Sales Chapter 15 Homework Front Line Solid Set Cross Tabulations Critical

Unlock access to all the studying documents.

View Full Document

McDaniel & Gates – Marketing Research, 9th Edition Instructor’s Manual

Instructor’s Manual

CHAPTER 15

Data Processing and Fundamental Data Analysis

LEARNING OBJECTIVES

1. To get an overview of the data analysis procedures.

3. To understand the data entry process and data entry alternatives.

5. To learn how to set up and interpret cross tabulations.

6. To comprehend basic techniques of statistical analysis.

KEY TERMS

Validation

Editing

Skip pattern

Marginal report

One-way frequency table

Cross tabulation

CHAPTER SCAN

This chapter examines the data processing procedure and fundamental data analysis. Data

processing begins with the validation of the interviewing process. The researcher must ensure

that the data was collected and recorded accurately and correctly. Once that has been

Instructor’s Manual

research because of technological advances. Once data has been entered, it should be submitted

to machine cleaning. Once the data has been correctly entered, it can be tabulated. This can be

done in several ways. One is a sample one-way frequency table that shows the number of

CHAPTER OUTLINE

1. Overview of the Data Analysis Procedure

2. Step One: Validation and Editing

I. Validation

3. Step Two: Coding

I. Coding

4. Step Three: Data Entry

I. Data Entry

5. Step Four: Logical Cleaning of Data

6. Step Five: Tabulation and Statistical Analysis

7. Graphic Representations of Data

I. Graphic Representations of Data Defined

8. Descriptive Statistics

I Descriptive Statistics

9. Summary

CHAPTER SUMMARY

1. OVERVIEW OF THE DATA ANALYSIS PROCEDURE

Once data collection is completed and questionnaires returned, here’s a 5-step procedure for data

analysis:

2. STEP ONE: VALIDATION AND EDITING

I. Validation

A. Validation Defined–the process of ascertaining that interviews were conducted as specified–

goal of validation is solely to detect interviewer fraud or failure to follow key instructions

1. Telephone Validation–covers four areas

a. Was the person actually interviewed?

b. Did the person who was interviewed actually qualify to be interviewed?

c. Was the interview conducted in the required manner?

d. Did the interviewer cover the entire survey?

2. Checking for Other Kinds of Problems

a. Was the interviewer courteous?

b. Did the interviewer speculate about the client’s identity or the purpose of the survey?

c. Was the interviewer neat in appearance?

3. Purpose of the Validation–process is to ensure that interviews were administered properly

and completely

II. Editing

A. Editing Defined–involves checking for interviewer and respondent mistakes. The process of

ascertaining that questionnaires were filled out properly and completely.

1. Editing Process–involves manual checking for a number of problems including the following:

a. Whether the interviewer failed to ask certain questions or record answers for certain questions.

See Exhibit 15.1 Sample Questionnaire (p 438 – 440)

b. Questionnaires are checked to make sure that skip patterns were followed.

See Exhibit 15.1 Sample Questionnaire (p 438 – 440)

c. Whether the interviewer paraphrased respondent’s answers to open-ended questions are

checked.

See Exhibit 15.2 Recording of Open-Ended Questions (p 441)

3. STEP TWO: CODING

I. Coding

A. Coding Defined–refers to the process of grouping and assigning numeric codes to the

responses to a question

B. Coding Process

1. Four steps in the coding of responses

See Exhibit 15.3 Sample of Responses to Open-Ended Question (p 442)

c. Set codes.

See Exhibit 15.4 Consolidated Response Categories and Codes of Open-Ended Responses

from Beer Study (p 443)

d. Enter codes

1) Read responses to individual open-ended questions on questionnaires

3) Write numeric code on the questionnaire for the response to the particular question

See Exhibit 15.5 Example Questionnaire Setup For Open-Ended Questions (p 443)

C. Automated Coding Systems

1. CATI and Internet Surveys–data entry and coding are completely eliminated for closed-

ended questions

a. Open-ended Questions–text is captured electronically but the coding process is still required

2. TextSmart Module of SPSS–using algorithms based on semiotics will do automated coding

4. STEP THREE: DATA ENTRY

I. Data Entry

Instructor’s Manual

b. System can be programmed to avoid certain types of errors at the point of data entry

C. The Data Entry Process

1. Process of going directly from the questionnaire to the data entry device and the associated

storage medium has proven to be more accurate and efficient than transferring data from

questionnaires to coding sheets.

