Test 2 Chapter 7

Question 1

Defining research objectives

Answer 1

Three types of survey questions (dependent on what the researcher wants to know)

1. Attitudes and beliefs
   •ways people evaluate, think, and feel about issues
2. Facts and demographics
   •indicate what people know about themselves and their situation to describe your sample (age, gender), or other relevant information (income,marital status)
3. Behaviours
   •past or future behaviours (how many times did you workout last week, have you ever been so depressed you called in sick)

Question 2

Question wording

Answer 2

1. Unnecessary complexity
   •questions should be relatively simple
   •avoid jargon/terms people may not understand
   •avoid memory overload
2. Double-barreled questions
   •asking two things at once
   •difficult to answer (can think yes and no)
3. Loaded questions
   •leads people to respond a certain way
   •ex: “do you favour ___”
   •ex2: emotionally charging words: rape, waste, immoral, dangerous
4. Negative wording
   •”should not” leaves room for confusion
   •decreases reliability and validity
5. “Yea saying” and “Nay saying”
   •when asked several questions, a respondent may disagree or agree with all the questions
   •solution: reverse scoring (may reduce reliability)
6. Varying agree vs disagree in a list of questions
   •often done, but research shows causes of confusion

Question 3

Open vs closed ended questions

Answer 3

Can result in different outcomes
1. Open ended
•respondents are free to answer however they want
•takes time to code (qualitative), more costly, but more insight
2. Close ended
•limited number of response alternatives
•easier to code, more often used when variables are well defined

Way to measure close ended questions: rating scales
•ask people to prove “how much” judgements
•format depends on the topic
•most common is the strongly disagree to strongly agree

Question 4

Response alternatives

Answer 4

Fully labelled vs partially labeled
•fully labeled scales (words on each number value) are more reliable/less measurement error than partially labeled scales (words on first and last number value)

Number of response alternatives
•5-7 point scales allow for greater expression than yes vs no
•Middle can be seen as “i don’t know”: a problem for reliability/validity
•Time scales (sometimes. rarely, frequently) are vague and should be more precise:
•high frequency scales: less than twice a week–atleast once a day
•low frequency scales: less than once a month – more than once a week

Question 5

Types of rating scales

Answer 5

1. Graphic rating scale
   •a mark along a 100mm line that has wording at each end
   •ex: movie was enjoyable __100mm___the movie was not enjoyable
2. Semantic differential scale
   •way to measure the meaning that people ascribe to concepts (can measure anything with this technique)
   •have words then put checkmarks
   •ex: bad___/_____good
   •three types: evaluation (good/bad, wise/foolish, kind/cruel), activity (active/passive, slow/fast, excitable/calm), potency (weak/strong, hard/soft, large/small)
3. Nonverbal scale
   •uses images instead of words/numbers
   •ex: smily faces used for children ratings

Question 6

Repeated measures design

Answer 6

Alternative to random assignment by having the same individuals participate in both conditions (participants are measured on the DV after being in both conditions)

Advantages
1. fewer participants are needed because everyone participates in both conditions
•would be a good method when participants are scarce or when procedure is costly (fMRI)
2. extremely powerful/sensitive to detecting differences between levels of the IV
•when the same people are in both conditions, less of the variance is due to error because they take their problems with them

Disadvantages
•the order can become a confound affecting the DV (order effect)
•practice effect: they’re improving/getting better at the task
•fatigue effect: getting bored of repetitive questions
•contrast effect: they respond differently to the second condition because they experienced the first condition
*ex: perfume testing (fourth smell will be impacted by the third)
•this lacks external validity (perfume at home vs perfume in store)

Ways to combat order effects
1) ensuring time between conditions is long enough to minimize the influence of the first condition
2) counterbalancing: helps reveal contrast effects if they’re there
•randomly assign some participants to order 1 and some to order 2, allowing you to see the extent that order impacted results
•ex: randomly assign one group to drink coke first than the other to drink pepsi first and if the groups consistently like their first drink more than their second, theres a contrast effect

Question 7

Ways to achieve internal validity in a repeated measure design

Answer 7

1) Experimental control
•ensuring that only the IV changes across conditions with no confounds (ex: order)
•avoiding other alternative explanations: experimenter expectancy effects, demand characteristics, order effects
2) Minimize the effect of participant individual differences
•ensure all participants complete all conditions