Collecting Data 3

Question 1

What are surveys and how are they used in data collection?

Answer 1

Surveys are a method of primary data collection where information is directly gathered from people. They typically involve either interviews or questionnaires to collect responses from participants.

2ways conduct surveys & collect opinions: interviews and questionnaires.

Question 2

What are the main types of interviews used in surveys?

Answer 2

There are two main types of interviews:

• Structured Interviews: These involve asking the same set of questions to each respondent in a consistent manner, often using a questionnaire.
• Unstructured Interviews: These are more like open-ended conversations, where the interviewer has the flexibility to explore topics in depth, making them suitable for qualitative insights.

Question 3

How should interviews be structured for small-scale versus large-scale data collection?

Answer 3

For small-scale or qualitative research, interviews can be unstructured. However, for large-scale data collection, it is better to use a structured approach with a standardized questionnaire to ensure consistency and completeness.

Question 4

Interviews can be conducted through which modes?

Answer 4

• Face to face
• Technology based (such as telephone or skype)

Question 5

What is the principle behind collecting information through interviews?

Answer 5

The principle is straightforward: someone poses questions, and then listens to and records the answers.

Question 6

Why is the role of the interviewer crucial?

Answer 6

The interviewer’s role is crucial because they need proper training to:

• Avoid introducing bias (interviewer bias) by influencing responses through the way questions are asked or their non-verbal cues.
• Ensure accurate recording of answers.

Question 7

What is interviewer bias and how can it affect survey results?

Answer 7

Interviewer bias occurs when the interviewer unintentionally influences the respondent’s answers through their tone, expressions, or the way they ask questions. This can lead to skewed data, making it crucial for interviewers to be well-trained to conduct interviews without leading the respondents.

Question 8

What are the advantages and drawbacks of personal interviews?

Answer 8

Personal interviews offer depth and intimacy but are costly and time-consuming, as one interviewer can only handle one interview at a time.

Question 9

What are focus groups, and how do they fit into the survey process?

Answer 9

Focus groups are small groups of people interviewed together to generate discussion and insights. While they can provide in-depth qualitative data, focus groups are prone to bias due to group dynamics and are often non-representative of the broader population.

Question 10

Why do focus groups have a poor reputation despite their theoretical advantages?

Answer 10

Focus groups have a poor reputation due to several criticisms:

• Selection Bias: Participants are often not selected randomly and sometimes selection is deliberately biased.
• Overgeneralization: Results from a single, small focus group are sometimes treated as if they are from a large-scale investigation, leading to misleading conclusions.
• Manipulation: Focus groups can be conducted in a way that ensures results align with predetermined outcomes, rather than producing honest findings.

Question 11

How are questionnaires distributed, and what are the pros and cons of each method?

Answer 11

Questionnaires can be distributed in various ways:

• Postal Mail: Inexpensive but typically has low response rates (often below 20%).
• Online Surveys: Can reach a wide audience at a low cost, but may introduce bias depending on the demographic’s access to technology.

Question 12

What challenges are associated with response rates in questionnaires?

Answer 12

Low response rates can introduce bias, as certain types of people might be more likely to respond than others. This makes the results less representative of the entire population. Incentives can help increase response rates but must be carefully designed to avoid skewing the results.

Question 13

What are the potential issues with the accuracy of responses in questionnaires?

Answer 13

Since questionnaires are typically self-administered, there’s no control over how truthfully or accurately respondents answer the questions. Additionally, unclear questions can lead to misunderstandings, resulting in inaccurate data. This makes careful questionnaire design crucial.

Question 14

Why is questionnaire design so important?

Answer 14

The design of a questionnaire is vital because it directly impacts the clarity of the questions and the reliability of the data collected. Poorly designed questions can lead to misunderstandings and inaccurate responses, undermining the validity of the survey results.

Question 15

How can you keep a questionnaire engaging and efficient?