D. Scanning

1. Technology of this Type–has been around for years. It has not been economically feasible to

use in marketing research until recently. Now a computer can generate the forms needed and

2. Electronically Captured Data is Increasing

a. Computer-assisted telephone interviewing

5. STEP FOUR: LOGICAL CLEANING OF DATA

I. Logical (Machine) Cleaning of Data Defined

A. Final Computerized Error Check of the Data

1. Error Checking Routines–computer programs that accept instructions from the user to check

for logical errors in the data

2. Marginal Report–a computer-generated table of the frequencies of the responses to each

question, used to monitor entry of valid codes and correct use of skip patterns

3. Final Error Check in the Process–when this step is completed, the computer data file should

be ready for tabulation and statistical analysis

6. STEP FIVE: TABULATION AND STATISTICAL ANALYSIS

I. Tabulation and Statistical Analysis Defined

1. One-way frequency table is the first summary of survey results. These tables typically indicate

the percentage of those responding who gave each possible response to a question

2. An Issue that must be dealt with when one-frequency tables are generated is what base to use

for the percentages for each table. There are three options for a base:

a. Total respondents

b. Number of people asked the particular question

c. Number answering the question

See Exhibit 15.7 One-Way Frequency Table Using Three Different Bases for Calculating

Percentages (p 447)

3. The question is whether percentages in frequency tables showing the results for these question

be based on the number of respondents or the number of responses

See Exhibit 15.8 Percentages for a Multiple-Response Question Calculated on the Bases of

Total Respondents and Total Responses (p 448)

B. Cross Tabulations–examination of the responses to one question relative to the responses to

one or more other questions

See Exhibit 15.9 Sample Cross Tabulation (p 449). This cross tabulation includes frequencies

and percentage, with the percentages based on column totals–it shows an interesting relationship

between age and likelihood of choosing Minneapolis or St. Paul for hospitalization.

1. Following are a number of considerations regarding the setup of cross tabulation tables and the

2) row percentage–bases on the row total

3) total percentages–based on the table total

See Exhibit 15.10 Cross-Tabulation Table with Column, Row, and Total Percentages (p

449)

2. A common way of setting up cross-tabulation tables is to use columns to represent factors

3. A complex cross-tabulation, generated using the UNCLE software package is referred to as a

stub and banner table

See Exhibit 15.13 A Stub and Banner Table (p 500)

See Practicing Marketing Research: Six Practical Tips for Easier Cross Tabulations (p 451)

Here are six practical tips to improve your cross tabulation gleaning from Custom Insight, a

provider of Web-based survey software located in Carson City, Nevada

(www.custominsight.com).

2. Look for What is Not There

4. Keep Your Mind Open

5. Trust the Data

6 Watch the “n”–small totals should raise suspicions

See From the Front Line: Solid Set of Cross- Tabulations Critical (p 452-453)

1. Were all the responses to rating scale questions entered to ascending order (lowest rating is

lowest number, highest rating is highest number) with the exception of two or three questions?

2. Did an unusual number of respondents fail to answer only one particular open-end question?

7. GRAPHIC REPRESENTATIONS OF DATA

I. Graphic Representations of Data Defined

A. Line Charts

1. Simplest form of graphs

2. Useful for presenting a given measurement taken at several points over time

1. Appropriate for displaying percentages or proportions of a whole

See Exhibit 15.13 Three-Dimensional Pie Chart for Types of Music Listened to Most Often

(p 454)

C. Bar Charts

1. Plain Bar Chart–simplest form of bar chart

See Exhibit 15.14 Simple Two-Dimensional Bar Chart for Types of Music Listened to Most

Often (p 455)

See Exhibit 15.15 Simple Three-Dimensional Bar Chart for Types of Music Listened to

Most Often (p 456)

2. Clustered Bar Chart–one of three types of bar charts useful for showing the results of cross

(p 456)

3. Stacked Bar Chart–helpful in graphically representing cross tabulation results

See Exhibit 15.17 Stacked Bar Chart for Types of Music Listened to Most Often by Age (p

457)

4. Multiple-Row, Three-Dimensional Bar Chart–most visually appealing way of presenting

cross-tabulation information

See Exhibit 15.18 Multiple-Row, Three-Dimensional Bar Chart for Types of Music

Listened to Most Often by Age (p 457)

See Practicing Marketing Research: Expert Tips on Making Bad Graphics Every Time (p

455)

1. Occultation: hide everything important and avoid showing the relevant data.

• Reduce to a minimum the data density (put miniscule data on a large chart).

