Announcements Thursday, Survey Analysis II due Tuesday, November 25, Lab day in 250 Saturday, December 13, 3:00-5:00 – Survey Research, Chapters 9,10,11,12,13.

Announcements

• Thursday, Survey Analysis II due• Tuesday, November 25, Lab day in 250• Saturday, December 13, 3:00-5:00

– Survey Research, Chapters 9,10,11,12,13– Statistical Analysis, Chapters 16,17, &18

Items and Scales

Measurement Part 2

Levels of Measurement

• Nominal coded data: Ability to distinctly categorize, or conversely, allows the determination of equality.

• Ordinal coded data: Allows determination of magnitude, rank, greater or less than.

• Interval coded data: Captures the distance apart two or more respondents are with respect to an attribute.

• Ratio data possess a natural or absolute zero, indicating a true absence of a characteristic. Permits statements concerning the equality of ratios.

Objectives

• Characteristics of questionnaire items• Distinguish between single-item measures versus

composite scales, “batteries of items.”• Qualities of measurements

– Validity– Sensitivity– Reliability and error

• Examples of scales used in marketing research

Total Error

Total Error

Poor Questionnaire

Inadequate samplesize

Inadequate sampledesign

Poor data collectionImproper design

Poor problemformulation

Poorly written report

Improper Use of Statistics

Clarifications

• The term questionnaire item is used to denote a single question on a survey, corresponding to a single column in a dataset.

• Scales typically denote sets of questions which become mathematical combinations of survey items.

Scaled-Response Format Questions

• Labeled: All positions, especially mid-point and intermediate positions for the respondent are marked, or bear “labels”

• Unlabeled: No intermediate points are labeled, sometimes positions bear only numbers.

Likert item: Labeled People should shop at local merchants even though the prices may be significantly higher.

1 . . . . . . 2 . . . . . . 3 . . . . . . 4 . . . . . . 5Stronglydisagree

NeutralDisagreeStrongly

agreeAgree

Numeric differential item: Unlabeled Paying higher prices at local merchants...

1 . . . . 2 . . . . 3 . . . . 4 . . . . 5 Showsintelligence

Showsignorance

Adding too many labels

1 . . . . 2 . . . . 3 . . . . 4 . . . . 5 . . . . 6 . . . . 7

Stronglydisagree

NeutralDisagreeStrongly

agreeAgreeDisagreesomewhat

Agreesomewhat

1 . . . . 2 . . . . 3 . . . . 4 . . . . 5 . . . . 6 . . . . 7

Never IrregularlyRarely AlwaysRegularlySeldom Often

Single-items adequate for measurement?

• Suppose an instructor had single-question exams?• Suppose the ACT (or GMAT) had only 5 possible

scores (similar to A,B,C,D,F grades)?• Suppose the ACT had only 4 questions, 1 each for

mathematics, English, reading comprehension, and science?

These are the issues behind marketing research use of multiple-item scales.

Composite, or Multiple-Item Scales• Capture the sensitivity to the continuous nature

of many subtle differences in between consumers.

• Simultaneously address concerns of:– Accuracy: Just as a 25-question exam can ask

questions evaluating the many topics covered in a course, multiple-item scale can ask the many subtle aspects that underlie a consumer attitude or behavior.

– Consistency: Though we do not administer the same surveys repeatedly to the same consumers, we strive to develop scales that consumers would provide consistent responses to over time.

• All relate to larger issue of measurement error.

Semantic Differential Scale

Indicate your impression of shopping at Dillard’s by checking the box corresponding to your opinion for each pair of descriptions.

Good valueUnfamiliar brandsDistinctive fashionLow qualityHelpful staff

Poor valueBest brandsNo fashionHigh qualityNo staff

Brand Personality: CompetenceHow do the following terms describe your view of the Cingular brand?

Reliable

Secure

Intelligent

Successful

Technical

Confident

Not at all descriptive __:__:__:__:__Extremely descriptive

Not at all descriptive __:__:__:__:__Extremely descriptive

Not at all descriptive __:__:__:__:__Extremely descriptive

Not at all descriptive __:__:__:__:__Extremely descriptive

Not at all descriptive __:__:__:__:__Extremely descriptive

Not at all descriptive __:__:__:__:__Extremely descriptive

Brand Personality: ExcitementHow do the following terms describe your view of the Cingular brand?

Trendy

Exciting

Imaginative

Up-to-date

Cool

Contemporary

Not at all descriptive __:__:__:__:__Extremely descriptive

Not at all descriptive __:__:__:__:__Extremely descriptive

Not at all descriptive __:__:__:__:__Extremely descriptive

Not at all descriptive __:__:__:__:__Extremely descriptive

Not at all descriptive __:__:__:__:__Extremely descriptive

Not at all descriptive __:__:__:__:__Extremely descriptive

VALIDITY AND RELIABILITY OF MEASUREMENTS

• Validity: Accuracy– Conceptual Issue– Face Validity– Predictive Validity

• Reliability: Consistency– Proportion of statistical error– Test-Retest– Inter-Item Correlation

• Increasing the number of questions (items) can increase both the validity and reliability of a scale.

