Understanding Surveys and Research Data

• A. To manipulate participants into changing their behavior
• B. To collect data from a sample in order to describe or understand a larger population ✓
• C. To prove a single cause-and-effect relationship with certainty
• D. To replace the need for experiments in every field of study

Why: Surveys gather information from a sample of respondents so researchers can describe attitudes, behaviors, or characteristics and draw inferences about a broader population. They are descriptive tools rather than tools that by themselves establish causation.

• A. Describe your experience with the product in your own words.
• B. What do you think could be improved?
• C. How satisfied are you with the service? (Very satisfied, Satisfied, Neutral, Dissatisfied, Very dissatisfied) ✓
• D. Why did you choose this course?

Why: Closed-ended questions offer a fixed set of response options for respondents to choose from. Options A, B, and D all invite open-ended, free-text answers, which makes them open-ended questions.

• A. The question is written in a way that encourages respondents toward a particular answer ✓
• B. The question appears at the beginning of the survey
• C. The question asks respondents to rank several items
• D. The question uses technical vocabulary that respondents may not know

Why: A leading question nudges respondents toward a specific answer, for example by including a positive assumption like 'How much did you enjoy our excellent service?' Leading questions bias results and should be avoided in good survey design.

• A. To increase the number of respondents automatically
• B. To identify confusing wording, technical problems, or timing issues before the main data collection ✓
• C. To guarantee that the results will be statistically significant
• D. To avoid the need for informed consent

Why: Pilot testing with a small group reveals unclear questions, skipped items, and logistical problems. Fixing these issues beforehand improves data quality and saves resources once the full survey is launched.

• A. Reliability
• B. Validity ✓
• C. Response rate
• D. Sampling frame

Why: Validity refers to whether a survey actually measures the concept it claims to measure. Reliability, by contrast, refers to consistency of results across repeated measurements, while response rate and sampling frame relate to participation and population coverage.

• A. To collect data from every member of the population
• B. To study a smaller subset that represents the larger population ✓
• C. To guarantee that results will exactly match the population
• D. To avoid the need for any statistical analysis

Why: A sample is a smaller group selected from a population so researchers can draw conclusions about the whole population without surveying every member, saving time and cost while still providing useful estimates.

• A. The statistical formula used to calculate sample size
• B. The list or source from which the sample is actually drawn ✓
• C. The questionnaire used to collect responses
• D. The final report summarizing survey results

Why: A sampling frame is the list or database of all members of the population from which the sample is selected, such as a voter registry or a customer list. Its quality strongly affects how representative the sample will be.

• A. Researchers choose participants they believe are typical of the population
• B. Every member of the population has an equal chance of being selected ✓
• C. Participants are selected only from those who volunteer
• D. The population is divided into groups and one group is chosen

Why: Simple random sampling gives each member of the population an equal probability of selection, which reduces selection bias and allows researchers to use standard statistical methods to estimate sampling error.

• A. Random variation that naturally occurs in any sample
• B. A systematic error where some population members are more likely to be selected than others ✓
• C. The deliberate choice of a large sample size
• D. Errors made when entering survey data into a computer

Why: Sampling bias is a systematic distortion caused when the sampling method favors certain groups, so the sample no longer reflects the population accurately. Unlike random sampling error, it does not shrink as sample size increases.

• A. It ensures the survey will be cheaper to conduct
• B. It allows findings from the sample to be reasonably generalized to the population ✓
• C. It eliminates the need to write survey questions carefully
• D. It guarantees that all respondents answer honestly

Why: A representative sample mirrors the key characteristics of the population, so results observed in the sample can be generalized to the broader group. Without representativeness, conclusions may apply only to those who were surveyed.

• A. How satisfied are you with the service you received today?
• B. Don't you agree that our excellent service deserves a high rating? ✓
• C. How many hours per week do you spend on the internet?
• D. Which of these brands have you purchased in the last 12 months?

