Lesson 5: Basic Data Analysis: Descriptive Statistics & Qualitative Methodologies: Focus Groups & Observation 1 Types of Statistical Analyses Used in Marketing Research Five basic types of statistical analyses – Descriptive analysis – Inferential analysis – Differences analysis – Associative analysis – Predictive analysis Each one has a unique role in the data analysis process - usually combined into a complete analysis of the information in order to satisfy the research objectives 2 1) Descriptive Analysis The mean, median, mode, standard deviation and range are forms of descriptive analysis used to describe the sample data to portray the ‘typical’ respondent and to reveal the general pattern of responses 3 2) Inferential Analysis Using statistical procedures to generalize the results of the sample to the target population to make conclusions regarding the population Hypothesis testing based on sample population 4 3) Differences Analysis Sometimes, a marketing researcher needs to determine whether the two groups are different, for instance, the researcher may be investigating credit card usage and want to see if highincome earners differ from low-income earners in the frequency of usage of their American Express card 5 Differences Analysis… Cont’d Statistically, the researcher may compare the average annual dollar expenditures charged on American Express by high vs low income buyers - important market segmentation information may come from this analysis Includes the t test for significant differences among groups and analysis of variance 6 How satisfied are you with the services of Singtel? 1 – very satisfied , 5 – very unsatisfied Heavy Users Average Rating = 1.5 Light users Average Rating = 3.5 7 4) Associative Analysis Investigates if and how two variables are related – For example: are the scores for advertising recall, positively associated with the intentions to buy the advertised brand? – Are expenditures on customer service training positively associated with customer service excellence? E.g. spend more on customer service training = more customer service excellence? 8 Associative Analysis… cont’d The analysis may indicate the strength of the association and/or the direction of the association between two questions on a questionnaire in a given research study Analyses include correlation and crosstabulation E.g. spend more on customer service training = more customer service excellence? By how much? 9 5) Predictive Analysis Market Researcher can make use of statistical procedures and models to make forecasts about future events - useful as marketing managers are concern about what will happen in the future given certain conditions such as price increase or competitive pressures Regression or time series analyses Given the past sales trends and current conditions, what is the expected sales volume in the next 3 years? 10 Understanding Data via Descriptive Analysis With SPSS, these data could be input and assigned variable names such as ‘user’, ‘gender’, ‘payment mode’ etc We need to find some convenient means of summarizing the numbers in each column while communicating as much about the nature of this set of numbers as possible 11 Understanding Data via Descriptive Analysis… cont’d Two sets of measures are generally used to describe the information obtained in a sample 1st set - measures of ‘central tendency’ or measures that describe the ‘typical’ respondent or response 2nd set - measures of ‘variability’ or measures that describe how similar (or dissimilar) respondents or responses are to (from) ‘typical’ respondents or responses 12 Measures of Central Tendency Use to report a single piece of information that describes the most typical response to a question The term ‘central tendency’ applies to any statistical measure used that somehow reflects a typical or frequently response 3 measures of central tendency are commonly used – mode, median & mean 13 Descriptive Statistics: Key Terms Mean – The sum of the values for all observations of a variable divided by the number of observations. Median – The observation below which 50 percent of the observations fall. Mode – The value that occurs most frequently. 14 Measures of Variability Measures of central tendency do not indicate the variability of responses to a particular question or, the diversity of respondents on some characteristic measured in the survey Measures of variability is able to indicate the diversity or variability of values in a data set Knowing the mean or some other measure of central tendency is important, but to be aware of how close to that mean or measure of central tendency the rest of the values fall can be meaningful 15 Measures of Variability… cont’d For instance, in a survey of SMS users, Singtel might find that the average user send 5 SMS messages every day This measure of central tendency paints only part of the picture as some users may make more than 10 SMS messages daily, therefore knowledge of the variety of users can help the company target ‘heavy users’ with the right marketing strategies The 3 measures of variability: frequency distribution, range and standard deviation - each of them have their own unique version of information to describe the diversity of the responses 16 Frequency Distribution A frequency distribution is a tabulation of the number of times that each different values appears in a particular set of values Frequencies are raw counts and normally these frequencies are converted into percentages for ease of comparison Example: 40 respondents – 30 own a car and 10 do not Drive a Car – 75% Do not own a car – 25% 17 Range The range identifies the distance between the lowest value (minimum) and the highest value (maximum) in an ordered set of values The range specifies the difference between the endpoints in a distribution of values arranged in order The range does not tell you how often the maximum and minimum occurred but it does provide some information on the dispersion by indicating how far apart the extremes are found Example: in a data set – 1,1,2,3,4,5,6,6,7,8,9,10, 11 The range is between 1 to 11 18 Standard Deviation Indicates the degree of variation or diversity in the values that are translatable into a normal or bell-shaped curve distribution On a normal curve, the mid-point is also the mean Formula s = {sum (xi – x)2}/ n-1 where xi = individual observation x = mean n = no. of observations 19 Assuming if the responses approximate a bell shaped distribution, the size of the standard