Qualitative Research and Attitude Measurement


Published on

The elements of Qualitative Resaerch, Attitude Measurement and Using Scaling.

Published in: Business, Technology

Qualitative Research and Attitude Measurement

  1. 1. The MAANZ MXpress ProgramQualitative Research and Attitude Measurement TechniquesDr Brian MongerCopyright June 2013.This Power Point program and the associated documents remain the intellectual property and thecopyright of the author and of The Marketing Association of Australia and New Zealand Inc. Thesenotes may be used only for personal study and not in any education or training program. Persons and/orcorporations wishing to use these notes for any other purpose should contact MAANZ for written permission.
  2. 2. MAANZ International• MAANZ International, is a Not for Profit, internet based professional and educational institute which has operated for over 25 years.• MAANZ International offers Professional Memberships;• Marketing Courses (Formal and Short)• And Marketing Publications• www.marketing.org.au 2MAANZ International
  3. 3. • For more information about MAANZ International and articles about Marketing, visit:• www.marketing.org.au• http://smartamarketing.wordpress.com• http://smartamarketing2.wordpress.com• .  http://www.linkedin.com/groups/MAANZ‐SmartaMarketing‐Group‐2650856/about• Email: info@marketing.org.au• Link to this site ‐ ‐ http://www.slideshare.net/bmongerfor further presentationsMarketing In Black and White 3
  4. 4. Dr. Brian Monger• Brian Monger is the CEO of MAANZ International and a Professional marketer and consultant with over 40 years experience.MAANZ International 4
  5. 5. 5Introduction to Qualitative Research• A common feature of marketing research is theattempt to have respondents communicate theirfeelings, attitudes, opinions, and evaluations insome measurable form.• Qualitative research focuses on the nature orstructure of attitudes and motivations rather thantheir frequency and distribution.MAANZ International
  6. 6. 6There are two basic methods of qualitative research.  • The group depth interview, commonly called afocus group, assembles eight to twelve respondentswith a trained moderator who guides a discussion,generally lasting about two hours.• The individual depth interview (or IDI) collectsinformation serially in one-on-one sessions thatgenerally lasts about 45 minutes to an hour.MAANZ International
  7. 7. 7Why conduct qualitative research?  • Although qualitative research does not tell you howwidely distributed an attitude or motivation might be,it does tell you, and in ways surveys cannot-fromwhere those attitudes arise, how they are structured,and what broader significance they may have forconsumer behaviour.• The rich insights that come from close inspection ofindividuals can never be duplicated by large-scaleconclusive forms of research, which view the marketfrom a more distant vantage point.MAANZ International
  8. 8. 8Limitations on the Reliability of a Qualitative Study• Sample size• Sampling Procedure• Question structure and sequenceMAANZ International
  9. 9. 9Understanding Buyers• An understanding of buyers attitudes is meant to explainthe why aspect of buyers past, present and futurebehaviour.• Such understanding can be achieved through somecomplicated means, which necessitate a more subtleapproach than direct questioning.• It is so primarily important because attitude variables (e.g.belief, preference, motivation and intention) areconceptual ideas which are not only hard to explore, butalso difficult to record the responses accurately andcomprehensively.MAANZ International
  10. 10. 10Understanding Buyers• Attitudes are multidimensional, there is alwaysmore than one aspect that make up ones attitudes.• The researcher should examine all relevant aspectsif he or she wishes to evaluate buyers attitudestowards a product, brand or service.• The results of Qualitative Research can often berecorded quantitatively. That is by using scalingtechniques.MAANZ International
  11. 11. 11Attitude Measurement• Attitudes are usually viewed as a long term or enduring disposition toconsistently respond in a given manner to various aspects of the world(within a similar situation or context) including persons, events, andobjects. There are three components of attitude:• Affective component: reflects an individualsgeneral feelings towards an object.• Cognitive component: represents ones awarenessof and knowledge about an object.• Behavioural component: reflects buying intentionsand behavioural expectations.MAANZ International
  12. 12. 12Attitudes as Hypothetical Constructs• The term hypothetical construct is used todescribe a variable that is not directly observable,but is measurable by an indirect means such asverbal expression or overt behaviour.• Attitudes are considered to be such variables.MAANZ International
  13. 13. 13The Attitude‐Measuring Process• There are a variety of techniques that have beendevised to measure attitudes.• These techniques range from direct to indirect,physiological to verbal• Obtaining verbal statements from respondentsgenerally requires that the respondent perform atask such as ranking, rating, sorting, or making achoice or a comparison.MAANZ International
