Midterm 2 - Chpt. 7 Flashcards
You might need to ask your Ps some questions - why? (4)
- to analyze & understand trends
- for correlational research
- as the DV in an experiment
- as a manipulation check in an experiment
What kind of information might you need to find out from Ps? (3)
□ ₋ Attitudes and beliefs
□ ₋ Facts and demographics
□ ₋ Behavioral frequencies
Simplicity (Question wording considerations)
avoid jargon, abbreviations, complexity!
Double-barreled questions (Question wording considerations)
EX: Do you support free tuition and reduced healthcare?
- Do you feel anxious and depressed when thinking about exams?
Loaded Questions (Question wording considerations)
Are you in favor of having your terrific instructor Mark Lam teach another term of fun-filled 217?
Negative Wording (Question wording considerations)
Do you agree that the City of Vancouver should not allow the homeless to camp on the streets?
Creating Q’s - Question validity:
Does the question assess what you want it to assess? (more later in the term)
EX: A researcher want to measure the number of close friendships a person has, and asks:
How many friend requests do you have on FaceBook?
Is this a valid measure of close friendship?
How to validate/test question validity?
Pre-testing:
– Have others answer the questions
– Ask for feedback
– Revise questions that are unclear or misunderstood
- OR use an existing & validated questionnaire!
Response Sets (Additional Considerations for Surveys)
A tendency to respond to all questions from a particular perspective or in a particular way rather than to provide answers to the questions themselves.
Social Desirability (Additional Considerations for Surveys)
(aka “faking good”)
Tendency for a participant to respond in the most socially acceptable way, or as “most people” would. - more likely with sensitive topics
Additional Considerations for Surveys
- Response sets
- Extraversion
- Social desireability
- Open-ended vs. closed-ended
Pros of open ended Q’s:
- Can get rich, creative, unexpected, and/or complex answers
- Identify common themes
Cons of open ended Q’s:
- Reponses may be ambiguous
- P’s may interpret questions in different ways
- Time consuming to code
READINGS
Survey research uses questionnaires and interviews to ask people to provide information about themselves
○ Attitudes
○ Beliefs
○ Demographics
○ Past behaviours
○ Intended future actions
4 reasons to conduct surveys:
- Method to ask people to tell us about themselves (self-report)
- Society demands data about issues, rather than just intuition and anecdotes
- Survey data, if collected in a responsible fashion, can more useful information than individual anecdotes (□ For example, collecting student opinions to create effective curriculums)
- When using basic research, you can easily study many important variables using questionnaires or interviews
How can surveys be used? (3)
- To provide snapshots of how particular groups think and behave
- To gather data for studying relationships among variables
- Closed ended questions, with only certain numerical responses allowed, provide quantitative data that can be tracked over time, or to compare scores for one institution with similar institutions
- Open ended questions can be content analyzed for additional insight into what undergraduates think would enrich their education
Panel Study (How can surveys be used?)
when the same people are tracked & surveyed at two or more points in time, the design is sometimes called a panel study, or a longitudinal design
- Each period is known as a wave - a two wave study involves surveying people at two points in time
- Allow for research questions about the relationship between one variable (at time one) and another variable (at time two)
Response bias in survey research:
- If a minority of people provide inaccurate information, the data as a whole may remain valuable and informative
- Only when there’s an excess of inaccurate information that overshadows or masks the true variance in scores do have serious problems
Response Bias
a tendency to respond to all questions from a particular perspective, or in a particular way, rather than provide answers directly related to the questions themselves
Most common type of response bias?
social desirability: leads an individual to answer in the most socially acceptable way - the way the person thinks “most people” would respond, or the way that would present the person in the most positive light
- Occur during questions on sensitive topics
How can we prevent response bias?
- To ask sensitive questions in a way that hides their main intent
- Ensure privacy when completing the survey, by stressing confidentiality and anonymity, and increasing motivation to respond accurately
Defining the research objectives
Potentially unethical to ask q’s with no real use for information - more q’s asked, greater of a burden placed on the participant with little benefit
Attitudes & Beliefs (Defining the research objectives)
- Q’s focus on the ways people evaluate, feel, and think about issues
- Tap into important aspects of how people see the world and the value that different things hold for them
Facts & Demographics - Defining the research
Factual q’s ask people to indicate things they know to be true about themselves and their situation
◊ Can be necessary to describe your sample so that researchers can consider the population from which it was drawn
◊ Questions about ethnicity, income, marital status, employment status, etc.
