Funds of Market Research

Ordinal Scale

Ranking scale that maintains labeling characteristics of nominal scales and has the ability to order data. Can determine whether an object has more or less of a characteristic than some other object, but not how much more or less. The number do not indicate absolute quantities. Ordinal numbers do not imply equal interval between the numbers. Arithmetical operations cannot be used. Statistics: counting (as with nominal) but also the use of statistics based on centiles, e.g. percentiles, quartiles. Possesses description and order but not distance or origin. Ordinal scales are used to measure the order of consumer's preferences. Ex: please rank the 5 department stores in order of your preference. Assign 1 to your most preferred store and assign 5 to your least preferred store.

Itemized Rating Scale Decisions Forced or Nonforced Scale

1. Should your scale include a "no opinion" option? 2. Should you force a response -- e.g. must answer to continue (qualtrics will allow you to do this) In a forced rating scale, a neutral category, if included, attracts not only those who are neutral but also those who have no opinion. Marking a neutral point when in fact, "no opinion" is the desired responses will distort the measures. Without a "no opinion" option, the respondent has to concentrate on their feelings, they must answer However, a "no opinion" can be easy out for the lazy respondent.

3 steps to cluster sampling

1. divide the target population into mutually exclusive and collectively exhaustive clusters (e.g. neighborhoods) 2. select a random sample of clusters based on probability sampling. 3. For each stage selected cluster, either include all the elements (one - stage) or a random sample of elements (two - stage) -- cluster should be as homogeneous as possible. Ideally, each cluster should be small - scale representation of the population. -- elements within a cluster should be as heterogeneous as possible.

Paired Comparison Scaling:

A respondent is presented with two objects and asked to select one according to some criterion. It is the most widely used comparative scaling technique. The data obtained are ordinal in nature. Under the assumption of transitivity, it is possible to convert paired comparison data to a rank order. Pros: less order bias, it is easy for people to select one item from a set of two than to rank a large set of objects. Cons: interviewee fatigue -- with n brands, [n(n-1)/2)] paired comparisons are required.

Representativeness

A sample should accurately reflect distribution of relevant variables in population representativeness is essential to generalize. If the sample does not represent the population well, generalizations can not be made. Sampling error results when the sample selected is not perfectly representative of the population.

Sample

A subgroup of the population -- use statistics to make inferences about the population parameters -- inferences link sample characteristics with population parameters

Content Validity

A subjective but systematic evaluation of how well the content of a scale represents the measurement task. Does the scale provide adequate coverage of the topic under study? Systematic efforts including exhaustive literature review, focus groups, expert panels helps determine content validity. Measuring service quality Ex: literature review and experts have revealed that consumers tent to evaluate service quality along 5 dimensions: tangibles, reliability, empathy, assurance, and attentiveness.. does your measure of service quality include each of these?

Stratified Sampling

A two step process: 1. Divide the original population into homogeneous, mutually exclusive, and exhaustive subset (e.g. male and female) 2. Elements of each stratum are then randomly selected The stratification variables -- the stratification variable should be closely related to the characteristics of interest -- the variables should decrease the cost of stratification process by being easy to measure and apply. Advantages: -- can acquire information about whole population and individual strata -- can lead to increased precision Disadvantages: -- information necessary to properly stratify the sample is usually not available.

Which of the following statements is not true about the comparative scaling technique? A. an example of using comparative scales is asking respondents to evaluative Coke on a 1 to 6 preference scale. B. comparative scale data must be interpreted in relative terms and have only ordinal or rank order properties. C. comparative scales include paired comparisons, rank order, and constant sum scales. D. all of the above are true statements about comparative scaling techniques E. none of the above are true statements about comparative scaling techniques.

A. an example of using comparative scales is asking respondents to evaluative Coke on a 1 to 6 preference scale.

Validity

Are we measuring what we think we are measuring? (Does out personality scale actually measure "personality"

An interval level scale has which of the following characteristics? A. relative sizes or positions of descriptors B. Equal intervals between scale descriptors C. Meaningful/absolute zero point D. All of the above E. None of the above

B. Equal intervals between scale descriptors

How is cluster different from stratified?

