Marketing Research

1
Marketing Research

• Dr. David M. Andrus
• Exam 2
• Lecture 3

2
Themes of My Presentation

• Attitudes and Measurement
• Nominal, Ordinal, Interval, and Ratio Scales
• Components of Measurement
• Comparative and Noncomparative Rating Scales
• Reliability and Validity

3
What is an Attitude?

• A learned tendency to respond in a consistently
  favorable or unfavorable manner toward something.

• Three key components
• Cognitive A persons beliefs or perceptions
  about something.
• Affective An individuals positive or negative
  feelings toward something.
• Behavioral A persons behavioral intention
  toward something.

4
What is Measurement?

• Assigning numbers to characteristics according to
  specified rules to reflect the characteristics
  that the marketing variable possess.

5
Types of Measurement Scales

• Scale- Nominal
• Basic Empirical Operations- Classify into groups,
  assign numbers, and count frequency
• Typical Usage- Classification Gender,
  Purchase-Non-purchase
• Descriptive- Percentages, Mode, Lamda
• Inferential- Chi-Squared Test

6
Types of Measurement Scales

• Scale- Ordinal
• Basic Empirical Operations- Determination of
  Greater or Less
• Typical Usage- Rankings Brand Preference,
  Market Position, Attitude Measures
• Descriptive- Median, Gamma, Tau A,B,C
• Inferential- Mann-Whitney U, Two-way Anova, Rank
  Order

7
Types of Measurement Scales

• Scale- Interval
• Basic Empirical Operations- Determination of
  Equality of Intervals
• Typical Usage- Price, Attitude Measures, Level of
  Knowledge of Brands
• Descriptive- Mean, Range, Standard Deviation
• Inferential- Product-Moment Correlation, T-Tests,
  Factor Analysis, MANOVA

8
Types of Measurement Scales

• Scale- Ratio
• Basic Empirical Operations- Determination of
  Equality of Ratios, True Zero Point
• Typical Usage- Sales, Units Produced, Number of
  Customers, Marketing Costs, Age
• Descriptive- All previous Statistics
• Inferential- Coefficient of Variation

9
Components of Measurements

• True Characteristic- Direct reflection of the
  characteristic of interest.
• Additional Stable Characteristics of the
  Respondent- Reflection of other permanent
  characteristics such as social class.
• Short Term Characteristics of the Respondent-
  Reflection of temporary characteristics such as
  hunger or anger.

10
Components of Measurements

• Situational Characteristics- Reflection of the
  surroundings in which the measurement is taken.
• Characteristics of the Measurement Process-
  Reflection of the interviewer and interviewing
  method.
• Characteristics of the Measuring Instrument-
  Reflection of ambiguous or misleading questions.
• Characteristics of the Response Process-
  Reflection of mistaken replies caused by checking
  the wrong response.
• Characteristics of the Analysis- Reflection of
  mistakes in coding, tabulating.

11
Major Data-Collection Methods
TYPES OF SCALES
Comparative Rating Scales
Noncomparative Rating Scales
Paired Comparison
Rank Order
Graphic Rating
Itemized Rating
Constant Sum
Q-Sort
Likert
Semantic Differential
Stapel
12
Comparative Rating Scales

• Paired Comparison
• Rank Order
• Constant-Sum
• Q-Sort

13
Paired-Comparison

• Highly popular technique
• Respondents are asked to choose which item rates
  higher, according to a predetermined criterion.
• The requisite number of individual items
  demonstrates a shortcoming of the
  paired-comparison technique.
• When several comparisons are required, the
  technique becomes less effective and less
  accurate due to respondent fatigue.

14
Rank-Order

• Popular technique
• Respondents rank products according to some
  predetermined criterion
• Easy to implement
• Forces respondents to provide rankings only from
  the pairs provided
• Gives only ordinal scale data
• Reduced usefulness when many items must be ranked

15
Constant-Sum Scale

• Respondents allocate a predetermined number of
  rating points among several items, according to
  some criterion.
• Indicates relative importance or preference of
  each item compared to all others on list.
• Does not require a large number of individual
  comparisons.
• Number of alternatives must be manageable.
• Respondents must allocate points in a way that
  indicates their relative preference among items,
  which may not be realistic to require.
• It has not been established that the data
  produced uses an interval scale.