Other factual info will depend on the topic of your research
Behaviours - Defining the research objectives
Focus on behaviours; either things people have done in the past or intend to do in the future
What 5 issues arise from question wording?
- Unfamiliar technical terms
- Vague or imprecise terms
- Grammatical errors
- Awkward phrasing that taxes working memory
- Embedding the question with misleading information
Unnecessary Complexity (Question Wording)
Survey q’s should be as simple as possible
Avoid jargon and technical terms
Write in a way that those without advanced degrees or familiarity with the language can understand
Double-Barelled Questions (Question Wording)
Q’s that ask two things at once
If you need to ask about more than one issue, ask using more than one question
Loaded Questions (Question Wording)
Written in a way to try and bias people’s responses toward a particular answer
To detect, watch out for emotionally charged words - meant to create an emotional reaction
Negative Wording (Question Wording)
Avoid framing questions in the negative, which can confuse respondents as to how to correctly indicate their preference
Can reduce scale reliability and validity
“Yeah-saying” and “Nay-saying” (Question Wording)
“Yea-saying” and “Nay-saying” response sets: when you ask several questions about a topic, a respondent may either agree or disagree with all of the questions
Agreeing with all q’s is AKA acquiescence bias: bias towards accepting the statement
A form of inattentive responding: people providing answers without even really thinking about them
Responding carelessly for times sake
How to detect “Yeah-saying” and “Nay-saying” (Question Wording)
- Word q’s so that true agreement with most items is unlikely
- Include q’s that request a particular response
AKA conscientious responders scale
Closed ended questions (Responses to Questions)
Closed-ended: a limited number of response alternative are given, like with a multiple choice question
- Easier to code (categorizing), by converting the response options to numerical data that can be easily analyzed
- More likely. Used when the variables under the study are well-known and well-defined
- Favoured by quantitative approaches
Open ended questions (Responses to Questions)
Open-ended questions: respondents are free to answer in any way they like
Harder to code, using content analysis or other qualitative methods - ex: a participant’s answer doesn’t make much sense, requires more time to analyze
However, a much richer source of information because of the wide variety of responses possible
Avoid missing out on possibly useful info
Good for when you don’t know enough to make educated assumptions about what sorts of answers people might provide
Can be used by both, but favoured by qualitative
Rating scales for close-ended questions (Responses to Questions)
Rating scale: asks people to provide judgements of “how much” for a dimension - amount of ___
Easiest/most direct presents people with five or seven response alternatives
What type of labelled scale is most efficient?
Fully labelled scales are more reliable than are partially labelled scales
More clear definitions and reduce measurement error that comes from relying on people’s interpretations
What is the ideal quantity of a scale? What other feature should be added?
In basic research, it’s preferrable to have more alternatives to allow people greater opportunity to express themselves
5-7 point scales, with a middle “neutral” option
However, participants tend to use this middle point when they don’t know how to respond, which can cause issues for reliability and validity
Often preferred to have an “I don’t know” option
High/Low Frequency scales
Labelling alternatives can be tricky when asking a q about the frequency of some behaviour
- High frequency scales: most of the options measure a behaviour done very frequently
- Low-frequency scales: behaviour in question doesn’t happen very frequently
Graphic Rating Scale
asks respondents to make a mark along a continuous 100mm line, anchored with descriptions at each end
Researcher measures where the mark was made, converting to a score from 0-100
Semantic Differential Scale
a way to measure the meaning that people ascribe to concepts
Respondents rate concepts based on a continuum spanning two adjectives, using multi-point scales (Wearing shorts in the winter: “good”, “strong”, “smart”
Concepts using this scale are rated among 3 dimensions:
Evaluation: how something is valued using adjectives such as good-bad, wise-foolish
Activity: active-passive, slow-fast
Potency: weak-strong, hard-soft