chapter 5 and 6 Flashcards
define population in a research design
large group of interest to a researcher
sample
small set of individuals from a population who participate in a study.
Intended to represent population
target population
group defined by the researcher’s specific interests. Individuals usually share one characteristic
accessible population
portion of the target population consisting of individuals who are accessible to be recruited as participants in the study
representativeness of a sample
extent to which the characteristics of the sample accurately reflect those of the pop
representative sample
sample with the same characteristics as the pop
biased sample
sample with diff characteristics from those of the pop
selection bias or sampling bias
occurs when participants or subjects are selected in a manner that increases the probability of obtaining a biased sample
2 basic categories of sampling methods:
probability and non-probability
probability sampling
the odds of selecting a particular individual are known and can be calculated
3 conditions for probability sampling:
- exact size of pop is known and possible to list all individuals
- each individual in the pop must have a specified probability of selection
- when a group of individuals are assigned the same probability, the selection process must be unbiased. Must be a RANDOM PROCESS
nonprobability sampling
odds of selecting a particular individual are not known. Researcher does not know population size and cannot list the members of a pop
which runs a higher risk of producing a biased sample, probability or nonprobability sampling?
nonprobability
Simple random sampling
each individual in a pop has an equal chance of being selected.
Each selection is independent of the others.
2 principa methods of random sampling
- sampling with replacement
2. sampling without replacement
problem with simple random sampling
there is a chance of selecting a very distorted sample
Systematic sampling
Prob sampling.
selecting every nth person. The “nth” is calculated by dividing population size by the desired sample size.
principle of independence is violated, but ensures high degree of representativeness
Statified random sampling
Prob sampling
select equal-sized random samples from several SUBGROUPS in a population and combine them into one overall sample.
This ensures that individuals from each subgroup will be represented in the study
Cluster sampling
prob sampling.
randomly selecting pre-existing naturally formed groups.
e.g., researcher wants to obtain a large sample of third-grade students from the city school system. Instead of selecting 300 students one at a time, the researcher can randomly select 10 classrooms of 30 students.
advantage of cluster sampling
- relatively quick and easy
- measurement of individuals can be done in groups, which can facilitate research project.
proportionate stratified sampling
like stratified sampling, but instead of selecting an equal sized sample from each subgroup, you select the proportions seen in the pop. e.g., if a class has 75 females and 25 males, you select more females than males in the two subgroups
Convenience sampling
nonprobability sampling.
- researchers use participants who are easy to get, based on their ability and willingness to respond.
- easier, less expensive
Quota sampling
nonprobability sampling technique.
-generating subgroups in a convenience sample to increase representitaveness
what does WEIRD stand for?
Western Educated Industrial Rich Democratic
4 statistical values in power analyses
- sample size
- effect size
- significance level = P(type I error) (prob of finding an effect that is not there)
- power = 1 - P(Type II error) (probability of finding an effect that is there
3 OUT OF 4 WILL ALLOW YOU TO CALCULATE MISSING VALUE
2 options for selecting sample sizes
- power analyses
2. precedent (ie. what have other people done?)
T/F: LARGE sample size can make it easier to reach statistical significance, but this does not always mean practical significance
true
What kinds of variables/measurements can you use in descriptive research?
nominal, ordinal, interval/ratio
Research strategies
- Descriptive research
- Correlational research
- Experimental research
- Non-experimental reseacrh
- Quasi-experimental research
describe experimental research
- seeks to answer CAUSE-AND-EFFECT questions about relationship between 2 variables
- does X cause Y?
- there is a level of control over variables
Nonexperimental research
- demonstrate relationship between variables without explaining the relationship
- in what way does X affect Y?
- GROUPS ARE PRE-EXISTING (e.g., males and females)
- NO VARIABLES ADDED
- less control than in experiment
- e.g., developmental research (you can’t assign an age to someone)
Quasi-experimental research
-cause and effect relationships
-less control than experiment, but more than nonexperimental research
-add another level of control
-PRE-EXISTING GROUPS EXIST ALREADY
-
Experiment vs nonexperimental studies vs quasi-experimental
Experiments are designed to demonstrate cause-and-effect relationships by producing unambiguous explanations and demonstrating that changes in one variable are responsible for causing changes in another. Quasi-experimental studies aim to demonstrate cause-and-effect relationships but fall short of achieving this goal. Non-experimental research attempts to demonstrate that a relationship exists, without attempting to explain it
What is “external validity?”
extent to which we can generalize the results of a research study to people, settings, times, measures, and characteristics other than those used in the study
What is a threat to external validity?
any characteristic of a study that limits the ability to generalize the results from a research study
3 different kinds of generalization:
- generalization from a sample to the general population
- generalization from one research study to another
- generalization to a real-world situation
Internal validity
a research study has internal validity if it produces a single, unambiguous explanation for the relationship between two variables.
-In other words, change in IV is the ONLY explanation for a change in the DV
threat to internal validity
any factor that allows an alternative explanation for the results of a study
___ ____ is a variable (not of interest) that could impact results
extraneous variable
what is a confounding variable?
an extraneous variable which varies systematically with the variable of interest
selection bias is a threat to ___ validity
external
what is the problem with volunteer bias?
it threatens external validity. Hard to generalize results with volunteers to individuals who may not volunteer to participate in studies
novelty effect
experiments can be novel, anxiety-producing experiences for participants. This causes them to act differently than they would in real-life situations.
multiple treatment interference
after multiple treatment conditions, participants may become fatigued or practiced in the task. This is hard to generalize to people who do not have the same experience
Experimenter characteristics
can be a threat to external validity.
Results can be specific to an experimenter with a certain set of characteristics, both demographic and personality, which can limit the generality of results.
MRS SMITH
- Maturation
- Regression to the mean
- Selection of subjects
- Selection by maturation interaction
- Mortality
- Instrumentation
- Testing
- History
Maturation
physiological processes occurring within the participants that could account for any changes in their behaviour
Regression to the mean
outliers get closer to the average
Selection of subjects
bias in selecting and assigning participants to the groups that result in systematic differences between participants
Selection by maturation interaction
2 groups would have grown apart even without the treatment. They would have developed differently regardless.
Mortality
differential dropping out of some subjects from the comparison groups before experiment is complete
Instrumentation
changes in a measuring instrument over time.
Threat to internal validity because any observed difference between treatment conditions may be caused by changes in the measuring instrument instead of the treatment.
Testing effects
Threat to internal validity that occurs when participants are exposed to more than one treatment and their responses are affected by participation in an earlier treatment. e.g., fatigue and practice
History effect
threat to internal validity.
Refers to environmental events other than the treatment that change over time and may affect the scores in one treatment differently than in another treatment.
e.g., fire alarm goes off in the middle of the night and ppl must stay up for an hour. Later in the day, memory test might not be good because they didn’t get enough sleep
