Kardes, F.R. In Defense of Experimental Consumer Psychology.
The criticisms towards experiments are
a. experiments rely too heavily on artificial settings,
b. experiments rely too heavily on college student research participants,
c. experiments do not provide practical solutions to applied problems.
DIFFICULTIES ENCOUNTERED IN NATURAL RESEARCH SETINGS
Scientists deliberately construct artificial settings and conditions because this is an effective way
to isolate the effects of specific causal variables in a large, tangled web of probabilistically
related causes and effects.
Necessary and sufficient causes are quite rare in the behavioral sciences (unlike in the physical
sciences), and most robust behavioral phenomena are multiply determined.
Consequently, it is nearly impossible to learn much about specific cause-effect relations without the
benefit of a controlled artificial research setting.
From the elaboration likelihood model of persuasion:
Research participants are randomly assigned to strong or weak argument conditions.
The weak argument condition is a highly unnatural and artificial condition because advertisers do not
deliberately attempt to create advertising copy containing weak, uncompelling arguments.
However, comparing the attitudes of participants in strong versus weak argument conditions provides
researchers with a useful diagnostic tool for determining when people are likely to follow the effortful
central route to persuasion and when they are likely to follow the less effortful peripheral route.
The distinction between these two routes to persuasion is important because attitudes formed via the
central (vs. peripheral) route are more predictive of subsequent behavior and are less readily changed by
future communications.
The two routes to persuasion and their consequences would have been extremely difficult to discern
without the use of highly artificial and controlled experiments. Comparison, manipulation, and control
are necessary for learning about complex systems of interrelated causes and effects.
Intuitive Physics
People observe falling objects and objects in motion throughout their lives. Nevertheless, most people's
beliefs about the physics of motion are rather medieval.
Illusory Correlation
When attempting to assess the degree of relation between two variables observed in a natural setting,
people frequently use prior beliefs or expectations to guide information search, construal, and
integration processes.
Subjective expectation-based (or theory-based) correlation judgments are consistently greater than
objective data-based correlations primarily because people often prefer to rely on subjective beliefs
, rather than on objective data when forming judgments of correlation. This insensitivity to objective data
makes it very difficult to learn from experience.
Hindsight Bias (уклонение от ретроспективной оценки)
Another judgmental phenomenon that makes it difficult to learn from experience is the hindsight bias,
or the "I-knew-it-all-along" effect
When people believe that they knew all along that a particular event would occur, they greatly
exaggerate the validity of their prior beliefs and mаke no attempt to update their knowledge.
Because of the hindsight bias, decision makers tend to be evaluated on the basis of the quality of their
decision outcomes rather than on the quality of their decision processes.
When unforeseeable (непредвиденные) chance factors lead good decisions to turn our poorly, good
decision makers often are reprimanded or fired. After the fact, unforeseeable chance factors seem
foreseeable, and those who failed to foresee these factors are often unfairly punished.
Change-of-Standard Effect
Standards or reference points for judgment frequently change over time in complex natural settings.
The failure to make appropriate judgmental allowances for changing standards creates illusions of
stability and constancy that do not exist in nature. These illusions make it very difficult to learn
systematically from unstructured natural experience.
Me change-of-standard effect encourages people to use successful interpersonal, medical, and
managerial strategies repeatedly, even when the task characteristics have changed in important ways.
When task characteristics change so that previously successful strategies are no longer effective,
persistence leads to the "success breeds failure" paradox.
Selective Hypothesis Testing
The manner in which people generate and test hypotheses can also make learning from experience
difficult. Hypotheses refer broadly to alternative judgments, interpretations, categorizations, options, or
possibilities.
Scientific hypothesis testing involves generating as complete a list of alternative hypotheses as
possible and then performing a systematic evaluation of the merits of each.
Everyday hypothesis testing, however, is often highly selective: People often consider a single
focal hypothesis or a very limited set of hypotheses
However, because even weak or ambiguous evidence is often construed as supportive--due to biased
assimilation, discounting of inconsistent information, or the failure to consider the extent to which the
data support alternative hypotheses 1979tthe search for additional evidence often terminates
prematurely. Consequently, the validity of a focal hypothesis considered in isolation is often
overestimated.
Moreover, selective hypothesis testing appears to contribute to illusory correlation, hindsight bias, the
change-of-standard effect.
