Individuals may not admit their true beliefs about energy conservation when influenced by group consensus.
Respondents could answer positively due to perceived social norms rather than honest opinion, presenting a challenge for researchers in achieving truthful responses from participants.
Survey Design Issues
Double-Barreled Questions
Definition: A double-barreled question combines two separate queries into one, leading to confusion and inadequate responses.
Example: The question "Do you find using a cell phone to be convenient and time-saving?"
Issue: It inquires about two separate aspects: convenience and time-saving, which may elicit conflicting responses.
Consequences: If a participant feels the cell phone is convenient but not time-saving, they cannot express this distinction.
Recommendations: Split double-barreled questions into separate questions; e.g.,
"Do you think that cell phones are convenient?" (Response: Yes/No)
"Do you think that cell phones are time-saving?" (Response: Yes/No)
Clarity in Questions
Importance of phrasing questions clearly to avoid misunderstandings.
Example Question: "How long have you lived in San Angelo?"
Potential Responses: Participants may give vague answers (e.g., "since I was six years old") without the needed context.
Improved Question: Rephrase to "How many years have you lived in San Angelo?" to elicit precise, quantifiable responses.
Improved clarity enhances reliability of data collected.
Pilot Testing for Surveys
Advise conducting pilot tests of surveys by friends or family prior to releasing to actual participants.
Benefits:
Identifies issues in question phrasing and instructions
Ensures clarity and intelligibility of questions
Terms: Pilot Data refers to practice data collected from trial runs of the survey.
Response Bias
Definition of Response Bias
Definition: Occurs when participants do not genuinely express their thoughts, often opting for uniform answers just to complete the survey.
Example: Survey responses might be filled out quickly, selecting the same answer (e.g., always selecting 'A') to avoid engaging with the content.
Common Scenario: End-of-year evaluations where students circle the same grades quickly out of lack of interest.
Mitigation Strategies
Reverse Scoring/Reverse Coding: Rephrase questions to help identify participants who might be responding lazily or under constraints.
Example: Pair statements that require opposing responses (e.g., "I frequently feel sad" vs. "I am happy most of the time").
Purpose: Identifies inconsistent responses indicating lack of honest engagement by the participant.
Survey Administration Techniques
Advantages of Electronic Surveys
Time-efficient and cost-effective method for data collection.
Greater reach leading to a diverse respondent demographic, decreasing sampling bias.
Eliminates interviewer bias as participants respond privately without pressure.Shifts in privacy may lead to more truthful answers on sensitive subjects.
Disadvantages of Mail Surveys
Costly due to printing and postage.
Time-consuming, as responses must be manually entered into databases for analysis.
Risks of human error increase compared to electronic responses.
Correlation Concepts
Definition of Correlations
Correlation: A statistical measure that expresses the extent to which two variables are related.
Importance: Helps in observing relationships but does not imply causation ("correlation does not equal causation").
Correlation Coefficient
Description: A number between -1 and +1 indicating the strength and direction of a relationship between two variables.
Key Points:
Ranges from -1 (perfect negative correlation) to +1 (perfect positive correlation).
0 indicates no correlation.
Interpretation:
Strength: Closer to -1 or +1 indicates a stronger relationship, while closer to 0 indicates a weaker relationship.
Direction: Positive or negative sign indicates the direction of the relationship.
Examples of Correlation
Positive Correlation Example
Hypothesis: Length of lecture correlated with number of yawns from students.
Research: As lecture length increases, the count of yawns also increases.
Correlation Coefficient: If calculated, a value of +0.82 suggests a strong positive correlation.
Negative Correlation Example
Hypothesis: Length of lecture related to students' attention level.
Research: Longer lectures relate to reduced attention.
Correlation Coefficient: If found to be -0.78, demonstrates a strong negative correlation.
No Correlation Example
Hypothesis: Milk consumption related to academic grade.
Research Findings: Poor correlation observed; the pattern of data points is scattered with no significant relationship.
Correlation Coefficient: +0.05 indicates nearly no relationship between two variables.
Determining Strength of Correlation
Cutoffs:
Weak: r = 0.29 or lower.
Moderate: r = 0.30 to 0.69.
Strong: r = 0.70 to 1.0.
Application: Demonstrates the interaction or influence between variables; important for analyzing research results.