Why: A leading question like 'Don't you agree that our excellent service deserves a high rating?' signals the desired answer and pushes respondents toward agreement. Neutral wording such as the other options avoids steering the respondent.

• A. It uses too many words for respondents to read easily.
• B. It asks about two different things in one question, so a single answer cannot clearly represent both. ✓
• C. It is always answered more honestly than single-topic questions.
• D. It cannot be analyzed statistically.

Why: A double-barreled question combines two constructs, such as price and quality. A respondent who likes the quality but dislikes the price has no accurate way to answer, so the resulting data is ambiguous. Each concept should be asked separately.

• A. An open-ended text box with no guidance
• B. A yes/no question
• C. A Likert scale ranging from 'Strongly disagree' to 'Strongly agree' ✓
• D. A ranking of items from 1 to 10 with no ties allowed

Why: A Likert scale captures the intensity of agreement along an ordered set of response options, giving more nuanced data than a binary yes/no. It is the standard format for attitude and opinion measurement.

• A. So that respondents can select as many options as they like.
• B. So that every possible answer fits exactly one category without overlap or gaps. ✓
• C. So that the survey can be completed more quickly.
• D. So that the questionnaire looks more professional.

Why: Mutually exclusive options prevent a response from fitting more than one category, while exhaustive options ensure every respondent has an appropriate choice. Overlapping or missing categories make the data hard or impossible to interpret.

• A. Adding the word 'really' before 'exercise'
• B. Defining 'regularly' with a specific frequency, such as 'at least three times per week' ✓
• C. Asking it in a different language
• D. Making it a yes/no question with no other options

Why: The word 'regularly' means different things to different people, so answers would not be comparable. Specifying a concrete frequency gives every respondent the same reference point and produces more reliable data.

• A. Mail survey with no reminders
• B. Face-to-face interview ✓
• C. Online self-administered survey
• D. Telephone survey with no callbacks

Why: Face-to-face interviews generally achieve higher response rates than other modes because an interviewer can build rapport, explain the study, and persuade reluctant respondents to participate. Self-administered online surveys and mail surveys usually have lower rates, and telephone surveys have declined sharply due to call screening.

• A. They are always more expensive to conduct
• B. They cannot include any closed-ended questions
• C. They often suffer from lower response rates and coverage bias among people without internet access ✓
• D. They produce data that cannot be analyzed statistically

Why: Web surveys are inexpensive and convenient, but they often have lower response rates and can exclude people without reliable internet access or digital skills, creating coverage bias. They can include closed-ended questions and produce analyzable data, and they are typically cheaper rather than more expensive.

• A. The speed at which respondents complete the survey
• B. The percentage of sampled units that complete the survey as a proportion of eligible units ✓
• C. The percentage of questions answered correctly
• D. The number of follow-up reminders sent

Why: Response rate is the proportion of eligible sampled individuals or households who complete the survey. It is a key indicator of potential nonresponse bias, since low rates raise the risk that respondents differ systematically from nonrespondents.

• A. To change the answers respondents originally gave
• B. To increase the response rate by prompting people who forgot or delayed participating ✓
• C. To reduce the total number of questions
• D. To comply with legal requirements for every survey

Why: Many sampled people intend to respond but forget or postpone. Reminders are a low-cost way to raise response rates in self-administered modes. They do not change answers, and they are not a universal legal requirement.

• A. A short, clearly explained questionnaire
• B. A prior notification letter explaining the study's purpose
• C. A very long questionnaire with many sensitive questions ✓
• D. Offering a small incentive such as a gift card

Why: Long questionnaires increase respondent burden and fatigue, and sensitive questions can make people unwilling to answer. Short surveys, advance notice, and incentives are generally known to improve response rates.

