Experimental Research: Chapter 7 Flashcards
high constraint
- most powerful of the research methods
main goal of research
establishing a causal explanation
manipulation
manipulation of one variable (IV) to observe its effect on another variable (DV) while holding other potential influences constant
2 basic types of research
- between groups
- within groups
between groups design
- 2+ groups are formed at random from a
pool of subjects (independent groups) - each group receives a different experimental treatment (value of the IV) and groups are compared
within-subjects design
- only one treatment group, and each subject is
given all values of the IV - comparison is made between scores obtained at
different levels of the IV for same subjects
4 basic elements to research
- manipulation
- measurement
- comparison
- control
2 critical elements to research
- independent and dependent variables
- experimental and control groups
independent variable
- is manipulated by the experimenter
- can be thought of as the cause
- typically at least two levels of the IV
- each level corresponds to a treatment condition
dependent variable
- variable in which you are trying to effect some sort of change when you are manipulating
the independent variable - DV is measured for each level of the IV and compared across conditions
types of groups
- experimental groups
- control groups
experimental groups
participants are exposed to the manipulation
control groups
participants not exposed to the manipulation and that is used for comparison purposes
extraneous variables
- “extra” variables present in the study but are not studied
- don’t matter because they don’t affect the
outcome. - ex: age, gender, ethnicity
categories of extraneous variables
- environmental variables
- participant variables
- time-related variables
confounding variables
- type of extraneous variable
- acts as IV and can mask the true effect of the
IV under investigation - vary systematically along with IV
- affect the outcome.
- ex: music, math performance, researcher
- threat to internal validity – don’t know if it is IV
or CV that is causing the change in DV - introduces ambiguity
extraneous vs confounding
extraneous
- if it affects all conditions equally
confounding
- if only one condition is affected
- because the effect of the noise becomes confused with the effect of the IV
within-group design: confounding variables
- environmental conditions
- history
- instrumentation
- maturation
- repeated testing
- regression to the mean
between groups design: confounding variables
- assignment bias
- attrition
- compensation
- resentful demoralization
- environmental conditions
within: history
- external events occurring between 1st and 2nd tests
- longer the time interval between a pretest & a
posttest measurement, the greater the possibility that outside events will influence a particular treatment outcome
within: maturation
- processes within oneself produce changes in a
subject over time - not related to treatment - includes any systematic changes in biological or
psychological condition over time - (physical growth, cognitive development, wisdom, boredom, etc.)
within: repeated testing
- possible effects of the pretest on the posttest
- progress on test may be caused by experience
Solomon’s four group design
refer to slide 26
within: instrumentation
- changes in the characteristics of a measurement
instrument over time - ruler does not change over time under controlled
conditions - if measuring instrument is a human
observer, person may become more skilled over time
within: regression of the mean
- when participants have extreme scores on a
pretest measure, their scores will be more likely to
change on a subsequent measure than will scores that are closer to the mean - high scores = more likely to go down
- low scores are more likely to go up
- how much regression will occur depends on how much performance on the pretest was due to variable factors
between: assignment bias
there is a systematic difference between comparison groups before the administration of a treatment or a manipulation
how is bias introduced
when subjects have been formed into groups for reasons other than study participation
Teachers who volunteered to adopt a new teaching program vs. those who did not (control group)
Open-ward patients receiving supportive therapy vs. patients in closed-ward not receiving it (control group)
are both examples of _________
assignment bias
attrition
loss of participants
reasons for attrition
- vacation
- no longer interested
- too busy
- forgetting appointments
- sick
- etc.
differential attrition
- differences in the number of drop-outs or refusals
that affects one group more than another - varies with treatment group
- cause for dropping out is the study itself
between: differential attrition
- subjects with certain characteristics are lost
- it is important to note WHO drops and WHEN
differential attrition: reasons for dropping out of the study
- gender bias
- too difficult or demanding treatment
- subjects feel worse
- treatment may be inappropriate for some
why is it important to know about differential attrition
- can bring positive feedback to experiment
- can help improve the study
between: compensation
- compensatory equalization
- compensatory rivalry
compensatory equalization
- untreated individuals or groups learn of treatment received by others
- demand the same treatment or something equivalent
- example: Large-scale education enrichment program given in certain schools but not others. Parents and teachers in other schools asking for equivalence
compensatory rivalry
- untreated group learns of the treatment received by others
- untreated group works extra hard to see to it that the expected superiority of the treatment group is not demonstrated
between: resentful demoralization
- individuals in untreated or control group learn others are receiving special treatment
- creates false appearance of an advantage for the treated group (Type I error)
- reality is: the treatment itself actually has no beneficial effect
effects of resentful demoralization on participants
- less productive
- less efficient
- less motivated
controlling for confounding variables
- Remove them
- Hold them constant
- Use placebo control
- Match them
- Randomize them
removing the confounding variable
- remove the researcher
holding the confound constant
- if we can’t remove the confound
- can try holding the confound constant across situations and conditions
using a placebo control
- experimental method can be a confound itself
- example: Stress due to an injection is subtracted from total treatment effect by including an injection in the control group
matching across conditions
refer to powepoint on experimental research, slide 39
randomizing the confounds
- randomly assign participants to treatment conditions
- extraneous variables will balance out
- aim = disrupt systematic relationship between extraneous and independent variables
- done to prevent EV from becoming CV
- powerful method for controlling environmental and participant variables at the same time
is full control possible in an experiment?
no
external validity: categories of generalizations
- from the sample to the general population
- from one research study to another
- from research to real world
threats to external validity (from the sample to the general population)
- selection bias
- college students
- volunteer bias
- participant characteristics
- cross-species generalizations
threats to external validity (from one study to another)
- experimenter characteristics
- type of measures used
- time of measurement
threats to external validity (from study to real world)
- simulation vs. artificiality
- mundane realism
- experimental realism
research artifact
- non-natural feature accidentally introduced in the study
- general threat to internal and external validity
- can involved experimenter bias, participant reactivity
experimenter bias
inadvertent tendency to influence results in
expected direction
solution to experimenter bias
double-blind technique
participant reactivity
participants modify their natural behavior in
response to being in a study
* good subject role
* negativistic subject role
* apprehensive subject role
solution to participant reactivity
- double-blind
- deception
factorial design
2+ IV in a study
multivariate study
2+ DV in a study