Chapter 9 Flashcards Preview

Open: answer in own words

Closed: limits respondents to pre-determined alternatives 

ADVANTAGES: can answer more completely, reveal reasoning, more likely to discover something not anticipated in design

DISADVANTAGES: harder to code b/c answers in narrative form, categorize responses after (breaks cardinal rule) 

ADVANTAGES: easier to code, fewer off-the-wall answers, dont need to think as hard or be as articulate to formulate 

DISADVANTAGES: issues may be too complex to reduce to small set of alternatives, may not agree with any, puts words in mouth, misinterpretation of Q is hard to discover

by using a verification key - a set of questions designed to detect dishonest answers & if a person has tendency to be overly influenced by social desirability 

tendency to agree with statement on question regardless of its content

set of Qs that enable respondents to move through questionnaire in different ways depending upon responses

• can skip innapropriate/irrelevant Qs & more through more efficiently

1. face-to-face
2. written responses
3. computerized administration
4. telephone administration

• establish rapport
• direct attention to material
• motive them to answer carefully
• guarantee order of administration
• correct misinterpretations

• social desirability tendency
• expensive
• time-consuming

• efficient use of time & money
• low cost

• lower response rate
• possible misinterpretation
• can't determine seriousnes of respondent

• control sequencing & branching
• reduce social desirability
• available 24/7
• allow high level of participation without high cost

• same problems as written responses
• security of responding
  • storage of survey info must be ensured without violating privacy of respondent

• low cost
• computer-assisted interview
• conducted from central location & interviewers can be supervised

• reduced external validity →less people with home phones
•  

- data from people motivated to respond

- biased in direction of more vocal people

- quality of data is direct function of response rate

→researchers need at least 50% & prefer 90%

1) Haphazard

2) Purposive

3) Convenience

4) Probability

popn subgroup selected using hit-or-miss

→ does not follow any systematic way of selecting participants

(

non-random sample chosen for some characteristic it possesses

ex) survey opinions of presidents of several leading colleges about desirable changes in college curriculum

non-random sample chosen for practical reasons

sample selected in way as to be representative of population

- most rely on random sampling

- permit one to apply various statistics

sample in which every member of popn has equal & independant chance of being selected

population as it is defined for purposes of selecting subjects for study

→popn that you work with for particular study

individual member of a sampling frame

probability sample that is not randomly selected

structure may result in bias

taking every nth individual of structured list may result in bias

1) Simple Random

2) Stratified Random

3) Cluster

group chosen from entire popn such that every member has equal & independant chance of being selected in a single sample

2 or more subsamples are represented according to some predetermined proportion, generally in same proportion as they exist in popn

(treat popn as 2 or more subpopulations & create seperate random sample of each)

ex) same-sex ratio, age ratio etc.

group selected using clusters/groupings from larger population

ex) rather than taking 1/10 students of each class, randomly select 1/10 of classes to study

form of cluster sampling in which clusters are further broken down by taking samples from each cluster

ex) selecting one or more highschols within each geographical area & then selecting seniors within each

gathering info about people anonymously using standardized series of Qs

1) Preamble → intro of purpose & instructions

2) Identification Questions→ demographic Qs, background info

3) Research Questions → make up data

4) Branching Instructions

1) Fixed-Option

2) Rating Scale

3) Yes-No

4) True-False

5) Likert-Scale (Agree-Disagree)

6) Funnel Questions

7) Open-ended Questions

Qs with qualitatively distinct response alternatives

(type/kind)

Qs with quantitatively distinct response alternatives

• 3 types

1) Categorical → quantitative dimension but response based on adjective

2) Numerical → represented more explicitly in numerical form

3) Graphical → for responses on continuum (mark vertical line along horizontal line)

Binary-choice Qs on matters of fact

Binary decision Qs on particular theme

Qs on attitudes or opinions

series of related Qs soliciting progressively more detailed information

discourages dishonest answers b/c have to follow-up

branching items within

Qs that respondents can answer in their own words

1) Question Bias

2) Response Bias

1) Double-barreled

2) Loaded

3) Leading

4) Double Negatives

items that ask 2 seperate Qs simultaneously

items that contain emotionally charged or value-laden term

"do you feel sleazy magazine should be openly displayed at local newstands?"

items that tend to influence responses in a certain direction

do you agree with most federal policy analysts that reducing deficit is our most urgent fiscal priority?

items that use 2 negative expressions to convey a positive idea

1) Acquiescence ("Yea-Saying")

2) Dissension ("Nay-Saying")

3) Deviation

4) Social Desirability

1) tendency to agree with items regardless of their content

2) endency to disagree with items regardless of their content

Balanced Items

example:

T/F Qs:

some of my high school friends enjoy reading

none of my high school friends enjoy reading

tendency to answer items in extreme way regardless of their content

tendency to answer questions in socially acceptable way regardless of correctness of answer

Forced-choice Inventory

Example:

Choose A or B:

• A: I like to read in my spare time
• B: I am active in sports
• A: I feel like blaming others when things go wrong
• B: I feel that I am generally inferior to others

1) avoid words with double meaning

2) avoid overly lengthy questions

3) ask questions based on respondent's own experience

4) Design entire questionnaire around theoretical rationale

5) Pre-test questionnaire on test sample

administering interview/questionnaire to sample of individuals to estimate characteristic of population

probability that true population value falls within tolerance level 

range of values within which true population value falls with specified probability

1) Feasability → less expensive/time-consuming 

2) Accuracy → less errors (people may not answer seriously or misinterpret data) 

3) Unobtrusiveness

1) confidence & tolerance levels 

2) type of sampling design used

3) variability (increases with increased size)

Probability Sampling

Nonprobability Sampling 

Probability →known probability of each person falling in sample 

→allows for statistical analysis

Nonprobability → unknown probability of individual being in sample

1) Simple Random

2) Cluster

3) Stratified

Sampling such that every member of population has equal & independent chance of being selected

random sampling in (2) phases:

• primary selection units (cluster, groups) 
• secondary selection units (individuals within groups) 

random sampling from mutually exclusive & exhaustive strata in population 

1) Proportional Stratified Sampling

2) Equal Stratified Sampling 

sample size of each stratum is proportionate to population size

(each stratum has the same sampling fraction)

each strata is determined percentage of sample 

according to who is around 

non-random sample chosen for some characteristic it possesses

chosen for availability/convenience

picked in systematic way (violates independance principle)

try to match up characteristics (ex. male vs female)

administering standardized test to measure sample of individual’s behavior 

1) Aptitude

2) Achievement

3) Personality

4) Projective