16
Q-Sort Scale

• Discriminates among a large group of items in a
  relatively short time.
• Respondents rank a group of items into sets,
  according to some criterion, such as preference,
  attitude, or behavioral intent.
• Cards are the most popular and simplest items to
  use in the sorting process.
• Each card notes an item to be ranked.
• To increase statistical reliability, at least 60
  cards should be used, but no more than 140.

17
Advantages/Disadvantages of Q-Sort Scale

• Advantages
• Short time process to complete
• Systematic manner in which preferences are
  categorized
• Ease of implementation
• Disadvantages
• Vast number of choices requested from the
  respondent in a relatively short period of time
• Uncertain reliability as the number of cards in
  the process increases

18
Comparative Graphic Rating Scales
Compared to your favorite soft drink (Coca-Cola,
Mountain Dew , my ideal soft drink), how would
you rate the taste of Pepsi?
1.
Excellent
Very Poor
Compared to your favorite soft drink (Coca-Cola,
Mountain Dew, my ideal soft drink), would you
rate the taste of Pepsi?
2.
All right,
Not at all
Probably Very good, neither good good. I
do Probably the best I like It nor bad not
like it the worst
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
19
Noncomparative Rating Scales

• Graphic rating scales
• Itemized rating scales

20
Graphic Rating Scale

• Responses are indicated on a continuum.
  Respondents place a mark at a location on the
  continuum that reflects their response to the
  question. Presentation may be horizontal or
  vertical, qualitative or quantitative.
• Excellent----------------------Poor
• Best----------------------Worst

21
Noncomparative Graphic Rating Scales
Overall, how would you rate the taste of Pepsi?
Excellent
Very Poor
Overall, how would you rate the taste of Mountain
Dew?
All right,
Not at all
Probably Very good, neither good good. I
do Probably the best I like It nor bad not
like it the worst
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
22
Examples of Itemized Rating Scales
1.
Balanced, forced-choice, even-interval scale
focusing on an attitude.
2.
Overall, how would you rate Crest Toothpaste?
Extremely Very Somewhat Somewhat
Very Extremely
Good Good Good Bad
Bad Bad
23
Examples of Itemized Rating Scales
Unbalanced, forced-choice, odd-interval scale
focusing on an overall attitude.
3.
4.
Neither Friendly nor Unfriendly
Very Friendly
Moderately Friendly
Moderately Unfriendly
Very Unfriendly
Dont Know
24
Likert Scale

• Summated rating scale
• Respondents select from choices ranging from
  strongly agree to strongly disagree.
• Choices are assumed to have equal distances
  between them and may be numbered.
• When using Likert scales, a series of statements
  that relate to the topic of interest are
  developed and administered using the strongly
  agree to strongly disagree choices.
• Responses may be analyzed either individually or
  on a total (summated) basis.

25
Likert Scale Survey Without Numeric Scores
1. First Bank provides excellent customer
service. Strongly Neither agree Strongly agree
Agree nor disagree Disagree disagree -----
----- ----- ----- ----- 2. First
Bank has convenient locations. Strongly Neither
agree Strongly agree Agree nor
disagree Disagree disagree ----- -----
----- ----- -----
26
Example of Likert Scale
Neither Agree nor Disagree
Strongly Disagree
Strongly Agree
Disagree
Agree
1. The celebrity endorser is
trustworthy.
2. The celebrity endorser is attractive.
3. The celebrity endorser is an expert on
the product.
27
Semantic Differential Scale

• Five- or seven-point itemized ordinal scale with
  dichotomous (opposite) pairs of descriptive words
  or phrases representing the two extremes and a
  neutral point.
• Extreme points are meant to be interpreted as
  strongly or extremely.
• To guard against respondents marking responses in
  the same place down a column without carefully
  reading each descriptive pair, researchers often
  place favorable adjectives and phrases randomly
  on either the left or the right ends of the
  scale.
• Shortcomings include
• Developing dichotomous descriptive words and
  phrases is not always possible.
• Neutral responses are difficult to interpret.

