Reading...
Play button
Play button
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;
66 Cards in this Set
- Front
- Back
|
Sampling
|
The selection of a relatively small group of individuals from whom we obtain data in order to be able to generalize about a larger group.
|
|
|
Population
|
The term used to describe all the possible cases of interest
|
|
|
Stratum
|
A particular subgroup within the population
|
|
|
Element (of the population)
|
Any individual case in the population
|
|
|
Sampling Frame
|
list all the elements so that a sample can be drawn from the population. The sampling frame is such a list
|
|
|
Probability sample
|
every element of the population has a known, though not necessarily equal, chance of being,selected for inclusion. Furthermore, every element has at least some chance (a nonzero chance) of being part of the sample.
|
|
|
Non Probability sampling
|
particularly well . suited for exploratory studies, where the focus is on the generation of theory and research ideas. It is also useful in observational and qualitative
---Three most important types are: Accidental, Quota, Purposive |
|
|
Accidental Sampling (Convenience sampling)
|
Interviewing anyone who happens to stroll by
---There is no reliable basis for determining whether an accidental sample is representative. Therefore, we[~ must be particularly cautious about generaliz"i ing from the data acquired through accidental sampling. |
|
|
Quota Sampling
|
interviewers are told to screen potential respondents in terms of desired characteristics
--Ex: 50% black women, 50% white women, still basically stopping people to interview but with parameters. |
|
|
Purposive Sampling (judgment sampling)
|
The researcher purposely selects certain groups or individuals for their relevance to the issue being studied.
---This sampling method is used in studies of deviance or other social phenomena that are too rare to be dealt with effectively using a.representative Cross section of the population. |
|
|
Simple Random Sampling
|
each element of the population has an equal chance of being included.
---A complete listing of all population elements must be available. In other words, an adequate sampling frame must be located ---use of an incomplete sampling frame can lead to biased results. |
|
|
Parameter
|
Characteristic of total population
|
|
|
Statistic
|
Characteristics of sample
|
|
|
Systematic Sampling
|
A random starting point is selected on this list and every "nth" name or unit is selected from that point on.
---similar to simple random sampling |
|
|
Selection Interval (systematic sampling)
|
dividing the population size (16,000) by the desired sample size (400).
|
|
|
PERIODICITY (Systematic Sampling)
|
the sampling frame that is used has a regular, recurring pattern or cycle.
|
|
|
Stratified sampling
|
dividing the population into two or more strata and then taking either a simple random sample or a systematic sample from each stratum.
(Notice that quota sampling is very similar to this procedure. The difference is that accidental samples are taken from different strata in quota sampling.) The trick is to form strata that (lIeinternally homogeneous yet different from one another. |
|
|
Proportionate sampling
|
The general procedure in stratified random sampling is to sample from each stratum according to its percentage in the total population
|
|
|
Disproportionate sampling
|
the strata are not sampled according to their percentages in the population.
|
|
|
Cluster sampling
|
A probability sampling procedure in which elements of population are randomly selected in naturally occurring grouping (cluster)
|
|
|
Single Stage Cluster Sampling
|
Ex: Study homeless people living in shelters
---randomly select shelter D in City ---Include all homeless staying at Shelter D in study |
|
|
Two Stage Cluster Sampling
|
Ex: Steps 1 and two the same as Single stage CS
---However include additional shelters ---Random sampling from each cluster --- Random sampling from shelter D, K, include all these people in sample. |
|
|
Multistage Cluster sampling
|
Involves the repetition of two basic steps: listing and sampling
Ex: first select all university and 2 year college from Washington ---then simple random sampling from all the colleges and Universities from Wa ---Obtain list of students from these schools ---simple random sample from all students |
|
|
Non Sampling error
|
Distortion caused by various sources of non sampling error
---5 types |
|
|
Poor sampling frame (non sampling error)
|
You use a poor sampling frame that does not represent the population
Ex: Literary Digest Poll 1936 ---Sampling frame: Wealthy individuals who leaned republican ---Bias: excluded middle/working class, independents, democrats ---Good sample: cross section of American Society |
|
|
Non-Response Error (non sampling error)
|
People refusing to participate
poorly worded questions very long surveys privacy issues no incentive |
|
|
Bias in Panels (non sampling error)
|
---Occurs in Longitudinal studies
---People drop out of study over time ---drop out rates skew results |
|
|
Bias in respondent selection (non sampling error)
|
Interviewer problems
---failure to adhere to procedures discourage by payment by the hour |
|
|
Snowball Sampling (qualitative research)
|
It is especially helpful when a complete or reliable sampling frame is unavailable or when access to appropriate subjects for interviewing and observation is difficult.
EX: Secret societies, deviant groups At the end of interview ask if there are any other persons willing to be interviewed that share same experiences |
|
|
Coming up with a Research Question
|
Is the project doable given time and financial constraints?
