Unit 5: Non-Experimental Research: Observational, Archival, and Case-Study Research Flashcards
non-experimental research
When the researcher has less than complete control over the independent variable, or cannot randomly assign subjects to groups. Often qualitative.
correlational research
Another name for non-experimental research. Describes correlation between variables, but can not tell whether one causes the other. Establishes that there is a relationship, but can not tell what it is. (Can indicate possible causes - good for future research ideas).
naturalistic observation
Observing subjects in their natural environment without intervening.
participant observation
Researcher becomes a part of a group in order to make their observations. Deceit is often used = ethical issues around consent. Often for small, closed groups.
inter-observer reliability
To what degree do two observers’ observations agree. High agreement increases the validity of the findings.
role demand
When a subject changes their behaviour to what they think the researcher wants to see.
observer bias
Biases of the researcher that influence the observations they make. Leads to ‘systematic errors’. Best avoided by keeping the researcher blind.
unobtrusive measures
No attempt to intervene or influence the situation
reactive measures
subjects know there is a researcher and change their behaviour in response
physical traces
physical indicators of past behaviour. Ex. objects left lying around, wear and tear.
archival data
data obtained from records or documents kept by individuals/institutions that the researcher had no part in collecting.
content analysis
analysis of textual or graphic records (ex. newspapers)
manifest content
In content analysis, counting the occurrences of some objective measure
latent content
In content analysis, recording the themes as interpreted by the researcher
case study
an intensive study and analysis of an individual subject, case, event, or occurrence of a phenomenon.
What distinguishes non-experimental from experimental research?
In experimental, the researcher manipulates an independent variable and measures the effect on the independent variable, to describe and understand the cause-effect relationship between the two.
In non-experimental, the researcher does not have control over the independent variable, so can only establish correlation not causation, or has to give up some control in order to access the data (ex. archival).
Why is non-experimental research often called correlational research?
Because the researcher does not have control over manipulating the independent variable, cause and effect relationships can not be determined. Only correlational effects can be observed, in which other variable may be influencing the dependent variable.
What are several common varieties of non-experimental research.
- Observational (naturalistic observation, participant observation, physical traces)
- archival
- case study
- surveys
How is naturalistic observation different from the casual observation that most people do?
Has the intention of describing correlations in behaviour, but is non-intrusive, non-reactive. Participants do not know they are being observed, so they don’t respond to the presence of the researcher.
What are three common ways of sampling behaviours in observational research.
- time sampling
- event sampling
- situation sampling
What is time sampling?
A sampling technique in observational research.
Selecting time intervals at which points to make sample observations of behaviour
a) Systematic
b) Random
c) Random within systematic
What is event sampling?
A sampling technique in observational research.
Record all instances of a particular kind of infrequent event.
What is situation sampling?
A sampling technique in observational research.
Observing a selected behaviour in as many situations and circumstances as possible.
When is physical trace research useful?
Physical trace research is useful when you want to conduct observational research without subjects reacting the the researcher’s presence.
The activity has already happened - no confounding.
