute the mean, mode, and median. Which measure of central tendency is most appropriate in this case
and why?
5. A research project examining the impact of income on the consumption of gourmet foods was conducted. Each
variable was classified into three levels of high, medium, and low. The following results were obtained.
Income
Consumption of
Gourmet Foods
Low
Medium
High
Low
25
15
10
Medium
10
25
15
High
15
10
25
= 9.488
b. For the purpose of cross-tabulation, classify respondents as light or heavy users. Those reporting 5 hours or
less usage should be classified as light users and the remaining as heavy users. Run a cross-tabulation of sex
and Internet usage. Interpret the results. Is Internet usage related to one's sex?
df = (r-1)(c-1) = 1
= 3.841
b. Is the relationship between income and consumption of gourmet food strong?
Yes. It has strong relationship.
Reject Ho if
c. What is the pattern of the relationship between income and consumption of gourmet food?
Value
Pearson Chi-Square
df
Asymp. Sig.
Exact Sig.
Exact Sig.
(2-sided)
(1-sided)
3.333(b)
(2-sided)
.068
.143
.072
6. A pilot survey was conducted with 30 respondents to examine Internet usage for personal(non-professional)
Continuity Correction(a)
2.133
.144
reasons. The following table contains the resulting data given each respondents' sex, familiarity with the Internet,
Likelihood Ratio
3.398
.065
Internet usage in hours per week, attitude toward Internet and toward technology, both measured on a seven-point
scale, whether the respondents have done shopping or banking on the Internet.
Association
N of Valid Cases
3.222
.073
30
a. Obtain the frequency distribution of familiarity with the Internet. Calculate the relevant statistics.
So Ho will not be rejected. (in 95% confidence interval)
90% (a = 0.1) ,
Ho will be rejected.
.
; z-Test
; A univariate hypothesis test using the standard normal distribution.
; Hypothesis testing procedures that assume the variables of interest are measured on at least an interval
scale.
The t Distribution
Two-Sample t-Tests
Two Independent Samples
; Examples
; Parametric tests provide inferences for making statements about the means of parent populations.
; The populations of users and nonusers of a brand differ in terms of their perceptions of the brand.
; t-Test
; A univariate hypothesis test using that t distribution, which is used when the standard deviation is
; The proportion of brand loyal users in Segment I is more than the proportion in Segment II
; The proportion of households with an Internet connection in the US exceeds that in Germany.
; t statistic
; A statistic that assumes the variable has a symmetric bell-shaped distribution and the mean is known
; 2 .
.
(or assumed to be known), and the population variance is estimated from the sample.
; Independent samples
; t distribution
; A symmetric, bell-shaped distribution that is useful for small sample (n<30) testing.
; Two samples that are not experimentally related. The measurement of one sample has no effect on
the values of the other sample.
Means
; Hypothesis .
(t distributed with n-1 degrees of freedom)
[1] (If both populations are found to have the same variance) Pooled variance estimate is computed
One-Sample t-Tests
; examples of One-sample t-test
; The market share for the new product will exceed 15%.
; At least 65% of customers will like the new package design.
degrees of freedom = .
[2] (If the two populations have unequal variances) an exact t cannot be computed for the difference
in sample means.
; t . df
Paired Samples
; Examples
; Shoppers consider brand name to be more important than price when purchasing fasion clothing
; Households spend more money on pizza than on hamburgers.
; The proportion of households who subscribe to a daily newspaper exceeds the proportion
subscribing to magazines.
; The proportion of a bank's customers who have a checking account exceeds the proportion who
; F-test
; A statistical test of the equality of the variance of two populations.
; 2 .
; Paired samples
; In hypothesis testing, the observations are paired so that the two sets of observations relate to the
same respondents.
Hypotheses are:
; 2 .
Means
; Paired-samples t-test
; F statistics
; .
( paired-difference .)
n1 = size of sample 1
n2 = size of sample 2
Where
; F distribution
; A frequency distribution that depends upon two sets of degrees of freedom: the degrees of
freedom in the numerator and the degrees of freedom in the denominator.
Proportions
Proportions
; chi-square !
; Analysis of variance(ANOVA)
; A statistical technique for examining the differences among means for two or more populations.
; Ho mean equal .
The t-test is equivalent to a chi-square test for independence in a 2x2 contingency table. The
relationship:
; examples
; Do the various segments differ in terms of their volumes of product consumption?