Has question been answered before? Institutional constriants |
|
|
Unit of Analysis
|
Level of social life which a research question is focused--Individuals, groups, towns, nations
---Data that is used to describe differences between X, X=Unit of analysis Ex: Do countries with strong maternity leave policies affect women wages? Unit of analysis: countries. |
|
|
Units of Observation
|
Level at which you collect the data
--Ex: do states with anti abortion laws have higher number of illegal abortions? ---collect data on number of botched illegal abortion in each state Unit of observation: number of botched abortions. |
|
|
Types of research: Basic research
|
The goal of basic research is the production of knowledge.
Typically researchers who do this kind of work are not greatly concerned with such practicalities. Their goal is to add to our stock of social knowledge. |
|
|
Types of research: Applied research
|
Generally is done in response to a request from someone with a problem (a policy maker from some government unit)
Ex: What is the best response to the gang problem in our area? |
|
|
Evaluation research
|
Is used to monitor specific programs or policies and especially to assess their outcomes
|
|
|
Lit Review
|
Revising existing literature on your topic
--what has been done before --if your question has been answered --specify your research question (narrow as possible) |
|
|
Surveys (2 general types)
|
Interviews: writing in own answers to questions in survey
Self-administered questionnaires --phone --web |
|
|
Observation
|
Observe behavior in natural context
--labor intensive --only small groups can be observed --ability to influence individuals depending on context |
|
|
Use of Existing Statistics
|
Unobtrusive
--these data are readily available in libraries and online, and are great starting places for researchers |
|
|
Content Analysis
|
invloves subjecting some text to careful scrutiny to see what it reveals about its author
--common materials ---documents ---music Unobtrusive |
|
|
Longitudinal Research
|
Data collected that can be ordered in time
---changes in society/people over time ---same/similar questions asked over time |
|
|
Panel study (longitudinal research)
|
Is one that follows the same people over time
---ideal type of longitudinal study |
|
|
Cohort Study (longitudinal research)
|
Sampling of different cohorts of people over time to compare
---birth cohort: what year you were born in |
|
|
Trend studies
|
These are studies in which researchers ask the same questions periodically.
The general idea is to track certain issues or behaviors over time |
|
|
Cross-Sectional design
|
Study in which data is collected at only one point in time
---study of labor market and wages ---faster, cheaper, easier |
|
|
Quantitative research
|
Use of sampling techniques whose findings can be expressed numerically
--Data analysis mainly statistical Experience and situation as measurable Description --correlations: correlation coefficient between 0.00 and 1.00 |
|
|
Structure of quantitative research
|
Introduction--Lit review
Hypothesis--X causes Y Method--collection of data and how to analyze data analysis discussion of results conclude and recommendation |
|
|
Qualitative research
|
Descriptive--information is not in numerical form
diary accounts, open ended questions, unstructured interviews Data analysis includes the coding of data where little information is available |
|
|
Structure of qualitative research
|
Introduction--Lit review
Description--narrative analysis Discussion and conclusion --no testable hypothesis in research itself |
|
|
When to use quantitative or qualitative research methods?
Go for Quantitative |
--Your research question
--what are your aims for doing research --interested to know cause and effect --ability to collect data --Statistical data --knowledge of statistical methods and analysis |
|
|
For Qualitative
|
Group individual of interest, little research done
--groups that are difficult to gain access to --want in depth knowledge |
|
|
When to use experiments?
|
When you want to isolate the cause on the effect
--Experiments strength: being able to isolate cause |
|
|
Experiments not suitable for:
|
--broader questions more suited to non probability and probability methods
--Longitudinal studies in the social science --Broad societal trends/attitudes |
|
|
Experiments: what is it suited for?
|
--Very narrow questions
--only interested if IV is the ONLY factor causing DV --Where question can be tested in a place where experiments can be able to control the environment, setting. |
|
|
External validity
|
--The ability to generalize your findings to the population
--good sampling frame ----includes everyone in population --probability sampling |
|
|
Threats to external validity
|
People who volunteer vs. people who don't may have different characteristics that influence experiment outcomes
vetting process for volunteers: ensure similarities to population |
|
|
Internal Validity
|
You are sure that the outcome of your experiment is due to your treatment
Establishing control over settings and environment; having proper protocol Consistency --How your experiments are carried out; exactly the same every time you repeat your experiments with different group |
|
|
Threats to internal validity
|
Testing effect
--effect of knowing your going to a test; makes people nervous History effect --effected by environment at the time of the experiment |
|
|
Hawthorne effect
|
Where people are reacting not to the experimental stimulus but the experiment itself.
Could be that people watching you during experiment. |
|
|
John Henry effect
|
Where your control group tries to be better than your experiment group/feels like competition.
|
|
|
Treatment group
|
Group that receives actual medication
|
|
|
Control group
|
Nothing really happens to them
|
|
|
Comparison group
|
Receives another drug/not drug you are testing for
|
|
|
True/Classic experiment 6 steps
|
1) subject selected randomly from population
2) have at least two groups (treatment and control) 3) Randomly assign subjects to group 4) Pre-test people in all groups for baseline 5) Carry out/ treat people in experiment group 6) post-test all groups |
|
|
Myth busters (problems with)
|
--no sampling frame
--to broad --no comparison group! |
|
|
How to make myth buster experiment generalizable?
|
Sampling frame: all registered voters
Narrow conditions for selection criteria for your sample sample size probability sampling method |