  14. 14. 14The Attitude‐Measuring Process• Ranking tasks require that the respondent rank order a small number of objects inoverall preference on the basis of some characteristic or stimulus.• Rating asks the respondent to estimate the magnitude of a characteristic, or quality,that an object possesses. The respondent indicates the position on a scale(s) wherehe or she would rate an object.• Sorting might present the respondent with several product concepts typed on cardsand require that the respondent arrange the cards into a number of piles orotherwise classify the product concepts.• Choice between two or more alternatives is another type of attitudemeasurement-it is assumed that the chosen object is preferred over the other(s).• Physiological measures of attitudes provide a means of measuring attitudeswithout verbally questioning the respondent. For example, galvanic skin responses,measure blood pressure, etc., are physiological measures.MAANZ International
  15. 15. 15Behavioural Intention• The behavioural component of an attitude involves thebehavioural expectation of an individual toward anattitudinal object.• Category scales to measure the behavioural component ofan attitude ask a respondents "likelihood" of purchase orintention to perform some future action.• The wording of statements used in these cases oftenincludes phrases such as "I would recommend," "I wouldwrite," or "I would buy," to indicate action tendencies.MAANZ International
  16. 16. 16Projective Techniques• (1) Association Techniques• (2) Completion Techniques• (3) Construction Techniques• (4) Expressive TechniquesMAANZ International
  17. 17. 17SCALESA scale is any series of items which are progressively arrangedaccording to value or magnitude into which an item can be placedaccording to its quantification.MAANZ International
  18. 18. 18Scaling and Numbers• The term scaling is often used in the process ofmeasuring such things as attitudes and perceptions,where numbers are assigned to describe or representobjects, persons, households or events according to afixed set of rules.• Numbers may be numeric or simply symbolic;alternatively they can be classified under one of thefollowing four types; each possesses some inherentproperties which govern the type of mathematicaloperations that can be performed.MAANZ International
  19. 19. 19The differences between measurement and scaling• Measurement is the assignment of numbers or othersymbols to characteristics of objects according tocertain pre-specified rules.• Measurement precedes scaling in test construction.• Scaling is an extension of measurement where itinvolves the generation of a continuum upon whichmeasured objects are located.MAANZ International
  20. 20. 20The Main Scales of Measurement• The Nominal scale: • This is used as a labelling scheme where numbersserve only as labels or tags for identifying andclassifying objects.• The numbers in a nominal scale do not reflect theamount of a characteristic possessed by the objects,rather they are used only for identification.• For example, numbers on football players uniforms,street names, or licence plate numbers.MAANZ International
  21. 21. 21The Ordinal scale• This is a ranking scale in which numbers are assignedto objects to indicate the relative extent to whichsome characteristic is possessed.• It is then possible to determine whether an object hasmore or less of a characteristic than some otherobject.• For example, rankings of teams for the AFL or NRL,socioeconomic status and quality rankings.MAANZ International
  22. 22. 22The Interval scale• These numbers are used to rank objects such thatnumerically equal distances on the scale representequal distances in the characteristic being measured.• Examples include time and temperature.• The Ratio scale• This is used to identify or classify objects, rank orderthe objects, and compare intervals or differences.• For example, height, age, and income.MAANZ International
  23. 23. 23There are two main types of scales: comparative and non‐comparative.• Comparative scales - a direct comparison of stimulusobjects is elicited. Thus, two brands may be comparedalong a dimension such as quality.• Non-comparative scales - the respondent provideswhatever standard seems appropriate to him/her, thus,only one object is evaluated at a time.• This type of scaling does not compare the object againstanother object or some standard.• Rather, the rater uses whatever standard seems mostappropriate to him or her. In this case, one brand is ratedon a scale independent of other brands.MAANZ International
  24. 24. 24Comparative Scaling Techniques.• Paired comparison scaling• Here a respondent is presented with two objects at atime and asked to select one object in the pairaccording to some criterion.• The data obtained is ordinal in nature.• This is frequently used in marketing whencomparisons of products or brands are being made.MAANZ International
  25. 25. 25Rank order scaling• This is where respondents are presented with severalobjects simultaneously and asked to order or rankthem according to some criterion.• This is commonly used to measure preferences forbrands as well as the importance of attributes.• It is a simpler scale than the paired-comparison scaleas its procedure can easily be under-stood by therespondent.MAANZ International
