Seminar für Allgemeine BWL, Marketing und Marktforschung Prof. Dr. Marc Fischer

Worksheet No. 1 – Solution Outline (Design of Market Research Projects, sampling fundamentals and sampling procedures)
1.) Since Blu-ray players have only recently been introduced in the German market, only few experiences on the purchase and usage behavior of consumers exist. The company SimSing wants to determine whether it is worthwhile for the company to enter the market for Blu-ray players. Therefore, it is interested in the customer needs and expectations with respect to Blu-ray players. As a marketing expert, you have been assigned to design an appropriate market research project. a) Identify and briefly explain the stages of a market research project. Illustrate the different stages with regard to the determination of customer needs and expectations for Blu-ray players.  1st Step: Problem definition (Manager has a management problem that has to be transformed into a market research problem; formulation of market research project aims):      Formulation of project aim (e.g., determining customer demands with regard to Blu-ray players in order to open up new markets) Identification of responsible managers/departments (e.g., strategic marketing, R&D, engineering, management accounting) Definition of possible deliverables (e.g., catalog of requirements for Bluray players and weighting of requirements) Narrowing down the scope of analysis (e.g., narrowing down requirements like potential playback formats, price, user friendliness) Clarification of cost responsibility (e.g., strategic marketing, cost controlled by management accounting)

 2nd Step: Project design (Design of a detailed data collection and data analysis plan):   Selection of data collection type (primary versus secondary research; e.g., survey, interview, experiment, conjoint measurement study) Selection of data analysis method (e.g., univariate/descriptive: determining the rank order of preferences; correlations: determining potential relations between preferences) Determination of the composition and scope of the sample (e.g., N = 100 customers, who have once bought a DVD-player) Design of data collection instrument (e.g., design of questionnaire, number of questions, scaling; pretesting of the questionnaire)

 

© Dr. Alexander Himme

1

Seminar für Allgemeine BWL, Marketing und Marktforschung Prof. Dr. Marc Fischer
 Consult with stakeholders in the organization on the project concept (e.g., checking if the design meets the expectations of the stakeholders and the projects goals; checking if the design is in line with the cost budget)

 3rd Step: Execution and analysis (Planning, organizing, controlling of data collection):   Possible evaluation and selection of external service providers (e.g., GfK) Anticipation of potential obstacles and how to overcome them (e.g., dissatisfying response rate → change of questionnaire design or provide better incentives for the participation in the survey) Definition of the deliverables of the data analysis (e.g., aggregation level of the results) Determination and control of milestones (e.g., data collection = 3 months, data analysis = 3 months) Possibly integrate the external service provider (e.g., benefit from its experience with previous market research projects; benefit from its data analysis competencies)

  

 4th Step: Communication of results (Formulation of research report and presentation of results; the report and the presentation have to be adapted to the various interests of different stakeholders):     Identify the result recipients (see 1st step, engineers, strategic marketing, different marketing departments within the firm, etc.) Define the manner in which the results will be conveyed (e.g., presentation and written report) Define the contents of the communication (e.g., presentation of the crucial results) Monitor the use of the results (e.g., keep track of how engineers implement the results)

b) Explain the difference between primary and secondary data. Name potential primary and secondary data sources for this present Blu-ray player market research project. Primary data = Data does not yet exist and must be collected by the researcher or third parties; e.g., qualitative survey with industry experts; quantitative customer survey Secondary data = Data already exists within the company or is collected by third parties at specific intervals and in standardized form; e.g., publications by research institutes or media companies; comparable survey with regard to the use of DVD-players c) Discuss the advantages of primary and secondary data. Which data collection procedure would you recommend in this case? © Dr. Alexander Himme 2