Cluster: only a sample of subpopulations (clusters) is chosen, objective: increase sampling efficiency to decrease costs. Stratified -- all subpopulations (strata) are chosen -- objective: increase precision

Examples Noncomparative Scales

Continuous rating scale, likert scale, semantic differential, stapel scale.

Criterion Validity (convergent)

Criterion validity reflects whether a scale performs as expected in relation to other variable selected (criterion variables) as meaningful criteria. Measuring service quality theories predict that service quality is positively associated with customer loyalty. Measuring service quality and customer loyalty simultaneously, and assessing the correlation between them to determine the criterion validity of the scale.

Research design

Cross sectional (given sample only once but can be multiple different samples) vs. longitudinal (over time, SAME sample). Error: sampling (sample imperfect representation of popn) vs. non sampling error (problem defn, scales, interviewing, data) - non sampling: non response (not responding) vs response error (inaccurate answers, misrecording, measurement).

Sampling process

Define the target population --> determine the sampling frame --> select sampling techniques(s) ---> determine the sample size ---> execute the sampling process

Scale characteristics

Description - unique labels or descriptors for each value Order - relative sizes or positions of descriptors Distance - absolute differences between scale descriptors Origin - unique or fixed beginning

Itemized Rating Scale Decisions Nature and Degree of Verbal description

Description may be provided for each category or only at the end points of the scale. The strength of the adjectives used to anchor the scale has a considerable effect on the response. Please indicate your degree of agreement or disagreement with the statement: the moral values in America are declining Completely disagree -- completely agree.

The sampling distribution

Distribution of values of sample statistic computed for all possible samples that could be drawn from the target population.

Simple Random sampling

Each element in the population has known and equal probability of selection Each possible sample of a given size (n) has a known and equal probability of being the sample actually selected Advantages: -- simple -- sampling error easily measured Disadvantages: -- need complete list of units, which is not usually available -- does not always achieve best representativeness -- units may be scattered and poorly accessible

Defining the target population

Element: 18 year old males Sampling unit: households with 18 year old males Time frame: upcoming summer Extent (geographic): domestic United States

Measurement validity

Face: judgement call, does the measurement seem to measure what it is supposed to measure (face/image) Content: subjective but systematic evaluation of how well content of a scale represents measurement task (resume) Criterion: reflects whether a scale performs as expected in relation to other variables (criterion variables)(math test) Construct: how does your scale perform relative to other constructs (similar to different from, other related constructs)

Graphic Rating Scale

How would you rate Sears as a department store? You'll be given a line graph to drag a point up and down.

Itemized Rating Scale Decision Balanced vs. Unbalanced Scale

In a balanced scale, the number of favorable and unfavorable categories is equal. If the researcher suspect that the responses are likely to be skewed either negatively or positively, an unbalanced scale might be appropriate Under these conditions, more categories are included in the direction of the skewness. "Rate your performance" Much below average --> excellent

Nonprobability sampling

In nonprobability sampling, the change of being selected into the sample is not precisely known. The resulting sample is not representative of any definable target population Generalization can not be made. Statistic inference from sample to the population of interest is problematic. Low cost. Fast and easy implementation

Snowball Sampling

In snowball sampling, an initial group of respondents is selected, usually at random. After being interviewed, these respondents are asked to identify others who belong to the target population of interest. Subsequent respondents are selected based on the referrals Referred respondents tend to have demographic and psychographic characteristics that are more similar to the person referring them than would occur by chance. It is used when studying characteristics that are relatively rare or difficult to identify in the population.

Testing Reliability

Internal consistency reliability = cronbach's alpha Determine the extent to which different parts of a measurement scale are consistent in what they indicate about the characteristics being measured.

Test Reliability

Internal consistency reliability: determines the extent to which different parts of a measurement scale are consistent in what they indicate about the characteristic being measured.