  26. 26. 26Q‐sort scaling• This technique uses a rank order procedure in whichobjects are sorted into piles based on similarity withrespect to some criterion.• Magnitude estimation• Here numbers are assigned to objects such that ratiosbetween the assigned numbers reflect ratios amongthe objects on the specified criterion.MAANZ International
  27. 27. 27Non‐comparative Scaling Techniques• Continuous rating scale• In this measurement method the respondents rate the objects by placing amark at the appropriate position on a line that runs from one extreme of thecriterion variable to the other.• The form of the continuous scale varies considerably depending on theimagination of the researcher. Their use in marketing has been limitedbecause they are not as reliable as itemised scales, the scoring process iscumbersome, and they provide little additional information.MAANZ International
  28. 28. 28Itemised rating scales• Here respondents are provided with scales havingnumbers and/or brief descriptions associated witheach category.• The respondents are required to select one of thespecified categories that best describes the objectbeing rated.MAANZ International
  29. 29. 29Likert summated scale• Likert is based on summated ratings.• A list (or series) of attitude statements about theobject under study is compiled and the respondentindicates his or her degree of agreement ordisagreement with each of these statements that arerelated to the object in question on a five-point scale:• strongly agree• somewhat agree• neither agree nor disagree;• somewhat disagree• strongly disagree.MAANZ International
  30. 30. 30Semantic differential scale• This is a seven-point rating scale with end points associatedwith bipolar labels that have semantic meaning. As theterm suggests, this scaling technique measures thedifference between words. Respondents are required torate objects on a number of itemised, seven-point ratingscales bounded at each end by one of two bipolaradjectives.• The respondent places a cross X in the position whichindicates his or her thinking about a value offer in termsof the construct or dimension along a bi-polar adjective.MAANZ International
  31. 31. 31Bi-polar adjectives :• active/passive• unsuccessful/ successful• cruel/kind• important/ unimportant• curved/straight• angular/rounded• masculine/feminine• calm/excitable• untimely/timely• false/trueMAANZ International
  32. 32. 32Staple scale• This is a modified version of a semantic differentialscale, using a ten-point scale ranging from -5 to +5 tomeasure direction and intensity of attitudesimultaneously. It uses a single word adjective insteadof the polar pair of adjectives or descriptive phrases.Each point on a Stapel scale is assigned a number• An advantage of the Stapel scale is that there is noneed to worry about phrase bi-polarity - a +5 score istruly opposite to a -5 score and so on.MAANZ International
  33. 33. Other types of rating scales employed in marketing research include:
  34. 34. 34Pictorial rating scales•With regard to the piece of cake which youve just tasted, can youplease indicate which of the following facial expressions best describesyour reaction to its level of sweetness?•Another way is to present a set of word descriptions in varying printsizes for the respondent to express the intensity of his attitude or feeling:•In your opinion, do you think that abortion should be legally allowed?MAANZ International
  35. 35. 35Itemised (or verbal) rating scale• Itemised rating scale (or verbal rating scale) is onewhere the respondent is presented with a set ofresponse categories, ordered by the scale positions soas to reflect the degree of attitude held. Itemisedrating scale is most commonly used in marketingresearch.• Example• How likely would you buy this product if it is available in the market at $4.95 perbottle?• I definitely would buy 1• I probably would buy 2• I might or might not buy 3• I probably would not buy 4• I definitely would not buy 5MAANZ International
  36. 36. 36Itemised (or verbal) rating scale• When constructing an itemised rating scale, thefollowing major issues will need to be considered:• number of categories• odd or even number of categories• balanced or unbalanced scales• forced or unforced scales• comparative or non-comparative scales - which wehave previously described.MAANZ International
  37. 37. 37Number of categories • There is no fixed rule governing the exact number ofresponse categories required. The researcher can create asmany response categories as he deems appropriate,ranging from the simple dichotomous (e.g. Yes-Noanswer) to as many as 100 in number.• Odd or even number of categories• An even number of response categories refers when halfof the response categories are positive (favourable)statements while the remaining half are negative (orunfavourable) statements. An odd number of categories,on the other hand, includes an additional category whichis usually identified as the neutral position.MAANZ International
  38. 38. 38Balanced or unbalanced scale• In a balanced scale, the number of favourable responsecategories equals that of unfavourable responsecategories.•• Scales are unbalanced in that they may have a greater number of favourable scale points than unfavourable scale points (or vice versa). • Generally speaking, when prior knowledge throughexploratory research suggests a high likelihood offavourable (or unfavourable) responses to the attitudeunder study, it will be appropriate to adopt an unbalancedscale.MAANZ International