; Do the brand evaluations of groups exposed to different commercials vary?
; Do retailers, wholesalers, and agents differ in their attitudes toward the firm's distribution policies?
; Do the users, nonusers, and former users of a brand differ in their attitudes toward the brand?
= mean for category j
Measurement of Effects
; X Y (effect) SSx . SSx X SSx
X Y . X Y
; eta2( )
; Factor
; Categorical independent variable; the independent variable must be categorical (nonmetric) to use
ANOVA.
; Treatment
; In ANOVA, a particular combination of factor levels or categories.
Significance Testing
; one-way ANOVA Ho .
; In one-way ANOVA, separation of the variation observed in the dependent variable into the
variation due to the independent variables plus the variation due to error.
; SSy
; Mean square
; A test to determine whether some differences exist between some of the treatment groups.
; Also denoted as SSx, the variation in Y related to the variation in the means of the categories of
X. This represents variation between the categories of X, or the portion of the sum of squares in Y
related to X.
; (c-1), (N-c) df F distribution .
; SSwithin
; Also referred to as SSerror, the variation in Y due to the variation within each of the categories of
X. This variation is not accounted for by X.
; .
or
or
= individual observation
Sample
Test/comments
One sample
Means
[Discussion Problems]
2. The current advertising campaign for a major automobile brand would be changed if fewer than 70% of the
Proportions
z-test
Means
Proportions
Ho :
Ha :
Paired samples
Means
Paired t-test
Proportions
Chi-square test
Ho will be rejected if
b. Which statistical test would you use? Why?
One-sample t-test will be used. Because we test for a single proportion.
c. A random sample of 300 consumers was surveyed, and 204 respondents indicated that they liked the
campaign. Should the campaign changed? Why?
a = 0.05
Proportions
Chi-square test
= 0.74
3. A major computer manufacturer is having an end-of-season sale on computers. The number of computers sold
during this sale at a sample of 10 stores was 800 1100 0 400 700 800 1000 500 800 300
a. Is there evidence that an average of more than 500 computers per store were sold during this sale? Use a
= 0.05
a = 0.05
Ho will be rejected if
[SPSS ]
Test Value = 500
t
Sales
1.313
df
Sig. (2-tailed)
.222
Mean
Difference
the Difference
Lower
Upper
140.000
-101.30
381.30
Ho )
c. If the average importance was computed or each group, would you expect the sample means to be similar
or different?
4. After receiving complaints from readers, your campus newspaper decided to redesign its front page. Two new
formats, B and C, are developed and tested against the current format, A. A total of 75 students are randomly
selected, and 25 students are randomly assigned to each of three format conditions. The students are asked to
evaluate the effectiveness of the format on a 11-point scale.
a. State the null hypothesis.
6. In a pilot study examining the effectiveness of three commercials (A, B, and C), 10 consumers were assigned to
view each commercial and rate it on a nine-point Likert scale. The data obtained are shown in the following
c. What are the degrees of freedom that are associated with the test statistic?
One-way analysis of variance testing will be used (because more than 2 variables are involved)
a = 0.05
Reject Ho if P(F) < 0.05
Reject Ho if = 3.10
comm
Mean
1.00
4.0000
10
Std. Deviation
.81650
2.00
5.0000
10
1.05409
3.00
7.0000
10
1.05409
Total
5.3333
30
1.58296
* :
Sum of
df
Mean Square
Sig.
Between Groups
Within Groups
Squares
46.667
26.000
2
27
23.333
.963
24.231
.000
Total
72.667
29
importance attached to shopping by consumers living in the northern, southern, eastern, and western United States.
A study is conducted and analysis of variance is used to analyze the data. The results obtained are presented in
the following table.
c. Calculate .
= 0.642
Source
df
Between groups
70.212
23.404
Within groups
996
20812.416
20.896
F ratio
F probability
1.12
0.3
a = 0.05
a = 0.05
Reject Ho if
Reject Ho if
Reject Ho if
So Ho will be rejected.
df
Sig. (2-tailed)
Mean
Difference
the Difference
Lower
Upper
1.00
10
4.0000
2.00
10
5.0000
3.00
10
3.225
29
.003
1.10000
.4025
1.7975
7.0000
Sig.
.093
1.000
1 2 , 3 .
