Module 5-8 Flashcards
ecological validity
external validity
naturalistic observation
direct data collection in the wild withour intervention
participant observer
hidden or disgused or undisguised person interact with the community and particpants
structured observations
participants repsond to situtation formed by researchers
field experiment
carried out in natural settign rather then a lab
ex: dressing homeless and seeing ppl react
structured observation
weak internal, strong external
population validity
the extent we can generalize our finding to the public
systematic sampling
selecting every nth person
Event sampling
Event sampling, also called frequency counts, involves observation of targeted behaviours or specific events
Non probability sampling
finding ppl to survey that are easy to reach
snowball sampling
particpants recruit other ppl
stratified sampling
divides group in non overlaping groups
categorical variables+ ex
Variable representing distinct, non-overlapping categories or groups of individuals, which are distinguished by a specific characteristic
Ex: These can be simple binary categories like whether participants drink alcohol (1 = No, 2 = Yes) or more complex variables like drinking habits (1 = Never, 2 = Rarely, 3 = Frequently).
Nominal variable
most basic type of variable and simply name categories. There is no numeric or ordered relationship between the categories or values, and as such, the order in which they are presented and how they are numbered is based on convenience.
Ordinal Variable
also involve naming categories, although these variables have an ordered relationship. These would be variables like income level (low, medium, or high) or medals at the Olympics (bronze, silver, gold). A category of an ordinal variable can be ranked higher or lower than another, but this does tell us anything about the numerical difference between each category.
Interval Variable (2)
- Variable measured by order with equal distances between each interval on the scale.
- do not have a true zero
Ratio Variable (2)
- represent real numbers with consistent intervals between values. However, numbers used to describe ratio variables also have a real zero
- can perform multiplication and division
Opinion and belief data have several limitations
reliability, scocial desirability
social desirability
A kind of response bias wherein study participants maximize what they believe are “good” responses and minimize what they believe to be “bad” or undesirable responses.
Behaviour data+ex+limitations
The behavioural data we collect should reflect an individual’s actions or activities. If we are interested in substance use, we may ask an individual how many times they have drunk alcohol in the last two weeks, and how many drinks they had each time on average. The goal is to develop measures of the activities of our participants.
- limitations: memory+ behavior does not always imply intention
Biological data
Consists of biological markers that can be used in research. These can include bodily fluids such as blood, recordings of bodily reactions such as skin conductance (electrical reaction of skin) or electrocardiogram (ECG- a type of heart monitor), or scans of the body such as with functional magnetic resonance imaging (fMRI- a scan of brain activity).
Primary Data
Data collected by the researchers themselves rather than obtained from others
Inter-rater reliability
The extent to which raters agree and provide consistent estimates of the same behavior.
Surveys are typically designed to collect information in three main areas: (3)
- demographic
- dependent variable as well as the independent variable (IV)
- Confounds
Secondary data
Data collected by another person for another purpose that is used in new research.
Replication Crisis
Ongoing discussion in the field about the challenges of replicating findings using similar techniques but different groups of people.
Low dispersion is problematic because
it makes it more difficult to compare the difference between groups when they are of very different sizes.
For quantitative variables, both interval and ratio, we are able to use a statistic called
Standard deviation
Nominal data strongest measure of Central tendency)
Mode
Ordinal data strongest measure of Central tendency)
Median
Interval/ratio data strongest measure of Central tendency)
mean
Methods for identifying dispersion for ordinal/nominal
graph/chart
Methods for identifying dispersion for interval/ratio
Standard deviation
Web scraping techniques
Data collection techniques that download feeds from social media sites like Facebook and Twitter.
evaluation
a process of forming judgments and making decisions based on a variety of measurements
Scientific bias
any systematic differences between the results of a study and the “truth,” and is assessed by examining test data.
two types of test bias:
content bias and testing situation bias
content bias
occurs if test items favor one group of individuals over another group
Item response theory
models the probability that a person with a given level of ability will correctly answer an item with certain characteristics, such as the spatial nature of this item.
computerized adaptive testing
allows for adjusting item difficulty to the ability level of an individual.
Distinctiveness
distinct numeric values may be assigned to individuals who belong to different classes of a characteristic being measured.
ordering in magnitude
assigned numeric values represent the degree of measured characteristics, with lower values signifying a lower degree of a characteristic and higher values signifying a higher degree of that characteristic.
classical test theory
postulates that every measurement consists of two elements: the true score and a measurement error
measurement reliability
the extent to which a measurement is repeatable and free of measurement error
construct validity:
The degree of overlap between a theoretical construct and the measurement instrument.
content validity:
Extent to which a measure represents all features of a given construct.
criterion validity:
A relation between a measure intended to assess a certain construct and other measures that assess the same construct.
comprehensive narrative:
Recording as much data as possible.
participant observation
Disguised or undisguised observer interacts with a community of interest.
structured observations:
Participants respond to a situation framed by the researcher.
systematic sampling
Selecting every nth person, behavior, or situation to observe.
- researchers select members of the population at a regular interval
