The Analysis Process Basic Concepts of Editing, Coding and Descriptive Analysis

1
The Analysis ProcessBasic Concepts of Editing,
Coding and Descriptive Analysis

• Chapter 14

2
The Analysis Process

• Categorization, aggregation and manipulation of
  data to obtain answers to the research questions.
• Special aspect the interpretation taking the
  results, making relevant inferences and drawing
  managerial useful conclusions.

3
Overview of the Analysis Process

• Tabulation
• Identify categories, sort data into categories,
  make initial counts and use summarizing measures
  to facilitate understanding.
• 2. Formulating additional hypotheses
• Inductions concerning variables, their
  parameters, differences and relationships.
• 3. Making inferences
• Conclusions about the important variables, their
  parameters, differences and relationships.

4
Steps of Data Tabulation

• Categorize
• Code
• Create data file
• Error checking
• Generate new variables
• Weight data subclasses
• Tabulate

5
Tabulation 1) Defining Categories

• Identification of response categories early in
  the study results in higher consistency
• Pre-coding can eliminate transcription
• Conditions for useful classifications
• Similarity of response within the category
• Differences of responses between categories
• Mutually exclusive categories
• Categories should be exhaustive

6
Tabulation 2) Editing and Coding

• Editing should be centralized and conducted as
  the data is being collected to ensure maximum
  accuracy and clarity
• Entries must be evaluated for (1) Legibility,
  (2) Completeness, (3) Consistency and (4)
  Accuracy.
• Coding is the process of assigning data to
  categories. Types pre-coding, post-coding.

7
Tabulation 3) Cleaning the Data

• Data matrix rectangular array of entries
• Basic characteristics central tendency,
  dispersion and categories of responses.
• Dealing with
• Missing data use it as it is, delete the
  respondent, or use statistical imputation
• Outliers answers inconsistent with the data set
  can be discarded only after careful examination
• Multiple coders edit the data file or make
  corrections in the analysis program.
• Weighting the sample data adjusting the final
  sample so that specific subgroups are found in
  identical proportions to those in the population.

8
Basic Tabulation Analysis

• Final step in data collection and first step in
  the analytical process counting the number of
  responses in each data category.
• Simple (or marginal) tabulation the frequency
  distribution
• Cross-tabulation simultaneous counting of the
  number of observations in each of the categories
  of two or more variables

9
Summarizing Data

• It is desirable to summarize data by computing
  descriptive measures.
• Descriptive measures reduce the data set into
  simple, precise and meaningful figures
• Measures of Central Tendency
• Arithmetic mean Semi-interquartile range
• Median Variance
• Mode Standard deviation
• Range Coefficient of variation

10
The shape of the distribution
Skewness - gives information about the tails of
the distribution Types positive (tail to the
right) and negative (tail to the left) Kurtosis
- shape of the distribution in terms of height or
flatness
11
Analyzing Associative Data

• Cross-tabulation the simplest form of
  associative data analysis that allows further
  insight into lower-order (e.g., two-variable)
  associations.
• It provides a means of data display and analysis
  that easily interpretable
• Provides clear insights into complex marketing
  phenomena
• Affords a more readily constructed link between
  research and action
• May lessen the problems of sparse cell values
• The entities being cross-classified units of
  association.
• Cross-tabulation shows frequency data and row and
  column percentages.
• How to identify the direction in which
  percentages should be computed
• How to you interpret percentage of change.
  (absolute and relative difference the percentage
  of possible change in percentages)

12

• For test area
• Absolute increase 24 percentage point
• Relative increase (66-420)/42x100 57
• Percentage Possible Increase 100 /
  24/(100-42) 41 of the maximum possible

13
Introducing a third variable

• Example male/female
• Other possible relationships in cross-tabulation
  suppressor effects, explanation.
• Proper formulation adoption is associated with
  gender (rather than caused by)

14
Presentation of Descriptive Analyses

• Visual representations are critical and almost
  mandatory
• Examples
• stacked bar and pie charts
• plotting and smoothing techniques
• enhancement and slideshow capabilities
• links to geographical area maps (used in market
  segmentation analysis), etc.