  39. 39. 39Forced scale and Unforced scales• The provision, or the lack of it, of dont know or noopinion category provides the distinction between aforced scale and an unforced scale.• If such a category is not provided, thus forcing therespondent to take one side or the other, we have aforced scale. Otherwise, we have an unforced scale.• A dont know answer should not be looked upon as aneutral response, since a respondent who is willing toexpress his feeling towards the research question mayfail to do so due to a genuine lack of knowledgeabout the attribute under study.MAANZ International
  40. 40. 40Verbal Protocols• A Verbal protocol is a method of understanding thethought processes of a consumer by having them speakevery thought which comes into their head, no matterwhat it may be. The objective is to understand thethought processes of the consumer.• Protocols are useful when a researcher wants to know theconsumers cognitive responses to marketing stimuli.Protocols have been used to determine the attributes andcues used in making purchase decisions, product usagebehaviour, and the impact of the shopping environmenton consumer decisions.MAANZ International
  41. 41. Criteria Used to Evaluate a Multi Item Scale
  42. 42. 42What Is To Be Measured?• The first question the researcher must ask is "What is tobe measured?". This is not as simple a question as it mayat first seem.• A precise definition of the concept may require adescription of how it will be measured, and there isfrequently more than one way of measuring a concept.Further, true measurement of concepts requires a processof precisely assigning scores or numbers to the attributesof people or objects.• To have precise measurement in marketing researchrequires a careful conceptual definition, an operationaldefinition, and a system of consistent rules for assigningnumbers or scales.MAANZ International
  43. 43. 43Concepts • Before the measurement process can occur, theresearcher has to identify and define the conceptsrelevant to the problem.• A concept (or construct) is a generalised idea about aclass of objects, attributes, occurrences, or processes.• Concepts such as brand loyalty, personality, and so onpresent great problems in terms of definition andmeasurement.MAANZ International
  44. 44. 44Operational Definitions• Concepts must be made operational in order to bemeasured. An operational definition gives meaningto a concept by specifying the activities or operationsnecessary to measure it.• It specifies what the investigator must do to measurethe concept under investigation.• An operational definition tells the investigator to "dosuch-and-such in so-and-so manner.".MAANZ International
  45. 45. 45Rules of Measurement• A rule is a guide instructing us what to do. Anexample of a measurement rule might be "assign thenumerals 1 through 7 to individuals according to howbrand loyal they are.• If the individual is an extremely brand loyalindividual, assign a 1. If the individual is a total brandswitcher with no brand loyalty, assign a 7."• Operational definitions help the researcher specify therules for assigning numbers.MAANZ International
  46. 46. 46Index Measures• Measuring more complex concepts may require more than one questionbecause the concept has several attributes. An attribute is a singlecharacteristic or fundamental feature pertaining to an object, person,situation, or issue.• Multi-itemed instruments for measuring a single concept with severalattributes are called index measures, or composite measures.• For example, index of social class may be based on three weightedaverages: residence, occupation, and residence. Asking different questionsin order to measure the same concept provides a more accurate cumulativemeasure than does a single-item measure.MAANZ International
  47. 47. 47Measurement accuracy• This refers to capturing the responses as therespondent intended them to be understood.• Errors can result from either systematic error, whichaffects the observed score in the same way on everymeasurement, or random error, which varies withevery measurement.MAANZ International
  48. 48. 48Reliability:  • Reliability refers to the extent to which a scale producesconsistent results if repeated measurements are made onthe characteristic.• Another dimension of reliability concerns thehomogeneity of the measure. An attempt to measure anattitude may require asking several questions or a batteryof scale items.• To measure the internal consistency of a multiple-itemmeasure, scores on subsets of items within the scale mustbe correlated. The split-half method, when a researcherchecks the results of one half of the scale items to theother half, is the most basic method for checking internalconsistency.MAANZ International
  49. 49. 49Validity:  • Validity refers to the accuracy of measurement.• Validity of a scale may be defined as the extent towhich differences in observed scale scores reflect truedifferences among objects on the characteristic beingmeasured, rather than systematic or random errors.MAANZ International