Answer 15

To keep a questionnaire engaging and efficient:

• Ask Related Questions in Series: Group related questions together for coherence.
• Be Brief: Make the questionnaire as short as possible to maintain respondent interest.
• Follow a Logical Sequence: Arrange questions in a logical order to help respondents follow the flow.
• Avoid Irrelevant Questions: Include only questions that are relevant to the objectives.
• Focus Each Question on One Topic: Ensure each question addresses only one issue to avoid confusion.

Question 16

How can you ensure questions are clear and understandable?

Answer 16

To ensure clarity and understandability:

• Use Simple Language: Avoid jargon and complex terms; use everyday language.
• Avoid Calculations and Memory Tests: Do not include questions that require complex calculations or memory recall.
• Avoid Vague Questions: Be specific in questions to avoid ambiguity. For example, ask about specific quantities or frequencies instead of vague comparisons.
• Handle Sensitive Topics Carefully: Phrase personal questions, such as those about age, religion, or sex, delicately to avoid offending respondents.

Question 17

How can you get honest and accurate responses from respondents?

Answer 17

To obtain honest and accurate responses:

• Use Neutral Wording: Avoid leading questions. For example, instead of asking if someone “likes” something, ask them to rate it on a scale.
• Minimize Bias: Ensure that the wording of questions does not influence or bias the responses, aiming for impartiality to get genuine feedback.

Question 18

What are some guidelines for designing effective questionnaires?

Answer 18

• Avoid open-ended questions where possible. They require more effort to analyze and categorize responses.
• Use questions that have numeric answers or are selected from a restricted list (e.g., scales of “one to five” or “yes or no”).
• For opinions, use a rating scale with predefined responses, such as “Strongly agree” to “Strongly disagree.” This type of scale produces data that is easier to analyze statistically.

Question 19

What is a Likert scale and how is it used in questionnaires?

Answer 19

A Likert scale is a type of rating scale used to measure attitudes or opinions. Respondents select from a range of predefined responses, such as “Strongly agree” to “Strongly disagree.” This type of scale helps produce structured data that can be analyzed more easily.

Question 20

What is a major problem with using questionnaires, regardless of how they are delivered?

Answer 20

A major problem is that there is no way to verify if a questionnaire has been answered truthfully or if the respondent has understood the questions properly. There’s no opportunity to explain the questions to respondents or clarify points they may not understand.

Question 21

What is response bias in the context of questionnaires?

Answer 21

Response bias refers to inaccuracies in answers that can occur due to various factors, such as misunderstanding questions, social desirability (giving answers that are perceived as more acceptable), or guessing.

Question 22

How can the design of a questionnaire help reduce response bias?

Answer 22

To reduce response bias, a questionnaire should be well-designed by:

• Crafting clear and precise questions.
• Pilot testing the questionnaire with a small sample to identify and address issues.
• Avoiding leading questions that could bias responses.
• Providing balanced response options to prevent biasing answers.

Question 23

What is a pilot survey?

Answer 23

A pilot survey is a preliminary test of your questionnaire conducted on a small sample of respondents before using it for full data collection.

Question 24

Why should you conduct a pilot survey?

Answer 24

Conducting a pilot survey helps identify issues with the questionnaire, such as confusing questions or unclear instructions, which can save time and money by addressing these problems early on.

Question 25

What are the benefits of a pilot survey?

Answer 25

• Identifying Issues: Uncovers problems with the questionnaire that need fixing.
• Saving Time and Money: Allows for corrections before full-scale data collection, avoiding the cost and effort of redoing data collection.
• Refining the Questionnaire: Provides an opportunity to make adjustments based on feedback, improving the clarity and effectiveness of the questions.

Question 26

How many times should you pilot the questionnaire?

Answer 26

You may need to adjust and re-pilot your questionnaire several times to get it right.

Question 27

Why is it important to use a different group of people for each pilot test?

Answer 27

Using different groups ensures that the changes you make are genuine improvements and that the questionnaire performs well across diverse respondents.

Question 28

What are the 3 main types of error that can appear in survey methods?

Answer 28

• Sampling error
• Response error
• Non response error

Question 29

What is sampling error in survey methods?