Seminar für Allgemeine BWL, Marketing und Marktforschung Prof. Dr. Marc Fischer
Advantages of primary data: Optimal adaptation of the data to the project goals; greater control over the data quality; exclusiveness of the data Disadvantages of primary data: Higher costs (money and time); uncertainty with regard to the response rate Advantages of secondary data: Lower costs (money and time), especially since the advent of the Internet; can be the background/indication for primary research Disadvantages of secondary data: Fit with the research goals may be doubtful; no control over the data quality; data can be outdated Recommendation: Blu-ray players are still a rather new product; this would imply that a primary research project is necessary (exploratory market research project); anyway, it should always be checked if secondary data (e.g., comparable surveys) already exists! 2.) Boss has hired you to determine the color preferences of customers for the next summer fashion. Describe potential goals pursued by this market research project. Depending on the goal, which market research project design would you recommend?  Possible goals of the market research project: input for marketing activities (advertising, promotion, etc.); prediction of sales/input for the production process (quantity of materials to be provided, material planning)  Marketing-oriented goal: e.g., experiment (determine, how different and new fashion collections appeal to the customers)  Production-oriented goal: e.g., quantitative survey to determine purchase probabilities to predict sales  To sum up, the goal of the market research project always determines its set-up (sample, data collection method, data analysis method, stakeholders, etc.) 3.) The CEOs of the S&P 500 have been asked for their assessment regarding the future performance prospects of their firm in the next business year. The results of the survey were as follows (scale: 5 = very positive, …, 1 = very negative):
How do you assess the prospects of your firm performance for the next business year? 160 120 100 80 40

180 160 140 120 100 80 60 40 20 0

Frequency

Very positive

Positive

Average

Negative

Very negative

© Dr. Alexander Himme

3

Seminar für Allgemeine BWL, Marketing und Marktforschung Prof. Dr. Marc Fischer

a)

Determine the population mean. Attribute level ui 5 4 3 2 1 Absolute frequency ni(ui) 100 160 120 40 80 Σ = 500 Relative frequency fi(ui) 0.20 0.32 0.24 0.08 0.16 Σ = 1.00

µ=

5 1 5 1 ∑ niui = ∑ fiui = 500 (100⋅5 +160⋅ 4 +...+ 80⋅1) = 0.20⋅5 + 0.32⋅ 4 +...+ 0.16⋅1= 3.32 N i=1 i=1

b)

Determine the variance and the standard deviation of the population. ui-µ 1.68 0.68 -0.32 -1.32 -2.32
5

(ui-µ)2 2.82 0.46 0.10 1.74 5.38

(ui-µ)2·fi(ui) 0.56 0.15 0.02 0.14 0.86

σ 2 = ∑ f i (ui − µ ) 2 = 1.73 ⇒ σ = 1.32 i=1 c)

A simple random sample (N = 25) with the following distribution of response frequencies was taken from the population:
How do you assess the prospects of your firm performance for the next business year?

10 8

9 5 4 4 3

Frequency

6 4 2 0

Very positive

Positive

Average

Negative

Very Negative

© Dr. Alexander Himme

4

Seminar für Allgemeine BWL, Marketing und Marktforschung Prof. Dr. Marc Fischer

Determine the sample mean, the sample variance, and the standard deviation of the sample.

x=

1 N 1 ∑ x i = 25 (4 ⋅ 5 + 9 ⋅ 4 + 4 ⋅ 3 + 3⋅ 2 + 5⋅1) = 3.16 N i=1

s2 =

1 N 1 3.16)2 + (5 - 3.16)2 + ∑ (x i - x)2 = 24 ((5 -(5 - 3.16)2 + (5 - 3.16)2   N -1 i=1
=4(5−3.16)2

+ ... + 5(1- 3.16)2 ) = 1.97 ⇒ s = 1.40
Note: Different formula for variance calculation compared to previous question, because here the sample variance is calculated, while before, the population variance was asked for. d) Determine the standard error of the sample mean ( σ X ). How does this standard error change, if N = 100? Why? (Advice: Use s instead of σ in the calculation of σ X ).

With s = 1.40: σ x = With σ = 1.32: σ x =

s N σ

=

1.40

4.)

= = 0.264; for N = 100: σ x = 0.132 N 25 The students at the University of Cologne have been asked whether they would agree to a general start of the first lectures at 7.30 a.m. in the morning (instead of 8.00 a.m.) (Scale: +2 = „definitely yes“; +1 = „prefer yes“; 0 = „I don’t care“; -1 = „prefer no“; 2 = „definitely no“). The population mean (µ) is 0.3 (σ = 1.49). A sample (N = 100) with a mean of 0.52 and a sample variance of 2.62 was drawn.
a) Determine a 90%-confidence interval for µ.