• A. Running advanced statistical models
• B. Reviewing the raw data for obvious errors, inconsistencies, and missing values
• C. Publishing the results to stakeholders
• D. Weighting the responses by demographic group

• A. Keep it, since it is a valid number
• B. Round it down to 25
• C. Flag it as an out-of-range value and check the original response or treat it as missing
• D. Replace it with the average age of all respondents

• A. The respondent definitely selected that answer
• B. It preserves the information that the question was skipped or did not apply, so the value is not mistaken for a real answer
• C. It removes the case from the dataset entirely
• D. It indicates the data has been verified by a second coder

• A. Software cannot store more than one spelling of a word
• B. Inconsistent spellings would be treated as separate categories, distorting counts and frequencies
• C. It makes the survey shorter for respondents
• D. It is required by law for all research datasets

• A. Making all changes by hand directly in the original file without saving a copy
• B. Keeping the raw data unchanged and documenting each cleaning step in a script or log
• C. Deleting unusual responses so the analysis is easier to read
• D. Asking a colleague to remember which changes were made

• A. Mean
• B. Median ✓
• C. Mode
• D. Range

Why: The median is the middle value of an ordered data set. If there is an even number of values, it is the average of the two middle values. The mean is the arithmetic average, and the mode is the most frequent value.

• A. 3 ✓
• B. 4
• C. 3.8
• D. 5

Why: The mode is the most frequently occurring value. The number 3 appears four times, more than any other value, so the mode is 3.

• A. Variance
• B. Standard deviation
• C. Range ✓
• D. Interquartile range

Why: The range is the simplest measure of spread, found by subtracting the minimum value from the maximum value. Variance and standard deviation measure how far values typically fall from the mean, and the interquartile range measures the spread of the middle 50 percent of data.

• A. Line chart
• B. Scatter plot
• C. Histogram
• D. Pie chart ✓

Why: Pie charts display how parts make up a whole, making them well suited to showing proportions of categories. Line charts show trends over time, scatter plots show relationships between two variables, and histograms show the distribution of numerical data.

• A. The data has no outliers
• B. The data is skewed to the right by high-income outliers ✓
• C. The data is skewed to the left by low-income outliers
• D. The sample size is too small to be reliable

Why: When the mean is pulled well above the median, a small number of very large values is typically pulling the average upward. This indicates a right-skewed distribution caused by high-income outliers, and the median is often a better summary in such cases.

• A. Sampling bias
• B. Leading question bias ✓
• C. Nonresponse bias
• D. Recall bias

Why: The wording pushes respondents toward agreeing by praising the bike lanes ('wonderful improvement') and framing agreement as the expected answer. This is leading question bias, which distorts results by steering answers rather than measuring true opinions. A neutral version would be, 'How do you feel about the city's new bike lanes?'

• A. Measurement error
• B. Response bias
• C. Sampling bias ✓
• D. Confounding variable

Why: Sampling bias occurs when the people surveyed are not representative of the population being studied. Blog readers are likely more tech-savvy and more internet-active than the general population, so the poll overstates internet usage. A representative sample would require selecting participants randomly from the broader population.

• A. The sample size was too small
• B. The question wording was biased
• C. Correlation was mistaken for causation ✓
• D. The data was collected too recently

Why: The data shows a correlation, meaning two things occur together, but it does not prove one causes the other. People who buy gym memberships may already be healthier, wealthier, or more motivated, and these factors could explain the health difference. Establishing causation requires controlled experiments or careful analysis of confounding variables.

• A. Selection bias
• B. Social desirability bias ✓
• C. Survivorship bias
• D. Confirmation bias

Why: Social desirability bias happens when respondents answer in ways they believe will be viewed favorably by others. Admitting to heavy drinking may feel embarrassing, so people give answers that make them appear more moderate. Anonymous surveys and carefully worded questions can help reduce this bias.

• A. Nonresponse bias ✓
• B. Framing effect
• C. Ceiling effect
• D. Hawthorne effect

Why: Nonresponse bias occurs when people who choose not to participate differ systematically from those who do. For example, very dissatisfied or very busy employees might skip the survey, making overall satisfaction appear higher or lower than it truly is. A low response rate raises concern about how well the results represent the whole group.