8. In a pretest, respondents were asked to express their preference for an outdoor lifestyle(V1) using a seven point
scale. They were also asked to indicate the importance of the following variables on a seven-point scale.
a = 0.05
Reject Ho if = 2.0484 (df=28, a=0.05/2)
Equality of Variances
F
Sig.
df
Std. Error
95% Confidence
Sig.
Mean
Interval of the
Difference
a = 0.05
Lower
Upper
Equal
variances
Reject Ho if
2.746
.109
.092
28
.928
.06667
.72681
-1.42214
1.55547
.092
25.980
.928
.06667
.72681
-1.42737
1.56070
assumed
preferen
Equal
variances
not
Test Value = 3
t
preferen
2.893
df
29
Sig. (2-tailed)
assumed
Mean
Difference
the Difference
Lower
Upper
.007
1.03333
.3029
1.7638
Do not reject Ho. Keep Ho.('cause zero is included!)
So Reject Ho. ('cause zero is not included!)
d. Does the importance attached to V2 through V6 differ for males and females?
95% Lower~Upper 0 .
[SPSS] , category respondent Independent samples t-test .
b. Does the mean importance of enjoying nature exceed 3.5?
[SPSS] One-sample t-test .
correlation
coefficient .
; Examples
; How strongly are sales related to advertising expenditures?
; Is there an association between market share and size of the sales force?
[ ]
Significance . 95% CI 0
Significance .
nature : SIG
weather : N/S
environment : SIG
exercise : N/S
people : SIG
* SIG : Significant (0 is not included)
; Covariance
; A systematic relationship between two variables in which a change in one implies a corresponding
change in the other (COVxy)
; Example) Correlation [p499] [ ]
; r=1.0 . (-1.0
e. Do the respondents attach more importance to enjoying nature than they do to relating to the weather?
95% CI 0 . .
Yes.
; Since r indicates the degree to which variation in one variable is related to variation in another, it can
also be expressed in terms of the decomposition of the total variation.
f. Do the respondents attach more importance to relating to the weather than that they do to meeting other
people?
95% CI 0 . .
No. (no difference in preference)
; r r2 (symmetric) . X Y correlation Y
X correlation .
g. Do the respondents attach more importance to living in harmony with the environment than they do to
exercising regularly?
95% CI 0 . .
Yes.
; The statistic significance of the relationship between two variables measured by using r can be conveniently
; Least-squared procedure
tested.
; A technique for fitting a straight line to a scattergram by minimizing the vertical distances of all
(t distribution with df=n-2)
; Example) t test [p502] [ ]
Regression Analysis
; Regression analysis ( )
; A statistical procedure for analyzing associative relationships between a metric-dependent variable and
= the error term associated with the ith observation.
Bivariate Regression ( )
; Bivariate regression
; A procedure for deriving a mathematical relationship, in the form of an equation, between a single
metric-dependent variable and a single metric-independent variable.
Estimation of Parameters
; (Sample observation) .
;
Estimated or predicted value , a b estimator .
b nonstandardized regression coefficient .
; Can variation in sales be explained in terms of variation in advertising expenditure? What is the
structure and form of this relationship, and can it be modeled mathematically by an equation describing a
; a b
straight line?
; Can the variation in market share be accounted for by the size of the sales force?
; Are consumers' perceptions of quality determined by their perceptions of price?
; b a .
; Example) [p506-507] [ ]
; example) r [p510-511] [ ]
a variance of 1.
; X Y coefficient of determination
; z value .
.
;
; Also known as beta weight, used to denote the standardized regression coefficient
; Byx = the slope obtained by the regression of Y on X
Prediction Accuracy
; Standard error of estimate (SEE)
; The standard deviation of the actual Y values from the predicted Y values.
; "The larger the SEE is, the poorer the fit of the regression"
Significance Testing
; X Y (significance testing) .
Examination of Residuals
; Residual
; The difference between the observed value of Y and the value predicted by the regression equation
; X Y . t stastistic
Y.
. (df=n-2)
Multiple Regression ( )
; Multiple Regression
; A statistical technique that simultaneously develops a mathematical relationship between two or more
; Standard error
; SEb denotes the standard deviation of b and is called the standard error.
; Coefficient of determination
; The proportion of variance in one variable associated with the variability in a second variable.
; Total variation :
; Explained Variation :
; Are consumers' perceptions of quality determined by their perceptions of prices, brand images, and
; Residual Variation :
brand attributes?