The criticisms towards experiments are
a. experiments rely too heavily on artificial settings,
b. experiments rely too heavily on college student research participants,
c. experiments do not provide practical solutions to applied problems.
DIFFICULTIES ENCOUNTERED IN NATURAL RESEARCH SETINGS
Scientists deliberately construct artificial settings and conditions because this is an effective way
to isolate the effects of specific causal variables in a large, tangled web of probabilistically
related causes and effects.
Necessary and sufficient causes are quite rare in the behavioral sciences (unlike in the physical
sciences), and most robust behavioral phenomena are multiply determined.
Consequently, it is nearly impossible to learn much about specific cause-effect relations without the
benefit of a controlled artificial research setting.
From the elaboration likelihood model of persuasion:
Research participants are randomly assigned to strong or weak argument conditions.
The weak argument condition is a highly unnatural and artificial condition because advertisers do not
deliberately attempt to create advertising copy containing weak, uncompelling arguments.
However, comparing the attitudes of participants in strong versus weak argument conditions provides
researchers with a useful diagnostic tool for determining when people are likely to follow the effortful
central route to persuasion and when they are likely to follow the less effortful peripheral route.
The distinction between these two routes to persuasion is important because attitudes formed via the
central (vs. peripheral) route are more predictive of subsequent behavior and are less readily changed by
future communications.
The two routes to persuasion and their consequences would have been extremely difficult to discern
without the use of highly artificial and controlled experiments. Comparison, manipulation, and control
are necessary for learning about complex systems of interrelated causes and effects.
Intuitive Physics
People observe falling objects and objects in motion throughout their lives. Nevertheless, most people's
beliefs about the physics of motion are rather medieval.
Illusory Correlation
When attempting to assess the degree of relation between two variables observed in a natural setting,
people frequently use prior beliefs or expectations to guide information search, construal, and
integration processes.
Subjective expectation-based (or theory-based) correlation judgments are consistently greater than
objective data-based correlations primarily because people often prefer to rely on subjective beliefs
, rather than on objective data when forming judgments of correlation. This insensitivity to objective data
makes it very difficult to learn from experience.
Hindsight Bias (уклонение от ретроспективной оценки)
Another judgmental phenomenon that makes it difficult to learn from experience is the hindsight bias,
or the "I-knew-it-all-along" effect
When people believe that they knew all along that a particular event would occur, they greatly
exaggerate the validity of their prior beliefs and mаke no attempt to update their knowledge.
Because of the hindsight bias, decision makers tend to be evaluated on the basis of the quality of their
decision outcomes rather than on the quality of their decision processes.
When unforeseeable (непредвиденные) chance factors lead good decisions to turn our poorly, good
decision makers often are reprimanded or fired. After the fact, unforeseeable chance factors seem
foreseeable, and those who failed to foresee these factors are often unfairly punished.
Change-of-Standard Effect
Standards or reference points for judgment frequently change over time in complex natural settings.
The failure to make appropriate judgmental allowances for changing standards creates illusions of
stability and constancy that do not exist in nature. These illusions make it very difficult to learn
systematically from unstructured natural experience.
Me change-of-standard effect encourages people to use successful interpersonal, medical, and
managerial strategies repeatedly, even when the task characteristics have changed in important ways.
When task characteristics change so that previously successful strategies are no longer effective,
persistence leads to the "success breeds failure" paradox.
Selective Hypothesis Testing
The manner in which people generate and test hypotheses can also make learning from experience
difficult. Hypotheses refer broadly to alternative judgments, interpretations, categorizations, options, or
possibilities.
Scientific hypothesis testing involves generating as complete a list of alternative hypotheses as
possible and then performing a systematic evaluation of the merits of each.
Everyday hypothesis testing, however, is often highly selective: People often consider a single
focal hypothesis or a very limited set of hypotheses
However, because even weak or ambiguous evidence is often construed as supportive--due to biased
assimilation, discounting of inconsistent information, or the failure to consider the extent to which the
data support alternative hypotheses 1979tthe search for additional evidence often terminates
prematurely. Consequently, the validity of a focal hypothesis considered in isolation is often
overestimated.
Moreover, selective hypothesis testing appears to contribute to illusory correlation, hindsight bias, the
change-of-standard effect.
