Upgrade to remove ads
Research Methods in Psychology EXAM 2
Terms in this set (63)
What are the three characteristics of true experiments?
1. Some type of intervention is implemented
2. Marked by the high degree of control that an experimenter has over the arrangement of experimental conditions, assignment of participants, systematic manipulation of independent variables, and choice of dependent variables.
3. The experimenter establishes a proper comparison to evaluate the effectiveness of a treatment.
What are obstacles to conducting true experiments in natural settings?
-obtaining permission to do the research from individuals in positions of authority
-access to participants
-random assignment perceived as "unfair"
waiting-list control group
When there is more demand for a treatment than an agency can meet, these people complete the study measures while waiting for treatment.
our ability to generalize a study's findings to persons, settings, and times beyond those used in the study.
What is the best test of external validity?
Procedures that resemble characteristics of true experiments, for example, that some type of intervention or treatment is used and a comparison is provided, but are lacking in the degree of control that is found in true experiments
Just as randomization is the hallmark of true experiments, so (blank) is the hallmark of quasi-experiments.
lack of randomization
Why might researchers turn to quasi-experiments?
Sometimes true experiments are not feasible.
one-group pretest-posttest design
- 01 X 02
- known as a "bad experiment" or "pre-experimental design"
-Pretest is 1st observation (01)
-Treatment is implemented (X)
-Posttest is 2nd observation (02)
-None of the threats to internal validity are controlled
What does the prefix 'quasi' mean?
nonequivalent control group
-quasi-experimental procedure in which a comparison is made between control and treatment groups that have been established on some basis other than through random assignment of participants to groups
-pretest scores are used to determine whether the groups are equivalent
Threats to internal validity in the nonequivalent control groups design
-threats to internal validity that must be considered include additive effects with selection, differential regression, observer bias, contamination, and novelty effects
-although groups may be comparable on a pretest measure, this does not ensure that the groups are comparable in all possible ways that are relevant to the outcome of the study
simple interrupted time-series design
-quasi-experimental procedure in which changes in a dependent variable are observed for some period of time both before and after a treatment is introduced
-archival data are often used
-look for clear discontinuity in the time-series data for evidence of treatment effectiveness
time series with nonequivalent control group design
quasi-experimental procedure that improves on the validity of a simple time-series design by including a nonequivalent control group; both treatment and comparison groups are observed for a period of time both before and after the treatment.
What are the two disciplines of psychology?
1. Experimental psychology
2. Correlational psychology
-"Tight little island"
-measures impact of environment on individuals
-came into existence first
-"Holy roman empire"
-poorly chosen measures
What is the relationship between correlational and quasi-experimental studies?
From a statistical point of view, there is no difference
What is the only difference between correlational and quasi-experimental studies?
Quasi-experiments have a nominal independent variable (group membership) and correlational studies have an interval or ratio level independent variable.
With careful planning, cause can sometimes be ascertained when:
-Genetic variables are independent variables
-A set of correlations can together establish cuase by eliminated other possible explanations
limitations of correlational research
-one study can eliminate a cause-effect relationship
-the third variable problem
-one study cannot establish a cause-effect relationship by itself
-says little about interventions
-statistics can get very complex when there are moderator variables
a variable that has a different correlation between x and y, depending on the value of the moderator
process of recruiting and/or identifying study participants
the set of all cases of interest
-comprehensive list of population members
-most populations have incomplete or nonexistent sampling frame
subset of the population actually drawn from the sampling frame
each member of the population
(blank), not samples, are of primary interest
a sample exhibits the same distribution of characteristics as the population
the distribution of characteristics in the sample is systematically different from the target population
occurs when the procedures used to select the sample result in the overrrepresentation of some segment of the population, or, conversely, in the exclusion or underrepresentation of a significant segment.
What are the two categories of sampling methods?