28
Semantic Differential Scale for Pepsi-Cola
Sweet ____ ____ ____ ____ ____ Sour Tasty ____ ___
_ ____ ____ ____ Tasteless Satisfying ____ ____ __
__ ____ ____ Unsatisfying Expensive ____ ____ ____
____ ____ Inexpensive
29
Example of Semantic Differential Scale
Celebrity Endorser
Some- Some-
Extremely Quite what Neither what Quite
Extremely
Not Trustworthy Unattractive Not
Expert Not Knowledgeable
Trustworthy Attractive Expert Knowledgeable
30
Stapel Scale

• Mirrors the semantic differential scale, with a
  few exceptions.
• Instead of using two dichotomous descriptive
  words or phrases as choices, only one word or
  phrase is used.
• Although points are not assigned numbers in a
  semantic differential scale, they are assigned
  numbers in a Stapel scale.
• The downside of the Stapel scale is the potential
  biasing of the respondents by the word choice of
  the categories.

31
Stapel Scale for Pepsi-Cola
5 5 5 5 4 4 4 4 3 3 3 3 2 2 2
2 1 1 1 1
Sweet Tasty Satisfying Expensive
-1 -1 -1 -1 -2 -2 -2 -2 -3 -3 -3 -3 -4 -4 -4 -
4 -5 -5 -5 -5
32
Example of Stapel Scale
-5 -4 -3 -2 -1 1 2
3 4 5
1. The celebrity endorser is
trustworthy.
2. The celebrity endorser is attractive.
3. The celebrity endorser is an expert on
the product.
33
Reliability
Refers to the ability of a scale to produce
consistent results if repeated measurements are
taken.

• Test-Retest Reliability Conducted by measuring
  the same subjects at two different times and
  under similar conditions to determine the
  similarity of the scores.
• Often difficult to persuade the original
  respondents to take a test a second time.
• The halo effect - it is unreasonable to expect
  respondents to answer questions a second time
  without considering their initial responses.
• Some situations may be measured only once.
• Alternative-Forms Reliability (a.k.a.,
  equivalent forms reliability) Ability of two
  equivalent scales to obtain consistent results.
• Problem is constructing two scales that appear
  different yet have similar content.

34
Reliability

• Internal Consistency Reliability Two or more
  measurements of the same concept are taken at the
  same time and then compared to see whether they
  agree.
• Split-Half Technique Test items are randomly
  split into two equal groups and their degree of
  correlation is examined.
• Coefficient Alpha (a.k.a., Cronbachs Alpha)
  Technique for judging internal consistency of a
  measurement by averaging all possible ways of
  splitting test items and examining their degree
  of correlation the greater the correlation is to
  a score of 1, the higher the internal consistency.

35
Improving Reliability

• Increase the number of measurements
• Use good experimental controls
• Be careful to select only items relevant to the
  topic for measurement

36
Validity
Degree to which a test measures what it is
supposed to measure.

• Face Validity A measurement seems to measure
  what it is supposed to measure.
• Content Validity The degree to which the
  instrument items represent the universe of a
  concept under study.
• Do the items in your questionnaire constitute a
  representative sample of the content area
  studied? Ask professionals or experts on the
  test topic to assess the scale.
• Criterion-Related Validity Ability of a scale
  to perform as predicted in relation to a
  specified criterion. The criterion is the
  attribute of interest.
• Concurrent Validity Evaluates how well the
  results from one scale correspond with the
  results from another when the scales measure the
  same phenomenon at the same point in time.
• Predictive Validity Ability of a scale to
  predict a future occurrence or phenomenon based
  upon a current measurement scale. We often use
  purchase intention questions

37
Validity

• Construct Validity A construct is a highly
  abstract, unobservable, hypothesized concept. Due
  to their nature, there are no direct ways to
  measure constructs, so researchers measure
  observable phenomena that theoretically
  demonstrate the presence of the construct of
  interest. Construct validity assesses how well
  ideas or theories are translated into real
  measures. How well do your questionnaire items
  measure the construct?
• Convergent Validity Ability of a scale to
  correlate with other scales that purport to
  measure the same concept, the logic being that
  two or more measurements of the same concept
  using different scales should highly agree or
  covary if they are valid measures of the concept.
• Discriminant Validity Opposite of convergent
  validity. A scales lack of correlation with
  another scale that purports to measure different
  concepts. The results of two scales measuring
  unrelated concepts should display no correlation.

38
Relationship Between Reliability and Validity

• Reliability is a necessary, but not a sufficient
  condition for validity.
• Validity is NOT a necessary condition for
  reliability.
• Survey data can not be more valid than reliable
  because highly reliable data can be invalid due
  to systematic basis.
• Validity is a matter of degree. Marketing
  research is partly valid and partly invalid.