RMS lecture 17

Importance of Conserving Energy and Response Bias

  • 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.,
    1. "Do you think that cell phones are convenient?" (Response: Yes/No)
    2. "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.