25 1.32

= 0.28; for N = 100: σ x = 0.140 1

5 (90 % - confidence interval); N = 100 : 3 ⎡ 5 1.49 5 1.49 ⎤ ⇒ ⎢0.52 - ⋅ ;0.52 + ⋅ ⎥ = ⎡0.272;0.768⎤ ⎦ 3 100 3 100 ⎦ ⎣ ⎣ x = 0.52; σ = 1.49; z =
b)
1

Determine a 95%-confidence interval for µ.

Please note: If we have a different N, s will also be different. For this simple case, we just assume that for N = 100, s is still 1.40. Anyway, the case of N = 100 shall illustrate that increasing N leads to a smaller standard error of the sample mean no matter if we have σ or s.

© Dr. Alexander Himme

5

Seminar für Allgemeine BWL, Marketing und Marktforschung Prof. Dr. Marc Fischer

x = 0.52; σ = 1.49; z = 2 (95 % - confidence interval); N = 100 : ⎡ 1.49 1.49 ⎤ ⇒ ⎢0.52 - 2 ⋅ ;0.52 + 2 ⋅ ⎥ = ⎡0.222;0.818 ⎤ ⎣ ⎦ 100 100 ⎦ ⎣
c) Repeat a) and b) by using s 2 instead of σ 2 . Explain the difference. 90 %-confidence interval:

5 ; N = 100 : 3 ⎡ 5 1.62 5 1.62 ⎤ ⇒ ⎢0.52 - ⋅ ;0.52 + ⋅ ⎥ = ⎡0.250;0.790 ⎤ ⎦ 3 100 3 100 ⎦ ⎣ ⎣ x = 0.52; s = 2.62 = 1.62; z =
95 %-confidence interval:

x = 0.52; s = 2.62 = 1.62; z = 2 ; N = 100 : ⎡ 1.62 1.62 ⎤ ⇒ ⎢0.52 - 2 ⋅ ;0.52 + 2 ⋅ ⎥ = ⎡0.196;0.844 ⎤ ⎣ ⎦ 100 100 ⎦ ⎣
Difference is due to higher sample variance (standard deviation) compared to population variance (standard deviation). d) What sample size is needed for reducing the sample error to 0.1 at a confidence level of 95%? Sample error 0.1, confidence level (95 %, i.e. z = 2):

Δx = z

σ
N

⇔N=

z 2σ 2 4(1.49)2 = = 888.04 (Δx )2 (0.1)2

As an illustration: confidence level of 95 %, but sample error of 0.05:

Δx = z

σ
N

⇔N=

z 2σ 2 4(1.49)2 = = 3552.16 (Δx )2 (0.05)2

Doubling of precision (= bisection of sample error) ≈ Quadruplication of sample size!

© Dr. Alexander Himme

6

Seminar für Allgemeine BWL, Marketing und Marktforschung Prof. Dr. Marc Fischer

5.)

The State Library of Cologne worries about the citizens of Cologne using the library relatively seldom compared to the average library usage in Germany. Therefore, the library wants to work out how they can make their offer more attractive to the citizens of Cologne. Therefore, the State Library of Cologne plans to interview a sample of citizens who own a library card. How do you assess this plan?  Convenience Sample  However, the State Library has chosen a sampling procedure which is not reasonable for the objective of the research project; in order to determine the reason for the low utilization the library should interview users without library cards and especially NONUSERS.

6.)

After the successful implementation of the market research project, SimSing has launched a new Blu-ray player. The company is interested in how the purchasing managers of electronics stores (e.g., Saturn) will assess the new Blu-ray player. The company knows that the 1,000 existing stores can be divided into 100 wholesalers and 900 smaller merchants. A random sample (N = 200) was drawn, which includes 19 wholesalers. Yet the manager of the marketing department is dissatisfied with this result, as the wholesalers make up 30% of sales volume. An employee proposes to draw a second sample. What would you recommend in this case? Which information regarding the sampling procedure would you like to have in addition?  The heterogeneity in the population asks for a stratified sample.  Disproportional sampling since the information from wholesalers is more valuable regarding their sales volume share. E.g., if there has been approved a budget for a sample of N = 200, a share of 120 smaller merchants and 80 wholesalers should be aimed for (at least 60 wholesalers should be found in the sample, so that the wholesalers make up 30 % (equivalent to sales volume share) of the total sample size).  Further relevant information: costs for the sampling procedure and the data collection methods; variance in each of the specific groups.

© Dr. Alexander Himme

7