(Research Design) Validity

Internal validity: refers to whether the manipulation of the independent variables or treatment actually caused the observed effects on the dependent variables. Control of extraneous variables is a necessary condition for establishing internal validity. Did my manipulation produce the effects, or was it something else? External validity: refers to whether the cause and effect relationships found in the experiment can be generalized. To what populations, settings, times, independent variables, and dependent variables can the results be projected. External validity also applies to survey research. No internal validity = no cause and effect inference.

Nominal, ordinal, interval, or ratio How satisfied are you with your new purchase? 1 = very satisfied 2 = satisfied 3 = neutral 4 = dissatisfied 5 = very dissatisfied

Interval

Reliability

Is the measure consistent? (e.g. if we measured personality today and tomorrow how close would the two results be?

Itemized Rating Scales

Itemized rating scale have a limited number of ordered categories with brief description associated with each category. The respondents are required to select the specified category that best describes the object being rated. The commonly used itemized rating scales in market research are the: Likert (most popular) Semantic differential

Why do we need to know scales type?

Knowing scale type enables us to apply APPROPRIATE SCALES to appropriate marketing variables. Knowing scales type informs us what STATISTICAL ANALYSIS to performance in data analysis stage.

Nominal, ordinal, interval, or ratio Please circle the vegetables have you consumed in the last week? (you can circle more than one) Broccoli Carrots Beets Radishes Tomatoes

Nominal

Nominal, ordinal, interval, or ratio What is your gender? Male Female

Nominal Why? Only possesses description

Marketing examples of the 4 major scales

Nominal: Brand nos, store types. Ordinal: preference rankings, market position, social class. Interval: Attitudes, opinions, index. Ratio: age, sales, income, costs.

Levels of Measurement

Nominal: categories or labels (e.g. gender) Ordinal: order, relative position not magnitude (e.g. ranking) Interval: equal intervals (e.g. most scales) Ratio: meaningful zero (e.g. age, income)

Itemized Rating Scale Decisions

Number of Scale categories: Balancing the scale sensitivity and information processing demands imposed on the respondents. - respondent's interest and knowledge about the objects. - mode of data collection - space limitations - how the data are to be analyzed No single optimal number. Traditional guidelines suggest five to nine categories.

Itemized Rating Scale Decisions Odd or Even Numbers of Categories.

ODD or EVEN numbers of categories: Odd numbers of categories: a neutral category. Even number of categories: no neutral category. Respondents are forced to indicate some degree of positive or negative feelings on an issue. Use a neutral point if the researcher has reason to believe that a portion of the respondents is actual neutral on a particular subject. Be cautious of the meaning of the midpoint: neutral vs. no opinion.

A rank order question on a survey would be which type of measurement scale? a. nominal b. ordinal c. interval d. ratio

Ordinal

Nominal, ordinal, interval, or ratio Please allocate a total of 100 points to indicate your preference for these department stores (higher number = preferred) _______ Nordstrom _______ Sears _______ Kohl's _______ Macy's

Ordinal Does not have numerically equal distances or absolute zero

Nominal, ordinal, interval, or ratio? What is your age? what about if this is the response option? Please circle your age. 15 and under 16-30 31-50 51 or above

Ordinal Why not nominal? - because it has order. Why not interval? - because it doesn't have numerically equal distances on the scale, you cannot determine how much more or less of a characteristic an object possesses. Why not ratio? - because it doesn't have clear, true or origin.

Nominal, ordinal, interval or ratio What is your average annual income? $0 - $10,000 $10,001 - $20,000 $20,001 and above

Ordinal Why? It has description and order but not distance or origin

Can an unreliable measure be valid? Can an invalid measure be reliable?

Reliability is a necessary, but not sufficient, condition for validity.