deviation helps you visualize how similar or dissimilar the typical responses are to the mean E.g. Distance travelled by cars in Singapore annually E.g. mean = 500 km Standard deviation = 100 km We can say that 95% of the responses will fall between 500 - 1.96 (100) and 500 + 1.96 (100) = 304 km to 696 km 500 + 1.96 (100) 20 Focus Groups & Observations 21 What are Focus Groups? …a small group of people brought together and guided by a moderator through an unstructured, spontaneous discussion about some topic. Some Objectives to generate ideas to understand consumer vocabulary to reveal consumer needs, motives, perceptions and attitudes on products or services to understand findings from quantitative studies 22 Some Objectives of Focus Groups To generate ideas: – To use the focus group as a starting point for new product or service ideas, uses or improvements To understand consumer vocabulary: – To use the focus group to stay abreast of the words and phrases consumers use when describing one’s product so as to improve product or service communication with them 23 Some Objectives of Focus Groups To reveal consumer needs, motives, perceptions and attitudes on products or services – To use the focus group to refresh the marketing team as to what customers really feel or think about a product or service To understanding findings from quantitative studies – To use the focus groups to better comprehend data gathered from other surveys 24 Operational Questions About Focus Groups What Size (no. of people)? Who Should Take Part? How to Recruit and Select Participants? Where Should It Take Place? What are the Moderator’s Roles and Responsibility? How to Report and Use the Findings? 25 Operational Questions About Focus Groups What should be the size of the focus group? – The optimal size of a focus group is 8 to 12 people – A small group with fewer than 8 participants is not likely to generate the necessary energy and group dynamics – As the group becomes larger in size it tends to become fragmented and there may be inherent digressions and side comments 26 Operational Questions About Focus Groups Who should be in the focus group? – The best focus groups are ones in which the participants share similar characteristics – The requirement is sometimes automatically satisfied by the researcher’s need to have particular types of people in the focus group discussion 27 Operational Questions About Focus Groups How should focus group participants be recruited and selected? – If the purpose of the focus group is to generate new ideas on product packaging, the participants must be consumers who have used the brand – Companies may provide customer lists for the focus group recruiters to work with 28 Operational Questions About Focus Groups How should focus group participants be recruited and selected? – Strategies to entice prospective participants includes: Various types of incentives ranging from monetary compensation to free products or gift certificates Using call-backs the day prior to the focus group meeting to remind prospective participants that they have agreed to take part Over-recruiting is one way to deal with no shows 29 Operational Questions About Focus Groups Where should a focus group meet? – If a group discussion is to take place for a period of 90 minutes or more, it is important that the physical arrangement of the group be comfortable and conducive to group discussion – Ideally, focus group should be conducted in large rooms setup in a roundtable format 30 Operational Questions About Focus Groups The focus group moderator’s role & responsibilities: – The moderator is a person who conducts the entire session and guides the flow of group discussion across specific topics desired by the client – A successful focus group requires an effective moderator – A good moderator is experienced, enthusiastic, prepared, involved, energetic and open-minded 31 Operational Questions About Focus Groups The focus group moderator’s role & responsibilities: – Unless the moderator understands what information the researcher is after and why, he or she will not be able to phrase questions effectively – Also, the moderator’s introductory remarks are influential and they set the tone of the entire session 32 Operational Questions About Focus Groups Reporting and use of focus group results: – Factors to consider when analyzing data: The translation of qualitative statements of participants into categories and then reporting the degree of consensus apparent in the focus groups Demographic and buyer behaviour characteristics of focus group participants should be judged against the target market profile to assess to what degree the groups represent the target market 33 Operational Questions About Focus Groups Reporting and use of focus group results: – Focus group transcripts must be translated before they are reported – A focus group analysis should identify the major themes as well as salient areas of disagreement among the participants 34 Advantages & Disadvantages of Focus Groups Advantages Generate Fresh Ideas Allow Clients to Observe Generally Versatile Work Well with Special Participants Disadvantages Representativeness Subjective Interpretation High Cost per Participant 35 Advantages of Focus Groups Generate fresh ideas: Creative & honest insights are often the result of focus groups The interaction among respondents can stimulate new ideas and thoughts that might not arise during one-on-one interviews 36 Advantages of Focus Groups Allows clients to observe the group the ability to observe focus groups greatly facilitates client understanding an opportunity could be made available for them to observe their customers or prospects from behind one-way mirrors 37 Advantages of Focus Groups Generally versatile virtually unlimited number of topics and issues can be discussed in a focus group interview situation focus groups using video-conferencing and Internet ‘chat rooms’ demonstrate the versatility of the technique 38 Advantages of Focus Groups Work well with special respondents Focus groups permit the researcher to study respondents that might not respond well under more structured situations Sometimes involving ‘hard-to-interview’ groups such as doctors and accountants; the format gives them an opportunity to associate with their peers and compare notes 39 Disadvantages of Focus Groups May not represent the population May not be representative of the total group under study Interpretation is subjective Individuals with