; .
; How much of the variation in sales can be explained by advertising expenditures, prices, and level of
distribution?
; What is the contribution of advertising expenditures in explaining the variation in sales when the levels
distribution?
; What levels of sales may be expected given the levels of advertising expenditures, prices, and level of
; .
; The strength of association :
; .
=0 .
; F statistics .
...
; equation :
1) The sample size n <= The number of independent variables k
2) One independent variable is perfectly corrrelated with another.
Strength of Association
; .
; R2 r2 . .
R2 r2 . R2
. adjusted R2 .
; Adjusted R2
2
; The value of R adjusted for the number of independent variables and the sample size.
Adjusted R2 =
Significance Testing
[Discussion Problems]
1. A major supermarket chain wants to determine the effect of promotion on relative competiveness. Data were
obtained from 15 states on the promotional expenses relative to a major competitor (competitor expenses = 100)
and on sales relative to this competitor (competitor sales = 100). You are assigned the task of telling the manager
whether there is any relationship between relative promotional expense and relative sales.
f. If the company matched the competitor in terms of promotional expense (if the relative promotional expense
was 100), what would the company's relative sales be?
* .
2
a. Plot the relative sales (Y-axis) against the relative promotional expense (X-axis), and interpret this
diagram.
2. To understand the role of quality and price in influencing the patronage of drugstores, 14 major stores in a
large metropolitan area were rated in terms of preference to shop, quality of merchandise, and fair pricing. All the
ratings were obtained on an 11-point scale, with higher numbers indicating more positive ratings.
a. Run a multiple regression analysis explaining store preference in terms of quality of merchandise and
pricing.
Bivariate Regression .
c. Run a bivariate regression analysis of relative sales on relative promotional expense.
[ ] : , .
3. Imagine that you've come across a magazine article reporting the following relationship between annual
expenditure on prepared dinners (PD) and annual income (INC)
PD = 23.4 + 0.003 INC
The coefficient of the INC variable is reported as significant.
a. Does this relationship seem plausible? Is it possible to have a coefficient that is small in magnitude and
yet significant?
Sales = f(const, promo) = -9.768 + 1.175(promo)
b. From the information given, can you tell how good the estimated model is?
R2 = 0.969
It is plausible.
[ ]
Sales = f(const, promo)
But the problem is that it has "0.003" in coefficient. It is problematic that we don't know the range and
drivers.
5. Conduct the following analyses for the preference of the outdoor-lifestyle data.
6. In a pretest, data were obtained from 20 respondents on preferences for sneakers on a seven-point scale, 1 =
a. Calculate the simple correlations between V1 to V6 and interpret the results.
not at all preferred, 7 = greatly preferred (V1). The respondents also provided their evaluations of the sneakers on
comfort (V2), style (V3), and durability (V4), also on seven-point scales, 1 = poor, and 7 = excellent. The
resulting data follow.
a. Calculate the simple correlations between V1 to V4 and interpret the results.
Preference-people,
Nature-environment,
Nature-people,
Weather-exercise,
b. Run a bivariate regression with preference for sneakers (V1) as the dependent variable and evaluation on
comfort (V2) as the independent variable. Interpret the results.
Weather-people correlation .
b. Run a bivariate regression with preference for an outdoor lifestyle (V1) as the dependent variable and the
importance of enjoying nature (V2) as the independent variable. Interpret the results.
.
c. Run a bivariate regression with preference for sneakers (V1) as the dependent variable and evaluation on
style (V3) as the independent variable. Interpret the results.
R2=0.016, .
beta .132 + .
c. Run a multiple regression with preference for an outdoor lifestyle as the dependent variables. Interpret the
results. Compare the coefficients for V2 obtained in the bivariate and the multiple regressions.
.
d. Run a bivariate regression with preference for sneakers (V1) as the dependent variable and evaluation on
durability (V4) as the independent variable. Interpret the results.
.
[ ] (b-d)
R2
P(f)
.291
.008
(style)
.380
.002
(durability)
.274
.010
pref = f (comfort)
e. Run a multiple regression with preference for sneakers (V1) as the dependent variable and V2 to V4 as the
independent variables. Interpret the results. Compare the coefficients for V2, V3, and V4 obtained in the
bivariate and the multiple regressions.
[SPSS] Enter style , Stepwise style
.
[Enter ]
[Stepwise ]