• Minimize the data/ink ratio (use a little ink for the data but lots for the axis, reference grid, and

Instructor’s Manual

• Ignore the codification (make the lengths or data areas non-proportional to its values).

Discussion Questions

1. Come up with two more ways to hide data and two more ways to be inconsistent in presenting

data graphics.

8. DESCRIPTIVE STATISTICS

I. Descriptive Statistics

A. Descriptive Statistics Defined–most efficient means of summarizing characteristics of large

sets of data

B. Measures of Central Tendency

1. Four Basic Types of Measurement Scales

a. Nominal and ordinal–referred to as nonmetric scales

b. Interval and ratio scales–metric scales

2. Three measures of central tendency

a. Mean–properly computed only from interval or ratio (metric) data–sum of the values for all

observations of a variable divided by the number of observations (see p 506 for formula for

1. Measures of Dispersion Defined – measure how “spread out” the data are. They include the

standard deviation, variance, and range.

2. Measures of Central Tendency Defined – indicate typical values for particular variable,

measures of dispersion indicate how spread out the data is.

a. The dangers associated with relying only on measures of central tendency are suggested by the

example in Exhibit 15.20 (p 459)

See Exhibit 15.20 Measures of Dispersion and Measures of Central Tendency (p 459)

D. Percentages and Statistical Tests

1. The research analyst makes the decision of whether to use measures of central tendency

(mean, median, mode) or percentages (one-way frequency tables, cross-tabulations)

2. Responses to questions either are categorical or take the form of continuous variables

a. Categorical variables–categories such as “Occupation” (coded 1 for professional/managerial,

2) If categories–one-way frequency tables and cross tabulations are used for analysis

9. SUMMARY

QUESTIONS FOR REVIEW AND CRITICAL THINKING

1. What is the difference between measurement validity and interview validation?

2. Assume that Sally Smith, an interviewer, completed 50 questionnaires. Ten of the

questionnaires were validated by calling the respondents and asking them one opinion

question and two demographic questions over again. One respondent claimed that his age

category was 30-40, when the age category marked on the questionnaire was 20-30. On

another questionnaire, in response to the question, “What is the most important problem

facing our city government?” the interviewer had written, “The city council is too eager to

raise taxes.” When the interview was validated, the respondent said, “The city tax rate was

too high.” As a valuator would you assume that these were honest mistakes and accept the

entire lot of 50 interviews as valid? If not, what would you do?

The error on the first questionnaire may have been a judgment call on the part of the interviewer

or it may have simply been a missed stroke of the pencil or computer. The second mistake,

3. What is meant by the editing process? Should editors be allowed to fill in what they

think a respondent meant in open-ended questions if the information seems incomplete?

Why or why not?

4. Give an example of a skip pattern on a questionnaire. Why is it important to always

follow the skip patterns correctly?

1. Have you eaten at a fast food restaurant in the last two weeks? Yes No

5. It has been said that, to some degree, coding of open-ended questions is an art. Would

you agree or disagree? Why? Suppose that after coding a large number of questionnaires,

the researcher notices that many responses end up in the “Other” category, what might

this imply? What could be done to correct it?

6. Describe an intelligent data entry system. Why are data typically entered directly from

the questionnaire into the data entry device?

Intelligent data entry refers to the use of a computer and a software package to assist a keyboard

7. What is the purpose of machine cleaning the data? Give some examples of how data can

be machine cleaned. Do you think that machine cleaning is an expensive and unnecessary

step in the data tabulation process? Why or why not?

The purpose of machine cleaning is to ensure that all keypunches are in the proper columns and

8. It has been said that a cross-tabulation of two variables offers the researcher more

insightful information than does a one-way frequency table. Why might this true? Give an

example.

A cross-tabulation between two variables provides the researcher with data that looks at the