Preparing Survey Data for
Analysis
Created @January 22, 2023 1:54 PM
Class RW: Survey
Type Reading
Stage? Drafting
Reviewed
Column
Materials
Reading 1 Chapter 9 (Fowler)
Reading 2
For analysis:
Survey Answers ———> Data Files
Preparing Survey Data: The process of taking completed questionnaires and survey
interviews and putting them into a form that can be read, processed and analysed. This
is also called CODING or Data Reduction.
The Process of Coding/ Data Reduction is as follows:
1. Designing the code: The rules by which a respondent’s answers will be assigned
values that can be processed by machine.
2. Coding: The process of turning responses into standard categories.
3. Data Entry: putting the data into computer readable form.
Preparing Survey Data for Analysis 1
, 4. Data Cleaning: Doing a final check on the data file for accuracy, completeness, and
consistency prior to the onset of analysis.
Errors that can occur during Data Reduction:
1. Coding Decision Errors: Misapplication of rules for equating answers and code
values
2. Transcription or Entry Errors: These can happen when an answer or a number is
recorded.
<The options for quality control are tied to the particular data collection, data entry,
and coding procedures chosen>
Formatting A Data File
Record= All data that pertain to a single individual case or interview.
The following points highlight HOW one should format a data file and WHY:
1. A serial identifier for each respondent usually goes in the same location for each
particular questionnaire or interview, usually at the beginning of a record. These
numbers are critical for checking files for completeness.
2. It eases coding, data entry, and programming tasks if the data are coded in the
order that they appear in the survey instrument. This will reduce errors at these
stages and represents a relatively cost-free means of quality control.
3. It probably is best to put a single positive entry in each field that contains data.
Some computer programs interpret blanks as zeros, whereas others do not. If zero
is meant, it is best actually to code a zero rather than leave a blank field; if the
intention is to have a code for item nonresponse, when no codable answer was
given, some specific value should be used.
Constructing a CODE
A code is a set of rules that translates answers into numbers.
Preparing Survey Data for Analysis 2
Analysis
Created @January 22, 2023 1:54 PM
Class RW: Survey
Type Reading
Stage? Drafting
Reviewed
Column
Materials
Reading 1 Chapter 9 (Fowler)
Reading 2
For analysis:
Survey Answers ———> Data Files
Preparing Survey Data: The process of taking completed questionnaires and survey
interviews and putting them into a form that can be read, processed and analysed. This
is also called CODING or Data Reduction.
The Process of Coding/ Data Reduction is as follows:
1. Designing the code: The rules by which a respondent’s answers will be assigned
values that can be processed by machine.
2. Coding: The process of turning responses into standard categories.
3. Data Entry: putting the data into computer readable form.
Preparing Survey Data for Analysis 1
, 4. Data Cleaning: Doing a final check on the data file for accuracy, completeness, and
consistency prior to the onset of analysis.
Errors that can occur during Data Reduction:
1. Coding Decision Errors: Misapplication of rules for equating answers and code
values
2. Transcription or Entry Errors: These can happen when an answer or a number is
recorded.
<The options for quality control are tied to the particular data collection, data entry,
and coding procedures chosen>
Formatting A Data File
Record= All data that pertain to a single individual case or interview.
The following points highlight HOW one should format a data file and WHY:
1. A serial identifier for each respondent usually goes in the same location for each
particular questionnaire or interview, usually at the beginning of a record. These
numbers are critical for checking files for completeness.
2. It eases coding, data entry, and programming tasks if the data are coded in the
order that they appear in the survey instrument. This will reduce errors at these
stages and represents a relatively cost-free means of quality control.
3. It probably is best to put a single positive entry in each field that contains data.
Some computer programs interpret blanks as zeros, whereas others do not. If zero
is meant, it is best actually to code a zero rather than leave a blank field; if the
intention is to have a code for item nonresponse, when no codable answer was
given, some specific value should be used.
Constructing a CODE
A code is a set of rules that translates answers into numbers.
Preparing Survey Data for Analysis 2
