Methods of Administering Questionnaires

Structured Communication

• The degree of standardization employed with the data collection instrument or survey.

Fixed Alternative Questions

• Responses are limited to the provided options.
• This is a close-ended question type.
• It gathers primary data.
• It is the most reliable type of response.
• Polls at the beginning of chapters are an example.

Unstructured Communication

• Open-ended questions allow participants to respond in their own words.
• It uses disguised communication to conceal a study's purpose or sponsor.
• It fosters a more organic setting for data collection.

Disguise in Research

• Reveal the knowledge about the study's purpose or sponsor to the respondent.

Methods of Administering Questionnaires

• Personal Interviews are direct, face-to-face conversations
  • They can be biased and expensive.
  • Mall intercepts are an example.
• Telephone Interviews are conducted via phone
  • They are less expensive than personal interviews.
  • Gather a limited amount of information.
  • Yield good response rates and use random digit dialing.
• Mail Questionnaires dispatched through postal services
  • Provide limited control over responses.
  • Have no interviewer bias, are anonymous, and have low cost.
• Internet-Based Questionnaires are done online, often with recruitment via email
  • They are inexpensive
• Computer-Assisted Interviewing manages questions and records answers electronically.

Measurement

• Measurement follows rules for assigning numbers to represent attribute quantities.

Scales of Measurement

• Nominal Scale assigns numbers for identification
  • For example, gender can be coded as 1 for male and 2 for female.
  • The central tendency measure is the mode.
• Ordinal Scale uses numbers to indicate order
  • For example, ranking as 1st, 2nd, or 3rd
  • The central tendency measure is the median.
• Interval Scale allows comparison of the size of differences
  • It includes properties of both nominal and ordinal scales
  • The central tendency is measured by the mean.
  • A temperature scale is an example
• Ratio Scale includes an absolute zero, allowing for magnitude comparison
  • Examples are sales volume and weight
  • It contains properties of nominal, ordinal, and interval scales
  • Geometric mean measures central tendency.

Zero in Scales

• Interval Scale Zero indicates a point on the scale, but not an absence
  • Temperature is an example.
• Ratio Scale Zero signifies the complete absence of the attribute
  • No sales is an example

Attitude Measurement Techniques

• Self-Report relies on individuals reporting their own beliefs or feelings.
• Itemized Rating Scales present respondents with response categories to select.
• Summated Ratings Scale (Likert Scale) measures agreement or disagreement.
• Semantic Differential Scale uses bipolar adjective pairs.
• Snake Diagram is a visual representation of semantic differential scale responses.
• Graphic Rating Scale involves marking a position on a continuous line
• Comparative Ratings Scale involves relative judgments.
• Constant-Sum Method requires respondents to divide a fixed sum among attributes.
• Global Measure gives an overall assessment using one or two items.
• Composite Measure assesses all relevant aspects of an object.

Reliability & Validity

• Reliability means a measure's consistency over time and between evaluators.
• Validity refers to a measure's accuracy in reflecting true differences.

Survey Design Considerations to Avoid Common Errors in Questions:

• Use simple words.
• Avoid ambiguous words.
• Avoid leading questions.
• Avoid unstated alternatives.
• Avoid assumed consequences and generalizations.
• Avoid double-barreled questions.

Question Sequencing

• Start with simple, engaging questions
• Follow a funnel approach
• Design branching questions carefully and ask classification information last
• Sensitive questions should be placed near the end.

Sampling Procedures

• Define the Target Population by clearly specifying inclusion criteria
  • Parameter refers to a population characteristic.
  • Statistic refers to a sample characteristic.
• Identify the Sampling Frame as a list of population elements.
• Select a Sampling Procedure
  • Non-Probability Sampling is judgment-based and not generalizable
    • Convenience Samples use readily available participants.
    • Judgment Samples use handpicked respondents.
    • Snowball Samples recruit others from initial respondents.
    • Quota Samples match population proportions
  • Probability Sampling is randomized and generalizable
    • Each element in a Simple Random Sample has an equal chance for selection.
    • Systematic Sample selects every kth element.
    • Stratified Sample divides the population into groups with sampling within each.
    • Cluster Sample consists of randomly selecting entire groups.
  • Area Sample uses geographic regions.
• Determine Sample Size
  • Sample size is influenced by diversity/variation, precision needed, and the required confidence level.

Errors in Sampling & Surveys

• Sampling Error is the difference between sample and full population results.
• Non-Sampling Errors include
  • Noncoverage Error caused by missing elements in the sampling frame.
  • Nonresponse Error which occurs when failing to get responses
    • Can be caused by refusals or non-at-homes
  • Response Error occurs due to inaccurate answers.
  • Office Error consists of mistakes in data processing

Response Rate Calculation

• General Formula: response rate = (completed interviews) / (eligible units).
• Web & Mail Surveys: response rate = (usable questionnaires) / (contacts attempted - bad addresses).
• Telephone Surveys (No Eligibility Check): response rate = (completed interviews) / (completed interviews + refusals + not-at-homes).
• Telephone Surveys (With Eligibility Check): response rate = completed interviews / [completed interviews + (eligibility % × (refusals + not-at-homes))].

Questionnaire Design & Pretesting

• Physical Characteristics should be professional and easy to complete
• Pretest includes a trial run on a small sample to refine the questionnaire.

Other Definitions

• Sample is a subset of a population.
• Census involves data collection from every member of a population.

Explore structured vs. unstructured communication methods, including fixed alternative and open-ended questions. Understand the advantages, disadvantages, and examples of personal and telephone interviews. Learn about disguise in research.