Exam 2 Flashcards
Why are samples necessary?
save time & money
Why is sampling on the dependent variable a problem?
creates selection bias & can lead to incorrect conclusions
What are the potential problems with random sampling?
the sampling frame might not be complete & low response rates
What is a representative sample?
a sample looks like the population from which is was selected in all relevant respects
What are the factors affecting the likelihood of getting a representative sample?
sample size & population makeup
What are the four types of random sampling?
- Simple Random Sampling- selecting elements in such a way that each individual has an equal chance of being selected
- Systemic Random Sampling- elements are selected from a list where every nth element is selected
- Cluster Sampling- elements are selected in two or more stages; the first is random selection of naturally occurring clusters & the second is selection of elements within those clusters
- Stratified Random Sampling- sample elements are selected separately from population strata that the researcher identifies in advance
What are the four types of non-probability sampling?
- Availability/ Convenience Sampling- anyone available is included in the sample
- Quota Sampling- choose some characteristics that must be representative
- Purposive Sampling- sampling only people with a particular knowledge of a subject
- Snowball Sampling- Ask the person you talk with to tell you someone else you could interview
What are some reasons why true experiments may not always be feasible?
too long, too costly, unethical, & treatment in past
What are the five threats to internal (casual) validity? Know the definitions and examples of each!
- Non-comparable Groups- when the experimental group and the control group are not comparable
- Endogenous Change- when natural developments or changes in the subjects account for change rather than the experimental treatment
- History- events external to the study that influence post-test scores
- Contamination- treatment and control groups are aware of the other and this influences the post-test
- Treatment Misidentification- when effects are not due to treatment, but some unknown intervening process
What are the two ways we may get non-comparable groups?
selection bias (occurs when characteristics of experimental and comparison group subjects differ in any way that influences the outcome) & attrition (when groups become different because subjects in one group are more likely to drop out)
What are the four types of endogenous change?
- Testing- taking the pretest itself can influence the post-test scores
- Maturation- subjects change because of experience, not because of treatment
- Regression to the Mean- when a variable is extreme on its first measurement, it will tend to be closer to the average in subsequent measurements
- Instrument Decay- when research produces different results later in the research because the measurement tool changes
What are the three threats to external validity (generalizability)?
- Sample Generalizability- do the results hold outside of this lab?
- Cross-Population Generalizability- do the results hold over different populations over different time periods?
- Intention of Testing & Treatment- there is only an effect when there is a pretest taken
What are the six potential problems in survey research?
- Confusing Phrasing
- Biased/ Leading Phrasing
- Allow for Disagreement
- Ask Answerable Questions
- Exhaustive & Mutually Exclusive
- Allow for Uncertainty
What are the five ways to administer a survey? What are the advantages and disadvantages of each?
- Mailed (pro- cheap; con- low response)
- Group Administered (pros- large audience & large response rate; cons- coercion & difficult to get people in these types of settings)
- Phone (pro- reach a lot of people; con- cannot do random digit dialing)
- In Person (1:1) (pro- survey can be complex; con- very expensive)
- Online (pros- easy to record & composite answers, can use complex survey designs, & global reach; cons- internet access required & easy to misrepresent who you are)
What is the difference between a bar chart and a histogram? When should you use each?
Bar Chart= bars not connected
Histogram= bars connected
Bar Chart- used when the variable you are describing has discrete categories (typically normal)
Histogram- graphic for continuous variables displayed by adjacent bars
What are the three measures of central tendency?
- Mean
- Median
- Mode
What are the two measures of variation? Which is better?
- Range
- Variance
- Standard Deviation= best
What is a Cross-Tabulation Table? What can it tell us? What can it not tell us?
Cross-Tabulation Table- data tables that present the results of the entire group of respondents, as well as results from subgroups of survey respondents
can tell us: correlation
cannot tell us: causation
Define Population
the entire group you wish to learn about
Define Sampling Frame
a list of all the elements in a population
Define Sample
subset of the population about which we wish to learn
Define Element
the individual units of a sample
Define Sample Generalizability
the element to which a study can inform us about persons, places, or events that were not directly studied
Define Sampling Bias
something other than random chance determines selection into sample
Define Association
variation in one variable is empirically related to variation in another variable
Define Time Order
variation in the presumed cause must occur before variation in the presumed effect
Define Spuriousness
false/ non-genuine
Define Causal Bias
systematic error in the strength or sign of an estimate of a relationship of interest
Define Random Assignment
a procedure by which each experimental subject is placed in a group based solely on chance
Define Matching
a procedure to equate certain characteristics of individuals in different comparison groups
Define Quasi-Experimental Designs
a research design in which there is a comparison group that is comparable to the experimental group in important ways but subjects are not randomly assigned
Define Push-Polling
using a survey to distribute rather than collect information
Define Statistical Significance
the mathematical likelihood that an association is not due to random chance
ESSAY QUESTION
casual effect- variation in the dependent variable follows variation in the independent variable, when everything else is held equal
three necessary elements of causal effect:
- Association- variations in variable are actually related
- Time Order- said cause must happen before said effect
- Nonspuriousness- variation between two variables is not caused by variation in third variable
Why is each of the elements as an important aspect of establishing causal effect?
association- if we violated association, we would lack causation and instead have correlation
time order- if we did not have time order, than it would be impossible for the said cause to actually cause the said effect because it came after the effect was in place
nonspuriousness- if there was spuriousness and a third variable was the cause then there may not have actually been a relationship between the first two variables
two additional things to help strengthen an argument for causal validity:
- a causal mechanism- a describable process that creates a connection between two variables
- context- the larger set of interrelated circumstances in which a particular outcome should be understood