Constant Sum Scaling

Respondents allocate a constant sum of units, such as 100 points, to attributes of a product to reflect their importance. The number of points reflects both the rankings and the relative magnitude of each alternatives as perceived by the respondent. Pros: avoid long lists of paired items. Respondents are allowed to indicate equal value to two alternatives if they perceive so. Cons: respondents may have difficulty allocating the points to total 100 if there are a lot of items. 10 items is the upper limit. EX: Below are eight attributes of bathing soaps, please allocate 100 points among the attributes so that your allocation reflects the relative importance you attach to each attribute, the more points an attribute receives, the more important the attribute is.

Rank Order Scaling

Respondents are presented with several objects simultaneously and asked to order or rank them according to some criterion. Data obtained are ordinal in nature. Pros: instructions easy to understand. Less scaling decision (n-1) need to be made. Cons: possible that the respondent may dislike the brand ranked 1 in an absolute sense. Order bias may exist. EX: rank the various brands of toothpaste in order of preference. Being by picking one brand that you like most and assign it a number 1.

Noncomparative Scaling Techniques

Respondents evaluate only one object at a time, and for this reason non-comparative scales are often referred to as monadic scales. Non - comparative techniques consist of continuous and itemized rating scales.

Continuous Rating Scale

Respondents rate the objects by placing a mark on a graphic continuum anchored by two extremes. Once the respondent has provided the ratings, the researcher divides the line into as many categories as desired and assigns scores. This type of scale produces interval data. Pros: easy to construct. It can be easily implemented on the internet based survey. Cons: scoring graphic scale can be cumbersome and unreliable if using paper and pencil questionnaire. Limited use in marketing research.

Convenience Sampling

Sample is selected based on the convenience of the researcher. Often, respondents are selected because they happen to be in the right place at the right time. The resulting sample is not representative of any definable target population. It is not theoretically meaningful to generalize to any population from a convenience sample Ex: use of students, mall intercept interviews without qualifying the respondents.

Ratio Scale

Scales that have the characteristics of interval scale, plus a meaningful zero point. Possesses all the properties of the nominal, ordinal, and interval scales. It has an absolute zero point. It is meaningful to compute ratios of scale values. All statistical techniques can be applied to ratio data. Possesses description, order, distance, and origin. Ex: sales, costs, market share, and number of customers are variables measured on a ratio scale. EX: please indicate the market share of product A and product B in Northwest region in 2011 Product A: Product B:

Interval Scale

Scales that have the characteristics of ordinal scales, plus equal intervals between points. Equal distance on the scale represent equal values in the characteristics being measured. It permits comparison of the differences between objects. The location of zero point is not fixed. Both the zero point and the units of measurement are arbitrary. It is not meaningful to take ratios of scale values. Statistics: interval data permit the use of the arithmetic mean, standard deviation, correlation, and other statistics. Possesses description, order and distance but not origin. Ex: In marketing research, attitudes, motivations, and emotions are variables measured on a interval scale. Equal intervals is normally assumed in these measures. EX: Overall, how would you rate the the quality of the services at Flat Tail Pub and Brewery? 1: Poor 2: Fair 3: Good 4: Very Good 5: Excelent

Nominal Scale

Scales that partition data into mutually exclusive and collectively exhaustive categories. The number serve only as labels or tags for identifying and classifying objects. The numbers do not reflect the amount of the characteristic possessed by the objects. When used for identification, there is a strict one-to-one correspondence between the numbers and the objects. Statistics: limited based on frequency counts, percentage of objectives in each category. No arithmetic operation. Possesses descriptive but not order, distance, or origin. Many demographics variables such as gender, ethnic group, occupations, are measured by nominal scales, we also assign a unique number to each response. Ex: Sex: Male (1) Female (2) Occupation: (1) student (2) faculty (3) staff (4) otheers Of the 243 respondents, 123 are male, and 120 are female. 53% of the respondents are students, 22% are faculty members, and the remaining 25% are staff and others.