1. probability sampling
2. nonprobability sampling
-each population member has a known, nonzero probability of being selected for the sample
-requires an accurate sampling frame
-a sampling technique in which there is no way to calculate the likelihood that a specific element of the population being studied will be chosen
-generally produce biased samples
simple random sampling
type of probability sampling in which each possible sample of a specified size in the population has an equal chance of being selected (GOLD STANDARD)
stratified random sampling
type of probability sampling in which t he population is divided into sub-populations called strata and random samples are drawn from each of these strata
A procedure in which the selected sampling units are spaced regularly throughout the population; that is, every n'th unit is selected.
the tendency to recur at regular intervals
-A probability sampling technique in which clusters of participants within the population of interest are selected at random, followed by data collection from all individuals in each cluster.
-takes advantage of naturally formed groups
-a probability sampling technique involving at least two stages: a random sample of clusters followed by a random sample of people within the selected clusters
-hierarchy of clusters
selecting respondents primarily on the basis of their availability and willingness to respond
As a general rule, you should consider that convenience sampling will result in:
a biased sample unless you have strong evidence confirming the representativeness of the sample
-A nonprobability sampling technique in which researchers divide the population into groups and then arbitrarily choose participants from each group
-like strata sampling & convenience sampling combined
a nonprobability sampling method in which elements are selected for a purpose, usually because of their unique position
-recruitment of participants based on word of mouth or referrals from other participants
-for hard to find groups of people
extreme case sampling
sampling of unusual or special participants who exhibit the phenomenon of interest in its extremes
The best samples are usually:
The best sampling method depends on:
-Your sampling frame
-Size of population
the ability of your study to reject a false null hypothesis
Is statistical power good or bad?
GOOD! You want statistical power
5 influences on statistical power
1. large effect size
2. large sample size
3. equal group sizes
4. high reliability
Warne's First Law of Behavioral Interventions
Subtle or brief manipulations usually have small effects on people's behaviors
large numbers of participants score near the max. score
large numbers of participants score near the min. score
What do manipulation checks detect?
measurement problems and weak manipulations
planning for high power studies
-anticipate your effect size
-ensure that you have large enough sample size
-reduce measurement error
-consider a design that has higher power
-collection of research studies on the same topic
-one study cannot establish scientific truth alone
-literature can vary in size
-requires gathering knowledge from many articles and combining that information into one whole
-this is not a summary of every article. Rather, it's an overall understanding of the literature
-A quantitative synthesis of the literature
-articles are selected systematically
-effect sizes are extracted from each article and then averaged.
-researchers can also search for moderators. A moderator can help a researcher understand why some studies produce stronger effects than others.
Problems with Literature Review
-Highly subjective: often people disagree about whether a study is
"strong"; susceptible to confirmation bias
-Not systematic: how studies are included is arbitrary; no regular way to decide which studies to emphasize
Benefits of meta-analysis
-easier to read than the entire literature
-removes much of the subjectivity from research synthesis
-all research gets included in the meta, without regards to its quality
-moderators can explain why results seem contradictory
Drawbacks of meta-analysis
-does not fix publication bias
-often studies do not report enough information for them to be used in meta
-different decisions made at the start of a meta can result in different final results
-questionable research practices can occur in meta-analysis as well
-hard to judge whether studies are really homogeneous enough to combine
Developments in meta-analysis
-it has become possible to estimate the level of publication bias through meta-analysis
-methods are not perfect but improving
-most methods require many studies
THIS SET IS OFTEN IN FOLDERS WITH...
Research Methods in Psychology EXAM2
Research Methods in Psychology EXAM 1
Research Methods of Psychology Exam 2
Research Methods in Psychology: Chapter 1
YOU MIGHT ALSO LIKE...
Research Methods chs. 5-8
Health Research chp 8, 9 vocab
DPT 6140 EBP Quiz 2
Definitions for Exam 2
OTHER SETS BY THIS CREATOR
EXAM 3: TESTS AND MEASUREMENTS
EXAM 4D: Depth perception
EXAM 4C: Object & color perception (cont.)
EXAM 4B: Object & Color Perception