  50. 50. 50How is Validity Measured?• The purpose of measurement is to measure what weintend to measure. Validity addresses the problem ofwhether or not a measure does indeed measure what itpurports to measure; if it does not, there will beproblems.• Researchers attempt to provide some evidence of ameasures degree of validity. There are three basicapproaches to dealing with the issue of validity:MAANZ International
  51. 51. 51Face or content validity • This refers to the subjective agreement ofprofessionals that a scale logically appears to beaccurately reflecting what it purports to measure.• A subjective evaluation by experts in the domainbeing studied to determine if all relevant items arecovered in the study. Thus, in the auto industry forexample, car experts (experienced professionals,professors, etc.) would evaluate the scale on theitems it covers vis-à-vis what they consider relevantreal world factors.MAANZ International
  52. 52. 52Criterion validity • Criterion validity is an attempt by researchers to answerthe question "Does my measure correlate with othermeasures of the same construct?"• Consider the physical concept of length. If a newmeasure of length were developed, finding that the newmeasure correlated with other measures of length wouldprovide some assurance that the measure was valid.• Criterion validity may be classified as either concurrentvalidity (when the measure is taken at the same time asthe criterion measure) or predictive validity (when themeasure predicts a future event).MAANZ International
  53. 53. 53Construct validity • Construct validity is established by the degree to whichthe measure confirms a network of related hypothesesgenerated from a theory based on the concept.• In its simplest form, if the measure behaves the way it issupposed to in a pattern of inter-correlation with a varietyof other variables, then there is evidence for constructvalidity.• This is a complex method of establishing validity and isof less concern to the applied researcher than to the basicresearcher.MAANZ International
  54. 54. 54Construct Validity• Construct Validity entails a theoretical investigationof the construct the scale is measuring.• It tries to assess why the scale works and the natureof the theory underlying the scale.• Convergent, discriminant, and nomological validityare assessed.MAANZ International
  55. 55. 55Reliability versus Validity• The concepts of reliability and validity should becompared. Reliability, although necessary forvalidity, is not in itself sufficient.MAANZ International
  56. 56. 56Sensitivity• The sensitivity of a scale is important, particularly whenchanges in attitude, or other hypothetical constructs, areunder investigation.• Sensitivity refers to the ability of a instrument toaccurately measure variability in stimuli or responses.• The sensitivity of a scale which is based on a singlequestion or a single item can be increased by addingadditional questions or items.• In other words, because index measures allow for agreater range of possible scores, they are more sensitivethan single-item scales.MAANZ International
  57. 57. 57Ranking• Buyers often rank order their preferences. An ordinal scale may bedeveloped by asking respondents to rank order (from most preferredto least preferred) a set of objects or attributes. This technique is easilyunderstood by the respondents.• Paired comparisons: In paired comparisons the respondents arepresented with two objects at a time and asked to pick the one theyprefer.• Ranking objects with respect to one attribute is not difficult if only afew products are compared, but as the number of items increases, thenumber of comparisons increases geometrically (n*(n - 1)12). If thenumber of comparisons is too great, respondents may fatigue and nolonger carefully discriminate among them.MAANZ International
  58. 58. 58Sorting• Sorting tasks requires that respondents indicate their attitudes orbeliefs by ranking items.MAANZ International
  59. 59. 59Generalisability• This refers to the extent to which one can generalisefrom the observations at hand to the set of allconditions of measurement over which theinvestigator wishes to generalise, called the universeof generalisation.MAANZ International
  60. 60. 60Randomised Response Questions• In special cases, when respondents are asked to provide sensitive orembarrassing information in a survey, the researcher may utiliserandomised response questions. In this type of questionnaire, eachquestion has two possible questions associated with it-one sensitive andone non-sensitive.• The respondent will answer "yes" or "no" to the question asked. Thequestion asked is randomly selected by the respondent, whoconfidentially determines which of the two questions will be answered(e.g., by the toss of a coin).• A formula is used to estimate the proportion of "yes" answers to thesensitive question. While estimates are subject to error, the respondentremains anonymous, and response bias is therefore reduced.MAANZ International
  61. 61. 61ENDMAANZ International
  62. 62. • For more information about MAANZ International and articles about Marketing, visit:• www.marketing.org.au• http://smartamarketing.wordpress.com• http://smartamarketing2.wordpress.com• .  http://www.linkedin.com/groups/MAANZ‐SmartaMarketing‐Group‐2650856/about• Email: info@marketing.org.au• Link to this site ‐ ‐ http://www.slideshare.net/bmonger for further presentationsMAANZ International 62