Value Choices in the Mass Public: Testing for Framing Effects and Hierarchical Structure

Download data and study materials from OSF

Principal investigator:

William G. Jacoby

Michigan State University



Sample size: 649

Field period: 06/16/2005-06/21/2005


Longstanding psychological theories have stressed that individual value preferences are very stable, highly structured and broadly applicable to virtually all situational contexts. But, recent research suggests that value choices may be relatively malleable and susceptible to contextual effects. Furthermore, the apparent existence of individual-level value conflict raises questions about citizens� abilities to differentiate among values in the first place. This study uses some new internet survey items in order to address the issues raised by these potentially conflicting perspectives. The empirical results show that most people do make consistent pairwise choices among values. Furthermore, most individuals� sets of pairwise choices are fully transitive, meaning that they can be assembled into hierarchical structures. Finally, value choices are largely impervious to priming effects.


1.Individuals make reliable pairwise choices among values. That is, a person's revealed preference across a pair of values will remain consistent across replicated choices between those values.
2.Most individuals' sets of pairwise value choices are fully transitive. That is, each person's separate pairwise choices can be assembled into a fully-ranked preference ordering for that person.
3.Individual value choices should not vary across situational contexts. That is, a person should always select one value as more important than another, even in situations where the less-important value is given greater salience or emphasis.

Experimental Manipulations

There are eight different experimental conditions. In five of the conditions, survey respondents answered two questions intended to prime one of five specific values (liberty, equality, morality, economic security, and social order). Two more conditions primed a conflict between a pair of values (liberty versus social order, and liberty versus equality). The eighth condition was a control, in which no values were primed; respondents were asked about partisanship, instead.


1.Replicated pairwise choices among every possible pair of five values (liberty, equality, morality, economic security, and social order).
2.Dominant pairwise choices (i.e., the value that is selected two or three times on the three replications of that pair) are assembled into full preference orders. If a respondent�s pairwise choices are transitive, then preferences can be fully ordered, from the value that is considered most important, down to the one that is considered least important. If intransitivities occur among the dominant pairwise choices, then ties will occur in the preference ranking of the values.

Summary of Results

For each value pair, between 70% and 80% of the respondents made consistent choices across the three replications. This is far more than would be expected by chance, suggesting that value choices are very reliable. Individual value choices are highly variable. But, at the aggregate level, morality, economic security and liberty are considered more important than equality and social order. Most individuals� choices can be fully rank ordered, with only about 12% of the sample exhibiting any intransitivity. And, priming a particular value, or a conflict between two values, does not have any impact on individual choices.


Taken together, the empirical results from this study support the traditional view from social psychology that value preferences are hierarchically structured and largely impervious to external influences.

Additional Information

The data on the dependent variable were collected using the method of triads. In other words, all possible subsets of three values (from the full set of five values) were presented to the respondents. For each such "triad" of values, the respondents were instructed to indicate the one that they thought was most important, and the one that they thought was least important. These selections can be broken down into pairwise choices among the three values in that triad. Across the full set of triads, each value pair occurs three times. So, there are three replications for each pairwise value choice.


Jacoby, William G. (2006) "Testing for Hierarchical Structure and Priming Effects among Individual Value Choices: A Preliminary Report on a Study Supported by the Program for Time-Sharing Experiments in the Social Sciences." Unpublished paper provided as an attachment to this questionnaire, and also available on the web at