Answer 29

Sampling error occurs when the sample selected for the survey does not accurately represent the entire population you want to study. This can happen if the sample is too small, not diverse enough, or not selected randomly. To minimize sampling error, carefully define your target population and use appropriate sampling techniques to ensure the sample reflects this population as accurately as possible.

Question 30

What is response error in surveys?

Answer 30

Response error arises when there are issues with how respondents answer questions. This can include misunderstanding the questions, providing guesses rather than informed answers, or inaccuracies in recording responses. To reduce response error, ensure that questions are clear and easy to understand and implement measures to verify the accuracy of responses and data recording processes.

Question 31

What is non-response error in surveys?

Answer 31

Non-response error occurs when respondents either refuse to participate in the survey or answer only some of the questions. It includes complete non-response (where the individual does not participate at all) and partial non-response (where the individual participates but skips some questions). Analyzing patterns of non-response can help identify issues with the survey or questions and provide valuable insights for improvement.

Question 32

Why is missing data important in survey analysis?

Answer 32

Missing data should not be ignored as it can indicate underlying issues with the survey or sample. Examining missing data can reveal patterns and potential problems, such as biases or issues with the questionnaire. This analysis is crucial for improving survey design and addressing any issues with the sample or questions.

Question 33

What are the three main types of missingness in statistics?

Answer 33

• Missing Completely at Random (MCAR)
• Missing at Random (MAR)
• Missing Not at Random (MNAR)

Question 34

What does “Missing Completely at Random (MCAR)” mean?

Answer 34

MCAR occurs when the missing data is completely unrelated to both the variables in the dataset and the respondents themselves. The missingness is purely random, meaning there’s no systematic pattern to the missing data.

MCAR means that if some data is missing its missing for random reasons.

Question 35

Why is it important to identify if data is MCAR?

Answer 35

If data is MCAR, it means the missing data does not introduce any bias, and the analysis results can still be valid. However, this situation is quite rare in real-world data.

Question 36

How can you check if data is MCAR?

Answer 36

To check if data is MCAR, you need to look for patterns or lack thereof in the missing data. Here’s how you can do it:

• Compare Missing Data Across Questions:
  Look at the proportion of missing responses for each question in your dataset. If the missing data is MCAR, the proportion of missing responses should be roughly the same across different questions. For example, if one question has a high percentage of missing answers but others do not, this could indicate that the data is not MCAR.
• Compare Missing Data Across Groups:
  Compare the missing data across different groups within your sample (e.g., by gender, age, etc.). If data is MCAR, the missingness should be equally distributed among these groups. For instance, if missing responses are not more common in one gender or age group compared to another, it supports the idea that the missing data is random and not influenced by these characteristics.

Question 37

What does “Missing at Random (MAR)” mean?

Answer 37

MAR occurs when the missingness is related to the observed variables in the dataset but not to the missing data itself. For example, some questions might be skipped more often, but this is not related to the characteristics of the respondents.

Question 38

What are common reasons for data being MAR?

Answer 38

Data might be MAR if a question is poorly worded, sensitive, or requires a detailed response that respondents might avoid. It could also happen if a question appears at the bottom of a survey and is accidentally missed.

Question 39

What does “Missing Not at Random (MNAR)” mean?

Answer 39

MNAR happens when the missing data is directly related to the unobserved values. This means that the reason data is missing is tied to the specific data that is missing, which can introduce significant bias.

Question 40

Can you give an example of MNAR?

Answer 40

A classic example of MNAR is a question about age where older respondents are more likely to skip the question. The missing data is not random but is related to the specific characteristic (age) of the respondents.

Question 41

Why is MNAR the most problematic type of missingness?

Answer 41

MNAR is problematic because it introduces bias that is difficult to correct. Since the missingness is related to the data itself, it can lead to skewed results that do not accurately reflect the population.

Question 42

Why is it important to understand the type of missingness in your data?

Answer 42

Understanding the type of missingness helps in deciding how to handle the missing data and ensures the reliability of your analysis. MCAR is the least problematic, MAR requires careful handling, and MNAR is the most challenging due to the inherent bias.