Adjusting for non responses substitution

Substitute nonrespondent with substitutes who are similar to nonrespondents

The most common paired comparison selling

Taste testing, the consumer is asked to sample two different products and select the one with the most appealing taste. The test is done in private and a minimum of 1,000 responses is considered an adequate sample. A blind taste test for a soft drink, where imagery, self perception and brand reputation are very important factors in the consumer's purchasing decision, may not be a good indicator of performance in the marketplace. The introduction of New Coke illustrates this point. New coke was heavily favored in blind paired comparison taste tests, but its introduction was less than successful, because image plays a major role in the purchase of Coke.

Face Validity

The degree to which a measurement seems to measure what is is supposed to measure, as judged by researchers. This is the weakest way to try and demonstrate construct validity because it won't be very convincing to others. Dressing up for a job interview.

Measurement Reliability

The extent to which scale produces consistent results if repeated measurement are made on the characteristic. Reliability can be defined as the extent to which measures are free from random error. A key question to ask: if we measure same phenomenon over and over again with the same measurement device, will we get the same or highly similar results?

Construct Validity

The extent to which what the researcher was trying to measure was actually measured. Perfect validity requires that the measurement instrument and procedure are free from both systematic and random error. A key question: is what we observed actually measuring what we thought?

True experimental designs

The researcher can randomly assign test units to experimental groups and treatments to experimental groups. Pretest - posttest control group design Posttest: only control group design.

Judgmental Sampling

The researcher choose the sample elements because she believes they represent the population of interest Judgmental sampling is subjective, relying largely on the expertise of the researcher Generally, generalized to a specific population cannot be made Ex: test markets selected to determine the potential of a new product

Systematic sampling

The sample is chosen by selecting a random starting point and then picking every ith element in succession from the sampling frame The sampling interval, i, determined by dividing the population size N by the sample size n and rounding to the nearest integer. Advantages: -- systematic sampling is less costly and easier than SRS, because random selection in only done once. -- when the ordering of the element is related to the characteristic of interest, systematic sampling increases the representativeness of the sample. Disadvantages: -- Need complete list of units, which is not usually avaliable

Semantic Differential Scale

The semantic differential scale is seven point rating scale with end points associated with bipolar adjective labels. Pros: sufficient reliability, validity, and statistical robustness when applied to corporate image research. Widely used in comparing brand, product, and company images. Cons: Lack of standardization: the scale must be adapted for each research problem. Halo effect: the rating of a specific image component may be dominated by the interviewee's overall impression of the concept being rated. Modern 1 2 3 4 5 6 7 Old fasion

Define the target population

The target population is the collection of elements or objects that possess the information sought by the researcher and about which inferences are to be made The target population should be defined in terms of elements, sampling units, extent, and time An element is the object about which or from which the information is desired, usually the respondent e.g. male or female head of the household responsible for most of the shopping at department stores A sampling unit is an element, or a unit containing the element, that is available for selection at some stage of the sampling process. Extent refers to the geographical boundaries time is the time period under consideration

How we infer causality?

To infer: suggest, to hint, to derive as a conclusion from facts or premises. Must satisfy 3 conditions: 1. concomitant variation -- two variables must occur together. 2. time order of occurrence of variables -- causing event must occur before or simultaneously with effect; it cannot occur after. 3. Elimination of other possible causal factors -- eliminate as many causes as possible. All three of these must exist to infer causality.

Likert Scale

Typically, respondents are asked to indicate a degree of agreement or disagreement with each of a series of statements. Respondents readily understand how to use the scale. Suitable for mail, phone, personal interview, or online survey. It takes longer to complete than other itemized scales because respondents have to read each statement. The analysis can be conducted on an item - by - item basis (profile analysis), or a total (summated) score can be calculated. (this is when you are asked if you strongly disagree -- strongly agree scale.

Validity and Reliability

Validity: are we measuring what we think we are measuring (e.g. does our personality scale actually measure "personality" Reliability: is the measure consistent? (e.g. id we measured personality today and tomorrow how close would the two results be? THE TARGET EXAMPLE: All the dots are centered: highly reliable and valid. The dots are off the target but still bunched: highly reliable but not valid. The dots are all over the place: neither reliable nor valid.

