Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
46 Cards in this Set
- Front
- Back
Purpose of Experimental Research
|
to control relationships among a set of variables in order to enable evaluation of causal relationships between specific variables of interest to the study.
|
|
Basic Issues of Experimental
|
In designing an experiment, decisions have to be made with respect to the four basic issues; Independent variable(s) to be manipulated, Dependent variable(s) to selected and measured, Extraneous variables to be controlled and Assignment of test units to the experimental treatments
|
|
Manipulation of the Independent Variable
|
Experimental Treatment, Experimental Group, several experimental treatment levels, several independent variables, and control group
|
|
Selection and measurement of the dependent variables
|
based on the purpose of the research, measurement and latency
|
|
Test Units
|
persons or entities whose responses are to be measured—individuals, teams, departments, school systems, supermarkets, airports, etc.
|
|
Sample Selection Error
|
bias introduced by inappropriate selection of test units—MWC business students may not be representative of total student population
|
|
Random Sampling Error
|
statistical fluctuation resulting from chance in selecting test units for the sample—unavoidable
|
|
Randomization
|
the assignment of subjects to treatments on the basis of chance. Generally considered the best procedure for controlling extraneous variables. Scatters or evens out the potential effects of extraneous variables
|
|
Matching
|
assigning test units to experimental treatments so groups are matched for extraneous variables that might distort results such as age, gender, education, income, etc.
|
|
Repeated Measures
|
subjecting the same subjects to all experimental treatments. Eliminates error due to subject differences but may introduce other problems such as “practice” or “history” effects.
|
|
Constant error
|
term usually used for “systematic error” when discussing experiments—error attributable to flaw in design or execution of the experiment.
|
|
Extraneous variables
|
variables that, uncontrolled, distort results in a particular direction every time an experiment is repeated and which mask the true state of affairs
o Example – consistently administering one experimental treatment in the morning and the other one in the afternoon |
|
Demand characteristics
|
experimental procedures that hint to subjects the nature of the hypothesis and suggest—“demand”—that they respond in a particular way.
o Experimenter bias – a constant error caused by the experimenter’s presence, actions, or attitudes—appearing eager, authoritarian, etc. o Guinea pig effect – a constant error caused by subjects behaving abnormally in order to “cooperate” with or “please” the experimenter. o Hawthorne effect – a constant error caused by subjects’ knowing they are participating in “research”—a kind of status/motivational effect |
|
Constancy of conditions
|
when extraneous variables can’t be eliminated, subjects in various treatments are exposed to identical conditions except for the treatment itself—assumes the effects of extraneous factors are spread evenly across all treatment conditions
|
|
Order of presentation bias
|
error caused by subjects’ accumulating experience or savvy in the course of responding to multiple experimental treatments.
|
|
Counterbalancing
|
technique to reduce error caused by order of presentation—varying the order of experimental treatments for different experimental groups
|
|
Blinding
|
technique to control subjects’ knowledge of whether or not they have been given a particular experimental treatment.
|
|
Double-blind design
|
neither the subjects nor the experimenter know which are the experimental and which are the control conditions.
|
|
Random assignment
|
if extraneous factors can’t be controlled, it is assumed that their effects will be equally present in all experimental conditions through random assignment.
|
|
Debriefing
|
giving subjects all pertinent facts about the nature and purpose of the experiment after its completion.
|
|
Basic Design
|
a single independent variable is manipulated to observe its effect on single dependent variable
|
|
Factorial Designs
|
two or more independent variables are manipulated simultaneously to measure both main effects and the effect of their interaction on a dependent variable
|
|
Laboratory Experiment
|
an experiment conducted in an artificial setting in which the experimenter has almost complete control mover the research setting
|
|
Tachistoscope
|
a device that controls the amount of time a subject is exposed to a visual image.
|
|
Field Experiment
|
an experiment conducted in a natural setting—often for a long period of time—in which the experimenter has less control over extraneous variables.
|
|
Internal Validity
|
whether an experimental treatment was the sole cause of change in a dependent variable. Six common extraneous variables raise issues of internal validity ; Maturation Effect, instrumentaion effect, selection effect, testing effect, history effect, mortality effect
|
|
Maturation Effect
|
caused by changes in the experimental subjects over time such as boredom, aging, attitude.
|
|
Instrumentation Effect
|
a change in the wording of a questionnaire, a change in interviewers, or a change in procedures used to measure the dependent variable
|
|
Selection Effect
|
sample bias resulting in differential selection of subjects for the various experimental treatments—such as age, gender, or educational level.
|
|
Testing Effect
|
in a before- and after-study, an effect of pretesting that sensitizes subjects when tested for the second time
|
|
History Effect
|
a specific event in the external environment occurring between the first and second measurements of the experiment—beyond control of the experimenter—that affects the validity of the experiment
|
|
Mortality Effect
|
sample attrition that occurs when some subjects withdraw from an experiment before it is completed
|
|
External Validity
|
the ability to generalize the results of an experiment to similar conditions outside the experiment itself—transferring the results to the real world
|
|
The Trade-off
|
the choice between a laboratory experiment and a field study always involves a trade-off. A field study may achieve higher external validity at the expense of internal validity—meaning the results may be specific to the real-world field situation only. Likewise, the results of a laboratory experiment may not be generalizable to the real world.
|
|
One Shot Design
|
Flawed quasi experiment because it has no comparison group and no control over extraneous variables (exp. offering a prize to boost sales)
|
|
One-Group Pretest-Posttest Design
|
Flawed quasi experiment. SOUrces of error include maturation effect, history effect, mortality effect, instrumentation effect
|
|
Static Group Design
|
Flawed Quasi experiment. • Major weakness is that there may be systematic differences in the groups selected, e.g., volunteers used for the experimental treatment
|
|
Pretest-Posttest Control Group Design
|
It is assumed that the effects any extraneous variables are balanced by the randomization, but a testing effect is still possible if subjects are sensitized by the test questions to the purpose of the research
|
|
Posttest Only Control Group Design
|
Used when: (1) pretest not possible, or (2) groups are known to be equal. Eliminates testing effect and assumes extraneous variables operate equally on both groups.
|
|
Solomon Four-Group Design
|
Separates out maturation effect, testing effect, and treatment effect and is rarely used in business because of effort, time, and cost
|
|
Compromise Experimental Design
|
sometimes random assignment is not possible, e.g., when administering treatment to departments within an organization. In such cases compromise designs are unavoidable, e.g., assigning units as a whole to experimental treatments.
|
|
Time Series Design
|
repeated observations are taken over an extended period of time—e.g., as in political polls—enables researcher to evaluate: The immediate effect of an event or experimental treatment
The permanence of the effect. Time series designs are subject to, History effect and Loss of control (therefore are quasi-experimental designs) |
|
Completely Randomized Design
|
randomly assigns experimental units to treatments to investigate the effects of a single independent variable
|
|
Randomized Block Design
|
an extension of the completely randomized design in which each experimental treatment is administered to each value (block) of a particular extraneous variable.
|
|
Factorial Design
|
allow for testing the effects of two or more treatments (factors) at various levels
|
|
Latin Square Design
|
attempts to control or block out the effect of two or more confounding extraneous factors. The major assumption—and drawback—of the Latin Square design is that interaction effects are expected to be minimal or nonexistent. A Latin Square design can have any number of treatments
|