• “I have bought “Joy” every time I have purchased dishwashing liquid in the last two years.

• “I will buy “Joy” whenever I receive a cents-off coupon.

• “I really notice the television ads for “Joy” dishwashing liquid.

• “I have a bottle of “Joy” in my kitchen cabinet.

• “I plan to purchase “Joy” on my next trip to the grocery store.”

SCALED-RESPONSE QUESTION FORMATS

Modified LikertSemantic DifferentialNumeric Differential

Composite ScalesFormative Composite ScaleReflective Composite Scale

Formative Scale Items: Satisfaction

1. “My last flight on TWA departed on-time.2. “TWA did not lose my luggage on my last trip.3. “TWA has competitive fares.”4. “TWA ticketing personnel are polite.5. “TWA has friendly reservation operators.6. “TWA has ample leg-room for me in coach seating.7. “I have not been “bumped” from a TWA flight in

the last two years.”

Additional items related to satisfaction

1. “I know it’s not the airline’s fault when a flight is cancelled.

2. “It upsets me to know others on the same flight have paid a lower price for their seat.

3. “An airline could always be on-time if they made that their priority.

4. “The two-item restriction on carry-on luggage is insensitive to the needs of today’s passengers.”

Reflective Items: Materialism• “I admire people who own expensive homes, cars,

and clothes.• “Some of the most important achievements in life

include acquiring material possessions.• “I don’t place much emphasis on the amount of

material objects people own as a sign of success.*• The things I own say a lot about how well I’m doing in

life.• I don’t pay much attention to the material objects

other people own.** Reverse coded

Please check the items in your kitchen:

• Electric frying pan• Food processor• Spring-form pan• 9” cake pan• Flour sifter• Hand-held electric

mixer

• 10” frying pan• 2-quart sauce pan• Rotary blender• Pressure cooker• Cookie sheet• Cake icing spatula

Formative and Reflective Items

• Formative items: Can be combined to measure the multiple aspects of a construct, though not necessary that respondents answer each item similarly.

• Reflective items: Measures a single trait and respondents should answer each item similarly.

Correlation Coefficient

• rxy, “the correlation between x and y”

• The denominator is always positive.• The numerator will be negative when…• The numerator will be positive when…

yx

iixy ssn

yyxxr

Percent Shared Variance

• If two “reflective” items correlation (rxy) is squared, the correlation ranges from 0 to 1.00 (instead of -1.00 to +1.00).

• This squared correlation (rxy2) is equivalent

to a percentage of shared variance.• If the two items are asking about a

common subject, this is viewed as the degree they measure the same “trait.”

Example: Farm Credit Survey

Correlations

1 .553** .515**

.000 .000

531 529 530

.553** 1 .781**

.000 .000

529 533 532

.515** .781** 1

.000 .000

530 532 534

Pearson Correlation

Sig. (2-tailed)

N

Pearson Correlation

Sig. (2-tailed)

N

Pearson Correlation

Sig. (2-tailed)

N

easily understood loaninformation

easy to do business with

staff/office are easilyaccessible

easilyunderstood

loaninformation

easy to dobusiness with

staff/officeare easilyaccessible

Correlation is significant at the 0.01 level (2-tailed).**.

“Percent” shared variance

• Correlations, .669, .718, .764

• Squaring the correlations: .448; .516; .583 or 44.8%; 51.6%; 58.3% shared variance in pairwise combinations of items.

•

Cronbach’s Alpha for k items

ncorrelatio item-inter average

ns)correlatio if (1.0 varianceaverage

827.616.)13(0.1

)]781.515.553(.3/1[3

)1(

c

v

ckv

ck

Correlations

1 .553** .515** .801**

.000 .000 .000

531 529 530 528

.553** 1 .781** .896**

.000 .000 .000

529 533 532 528

.515** .781** 1 .887**

.000 .000 .000

530 532 534 528

.801** .896** .887** 1

.000 .000 .000

528 528 528 528

Pearson Correlation

Sig. (2-tailed)

N

Pearson Correlation

Sig. (2-tailed)

N

Pearson Correlation

Sig. (2-tailed)

N

Pearson Correlation

Sig. (2-tailed)

N

easily understood loaninformation

easy to do business with

staff/office are easilyaccessible

easy3

easilyunderstood

loaninformation

easy to dobusiness with

staff/officeare easilyaccessible easy3

Correlation is significant at the 0.01 level (2-tailed).**.

Reliability Statistics

.825 3

Cronbach'sAlpha N of Items