Data Collection Methods and Questionnaire Design Principles

Data Collection Methods and Types

Primary Data vs. Secondary Data

Data can be categorized based on its origin:

Primary Data

Data that previously did not exist in our research and, therefore, we had to collect ourselves (e.g., from surveys).

• Advantages: Tailored to suit specific research needs.
• Disadvantages: High cost in terms of time and money.

Secondary Data

Data that had been collected in a previous independent study and is repurposed for use (e.g., from official sources like the INE, IDESCAT).

• Advantages:
  • Fast and cheap.
  • May contain information difficult to obtain from primary sources.
  • Saves effort.
• Disadvantages:
  • May not align well with the objectives of the current study.
  • May be measured in inadequate units.
  • May refer to a previous, potentially outdated, period.

Transverse and Longitudinal Data

These terms describe how observations are structured over time:

• Longitudinal Data: Observations of a single individual or group tracked over time.
• Transverse (Cross-sectional) Data: Observations of a set of individuals observed at a single, given moment in time.

Types of Surveys

Surveys are classified based on the medium and the presence of a pollster:

• Paper Surveys:
  • With Pollster: Personal (face-to-face) or telephone interviews.
  • Self-Administered: Mail-in, handout, or postcard surveys.
• Electronic Surveys:
  • With Pollster: Personal computer-assisted interviewing (CAPI) or computer-assisted telephone interviewing (CATI).
  • Self-Administered: Email, web-based forms, or disk/file distribution.

Design of a Questionnaire

The process of designing an effective questionnaire involves several critical steps:

1. Identifying information needs.
2. Defining goals and reviewing similar studies.
3. Conducting previous qualitative research and in-depth interviews.
4. Conceptual and operational definition of variables to measure.
5. Selecting the appropriate response scale for the questions.
6. Question wording and specifications.
7. Determining the order of questions.
8. Pilot testing the questionnaire.
9. Measuring attitudes and abstract variables effectively.

Question Formats and Variable Types

Questions must be formatted according to the type of variable being measured:

• Qualitative Variables: Single answer or multiple answers (checklists).
• Quantitative or Numeric Variables: Require specific numeric data input.
• Attitudinal Variables: Use a battery of questions, estimates of magnitudes, or semantic differential scales (e.g., Likert scales: very poor, poor, good, very good...).
• Rankings: Require respondents to sort items in order of preference.
• Open Questions: Questions without predefined answer choices.
  • Advantages: Captures expression and emotion; answer choices do not affect the outcome; useful when possible answers are unknown.
  • Disadvantages: Requires greater respondent effort; treatment and coding costs are higher; answers can be ambiguous, incomplete, or hasty.

Modes of Survey Administration

The mode of administration dictates the format and complexity of the questions:

• Self-Administered: Must include clear instructions since there is no pollster present.
• In-Person (Personal): Combines oral delivery and visual aids.
• Telephone: Requires short questions and oral recording of responses.

Questionnaire Structure and Flow

The arrangement and order of questions significantly impact response quality:

• Questions should be ordered logically by theme.
• Start with simple questions concerning the main issue.
• Continue with more complex questions later.
• Questions that might cause abandonment or sensitivity should go toward the end.
• Instructions for entry and screening should be placed at the top.
• Early questions should not bias or discourage responses to later questions.
• The overall distribution and layout of the questions should be attractive and easy to follow.

Pilot Testing the Questionnaire

The pilot test is crucial for verification. It ensures that:

• The questions are clearly understood.
• The response options are adequate.
• Respondents answer in the way intended, and responses are consistent across different respondents.
• The pollsters perform their role well and maintain neutrality.
• Respondents and interviewers follow all instructions correctly.