pre-conceived notions can almost always find something to support their views, ignoring anything that does not support their opinions It is always advisable to use a trained analyst to interpret the findings from the focus groups Peer Pressure 40 Disadvantages of Focus Groups High cost per participant A number of expensive items contribute to the high cost per participant Participant recruitment and the numerous telephone calls are usually needed to recruit the desired number of required participants Incentive cost for all participants Moderator’s fees Facilities rental fees + videotaping equipments charges 41 Brief Steps in Conducting a Focus Group Prepare for the Group: Select a focus group facility and recruit the participants Select a Moderator: Create a Discussion guide Prepare the Focus Group Report Conduct the Focus Group 42 Observation Research Defined Instead of asking people questions, as in a survey, observation depends on watching what people do Observation methods often rely on recording devices as the researcher’s memory alone can be faulty McDaniel defines Observation Research as ‘the systematic process of recording the behavioural patterns of people, objects and occurrences without questioning or communicating with them’ 43 Types of Observation Observation studies may appear to be without any structure or plan, but it is important to stick to a formal structure and plan so that the observations are consistent and comparisons or generalizations can be made as accurate as possible 44 Types of Observation 4 general ways of organizing observations: Direct versus Indirect Disguised versus Undisguised Structured versus Unstructured Human versus Mechanical 45 Direct versus Indirect Observation Observing behaviour as it occurs is called direct observation – – Some examples: finding out how much shoppers squeeze tomatoes to assess their freshness through the observation of people picking up the tomatoes Kellogg’s used direct observation to understand people’s behaviour during breakfast 46 Direct versus Indirect Observation Observing different types of past behaviour (or hidden behaviour) we must rely on indirect observation With indirect observation, the researcher observes the effects or results of the behaviour rather than the behaviour itself – Example: In a product prototype test, it is important to know how much of the test product was actually used – the most accurate way to find out is to have respondents returned the unused product so that the researcher can measure how much is left 47 Types of Indirect Observation Archives – secondary sources such as historical records that can be applied to the present problem – – Examples: Records of sales calls may be inspected to determine how often sales people make cold calls Warehouse inventory movements can be used to study market shifts 48 Disguised versus Undisguised Observation With disguised observation the subject is unaware that he or she is being observed – – Examples: ‘Secret shopper’ that a retail store chain used to record and report on sales clerks’ assistance and courtesy One-way mirrors and hidden cameras are a few of the other ways that are used to prevent subjects from becoming aware that they are being observed 49 Disguised versus Undisguised Observation Sometimes, it is impossible for the respondent to be unaware of the observation and here is the case of undisguised observation – – Some examples: laboratory settings, observing a sales rep’s behaviour on sales call must all be used/done with the subject’s knowledge As people might be influenced by knowing that they are being observed, it is advisable to always minimize the presence of the observer to the maximum extent possible 50 Structured versus Unstructured Observation Observation may be structured or unstructured When using structured observation techniques, the researcher identifies beforehand which behaviours are to be observed and recorded, all other behaviours are ‘ignored’ Often, a checklist or a standardized observation form is used to isolate the observer’s attention to specific factors 51 Structured versus Unstructured Observation Unstructured observation places no restriction on what the observer would note: all behaviour under study is monitored and the observer just watches the situation and records what he or she deems interesting or relevant The observer is thoroughly briefed on the area of general concern (what to look out for during the observation process?) 52 Human versus Mechanical Observation With human observation, the observer is a person hired by the researcher or the observer is the researcher and When we replace the human observer with some form of observing device, we call it mechanical observation The substitution may be made because of accuracy, cost or functional reasons 53 Human versus Mechanical Observation Example: during rush hours, a human observer could not count the number of cars on most major cosmopolitan commuter roads Mechanical observation has moved into the hightechnology area recently and the combination of telecommunications, computer hardware and software programmes has created a very useful research tool 54 Advantages of Observational Data Ideally the subjects of observational research are unaware they are being studied as they should react in a natural manner, giving the researcher insight into actual, not reported behaviours Observers see what people actually do rather than having to rely on what they say they did Observation may be the only way to obtain accurate information 55 Limitations of Observational Data The limitations of observation are the limitations present in qualitative research in general With direct observation, only small numbers of subjects are studied and usually under special circumstances, so their representativeness is a concern Secondly, the subjective interpretation required to explain the observed behaviour usually forces the researcher to consider his or her conclusions to be tentative 56 Limitations of Observational Data The greatest drawback is the researcher’s inability to pry beneath the behaviour observed and to interrogate the person on motives, attitudes and all of the other unseen aspects of why what was observed took place Motivations, attitudes and other ‘internal’ conditions cannot be observed 57 Other Qualitative Research Techniques 1. Protocol Analysis 2. Projective Techniques Word Association Test Sentence Completion Test Picture Test Cartoon or Balloon Test Role-Playing Activity 58