Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
59 Cards in this Set
- Front
- Back
Observation |
Give us a sense of context but not an understanding of how people think and feel |
|
Descriptive Questions |
Provide sample of informants language and understanding of some phenomenon |
|
Structural questions |
Enable researcher to discover info about how informants have organized their knowledge |
|
Contrast Questions |
Help us with meanings "What's the difference between x and y?" |
|
Grand tour questions |
Asked to extract a verbal description of a consumption scene through space, time, events, people, etc. Can be typical or in regards to a specific event (ie"what was the first day of this class like for you?" |
|
Probing Questions |
In cases where responses are short or off base, asks additional or improvised questions to get more info Informational: "What do you mean by x?" Nudging: "I see", "go on", "what else?" Clearinghouse: "Is there anything else you remember?" Reflective: "So you thought this because of x?" |
|
Interview Protocol |
Should move from general to specific, begin with very open ended questions |
|
Nondirection |
When interviewing, you should not direct an informants responses (ie "Was the service too slow or too high priced?" bad bad bad) |
|
Discovery (interviewing) |
Revealing as little of your own way of thinking as you can while discovering an informants categories of thought (language, style, images) |
|
Markers |
Verbal and nonverbal signals that there is emotion behind a topic: " |
|
Opinions |
Never express these sh*ts while interviewing, never imply you know anything about the topic |
|
Survey |
Involves interviews with a large number of respondents using a predesigned questionnaire |
|
Person-administered surveys |
interviewer reads questions face to face or over the phone, they record respondents answers (+ feedback, rapport, quality control, adaptability) (- human error, slow, high cost, respondents may not want to answer truthfully) |
|
Computer assisted surveys |
Interviewer basically verbalizes questions while relying to some degree on computer tech to facilitate interview work (+ Helps cut human error, no need to transcribe) (- Still expensive, still has interviewer effects) |
|
Self-administered surveys |
Respondent completes the survey on his or her own (pen and paper, usually mailed NO oversight) (+ reduced cost, respondents control pace, no interviewee apprehension or embarrassment) (- respondent cannot clarify questions, lack of responses) |
|
Mixed mode (hybrid) surveys |
Multiple data collection methods, increasingly popular in recent years (online surveys for portion of population with internet and phone calling for those without) (- mode may affect response, more complex) |
|
Survey Advantages |
Standardization Ease of administration Ability to tap "unseen" Suitability to tabulation and statistical analysis(easy to analyze) Sensitivity to subgroup differences |
|
Computer administered survey |
The respondent completes survey on a computer with no oversight (+ User-friendly features, inexpensive) (- Needs computer-familiar respondents, can't always verify respondents or answers) |
|
Accuracy vs. precision |
Accuracy: Darts hit the bullseye Precision: Darts hit near one another |
|
Objective properties |
Physically verifiable (Age, Gender) |
|
Subjective properties |
Mental constructs (Preferred Brand, opinion[good, bad, fair]) |
|
Properties |
Specific features or characteristics of an object that can be used to distinguish it from another object. |
|
Construct |
Something we want to measure (customer satisfaction, purchase interest, consumption rate, attitude toward brand, age) |
|
Elements of a construct |
Must be carefully defined, know why you're measuring it, why it's important. |
|
Ways to measure (nominal) |
Numbers used as labels and used solely for identification, you can only count number of responses (i.e. most checked category) can't calculate averages, weakest scale |
|
Ways to measure (ordinal) |
Numbers possess property of rank order, can calculate median and mode (category into which 50th percentile response falls) |
|
Interval Scale |
Numbers possess property of rank order and differences in scale values are meaningful, can calculate mode, median, and mean plus standard deviation (degree of deviation of numbers from mean) |
|
Ratio Scale |
Numbers possess characteristics of interval scale plus ratios of numbers on these scales have meaningful interpretations Can calculate median, mean, mode, and standard deviation of responses Has true zero and interpretable ratios, (how old are you? Hourly earnings? |
|
Attributes |
Personal or demographic characteristics (education level, age, size of household) |
|
Behavior |
Relate to behaviors such as frequency of visits, extent of magazine readership |
|
Beliefs |
Relates to knowledge and what respondents believe to be true |
|
Attitudes |
Similar to a belief but reflect the respondents' evaluative judgement |
|
4 "Do's" of question wording |
Question is focused on single issue or topic Question is brief Question is grammatically simple Question is crystal clear |
|
4 "Don'ts" of question wording |
Don't use a "double barreled" question (2 different questions in 1) Don't use words that overstate the case (undue emphasis on some aspect of the question) |
|
Questionnaire organization |
pertains to sequence of statements and questions that make up a questionnaire (well-organized: motivate respondents to be conscientious and complete) (poorly-organized: discourage or frustrates, can cause respondent to stop early) Typically move from general to specific (funnel sequence) |
|
Screening question |
used to cut out respondents who do not meet qualifications necessary to participate in the study (less sensitive and helps build rapport) |
|
Question flow |
pertains to the sequencing of questions including any instructions Warm-up questions are simple and used to get respondents' interest and to demonstrate ease of responding to research request |
|
Transitions |
Used to let respondent know that a change in question topic or format are about to happen |
|
Skip logic |
Helps respondents from answering questions that don't pertain to them (directs online survey to ask questions based on previous answers) |
|
Classification questions |
used to classify respondents (generally done in the last section because they are considered personal and can be offensive, usually demographic based) |
|
Pretest |
Dry run of a questionnaire to find and repair where difficulties that respondents encounter while taking a survey |
|
Statistical inferences |
talk to a sub-group of a group we care about to make generalizations of the overall group based on that smaller groups responses |
|
Sample |
Sub-group of the group we care about |
|
Population |
Bigger group we care about |
|
Sampling |
Technique used to select the sub-group |
|
Probability |
All members of the population have a known probability for getting into the sample |
|
Non-probability |
We pick the people for our sample for a reason |
|
Sampling frame |
a listing of units from which the sample is chosen |
|
Census Study |
Draw inferences or conclusions from entire population (Best if population is small, easily accessible) (Necessary if population units are extremely varied) |
|
Proportionate Stratified Random Sampling |
ssdfd |
|
Disproportionate Stratified Random Sampling |
sdfssdf |
|
Cluster Sampling |
Clusters of population are selected at random then all or some of the units in chosen clusters are studied (best for naturally occurring groups like households) |
|
Systematic Sampling |
First unit is selected randomly then remaining are selected using sampling interval (X=Units in population/units required for sample randomly choose 1 unit between 1->X then take every Xth unit) |
|
Parameter |
Actual population mean value or population proportion (mew) |
|
Statistic |
Estimate of a parameter from sample data |
|
Sampling error |
Difference between statistic value and parameter value, can only be determined through census study |
|
Sampling distribution |
representation of sample statistic values, chosen from a population by using a sampling procedure along with relative frequency of occurrence of those statistic values (we ask 400 people how likely they are to buy, 1-7 401 possible results of survey represent vertical lines, their height represents probability of that result if 24% of people in population will say "definitely", then 400*.24=96 so we can expect 96 respondents to say "definitely" |
|
Standard error |
Standard deviation of different sample statistic values that will be obtained through repeated selection of samples from the same population Average amount of sampling error associated with sampling procedure |
|
Three Sigma Rule |
68-95-99.7, for a normal distribution almost all values lie within 3 standard deviations of the mean |