Quota Sampling

Viewed as two stage judgmental sampling The first stage consists of developing control categories, or quotas, of population elements In the second stage, sample elements are selected based on convenience or judgement. Attempts to obtain representativeness at relatively low cost but.. No assurance that sample is representative Does not permit assessment of sampling error

`Noncomparative Scales

`Measurement scales in which judgment is made without reference to another object, concept, or person. The resulting data are generally assumed to be interval or ratio scale.

Multi - Item Scales

Why is the same thing asked with different (but similar meaning) words? Attitudes: Very bad ..... Very good Susceptibility to interpersonal influence (likert) I often consult other people to help choose the best alternative available from a product class. If I want to be like someone, I often try to buy the same brands that they buy. It is important that others like the products and brands I buy. I rarely buy the latest fashion style until I am sure my friends approve of them;.

If respondents are asked to allocate 100 points to attribute of a toilet bowl cleaner in a way that reflects the importance they attach to each attribute, _______ scaling is being used. a. constant sum b. Q - sort c. paired comparison d. rank order

a. constant sum

How would you rate Sears a department store? Probably the worst ________________^_______________ probably the best A. continuous rating scale B. stapel scale C. semantic differential scale D. likert scale

a. continuous rating scale

Income is always a ratio - level scale of measurement a. true b. false

b. false

A ________ is a measurement scale with 5 response categories ranging from "strongly disagree" to "strongly agree", which requires the respondents to indicate a degree of agreement or disagreement with each of a series of statements related to the stimulus objects. a. semantic differential scale b. likert scale c. continuous rating scale d. stapel scale

b. likert

The ranking of teams in a tournament in which numbers are assigned to teams to indicate their placing constitute a ______ scale. a. nominal b. ordinal c. interval d. ratio

b. ordinal

In marketing research, attitudinal data obtained from rating scales are often treated as _______- data. a. nominal b. ordinal c. interval d. ratio

c. interval

When used for classification purposes, the _______ scaled number serve as labels for classes or categories a. ordinally b. intervally c. nominally d. ratio

c. nominally

Adjusting for non responses Subsampling of nonrespondents

contact a subsample of the nonrespondents

In a department store study, numbers 1 through 10 were assigned to identify the 10 stores considered in the study. Thus, store number 9 referred to Sears and store number 6 referred to Neiman Marcus. Using this information, which of the following statements is true? a. sears is in some way superior or inferior to Neiman Marcus b. it is meaningful to state that the number of the average store is 5.5 c. both a and b are true d. none of the above statements are true

d. none of the above statements are true

Cluster sampling

divide population into clusters --> random sample of clusters ----> include all elements from each selected cluster. (one stage cluster sampling) Random sample of clusters --> randomly sample elements from each selected cluster. (two stage cluster sampling)

Adjusting for non responses subjective estimates

evaluate likely effects of nonresponse based on experience and available info

Adjusting for non responses Trend analysis

examine trend between early and late respondents.

Comparative Scales

measurement scales in which one object, concept, or person is compared with another on a scale. Comparative scale data must be interpreted in relative terms and have only ordinal properties.

Population parameter

numerical characteristic of a population

Statistic

numerical characteristic of a sample (e.g. mean)

Adjusting for non responses replacement

replace nonrespondents in the current survey with nonrespondents from an earlier, similar survey

Potential sources of error on measurement

respondent factors: such as intelligence, social desirability, and education. short term: or personal factors, such as health, emotions, and fatigue. Situational factors: such as the presence of other people, noise, and distractions. Items: in the scale: addition, deletion, or changes in the scale items. Lack of clarity: of the scale, including the instructions or the items themselves. Mechanical factors: such as overcrowding items the questionnaire and poor design. Administration of the scale: such as difference among interviewers. Analysis factors: such as difference in scoring and statistical analysis.

Census

surveying the ENTIRE population of interest -- population parameters can be calculated directly from the census.