A Quarterly Review
of Relevant Research
From Other Disciplines
I-O Outside I-O
Mark Alan Smith
Society for Human Resource Management
If you been under the most all-encompassing rock for the last 2 years, you might not know that politics play a large role in the national discourse around every topic. No matter one’s leanings it is clear that much investment is made in trying to predict one act of national-level behavior and the motivations behind this behavior. Specifically, we are talking about voting and the motivations for voting: both why people vote and why they select a candidate on the ballot.
Political science/political psychology is the topic of the two focal articles. The first article looks at the name order effect on ballots (the primacy effect seems to really exist) and potential moderators of it. This effect can have misleading influences on voters and outcome of votes, as well as surveys and other tools that I-O psychologists typically use. The second article looks at general motivations for voting behavior and how moral foundation theory may provide insights into other kinds of behavior not previously associated with morality.
Moderators of Candidate Name Order Effects in Elections: An Experiment. Kim, Nuri; Krosnick, Jon; Casasanto, Daniel, Political Psychology, 36(5), 2015, 525-542. doi: 10.1111/pops.12178
In this recent article, the researchers studied the effects that the order of appearance on a ballot can have on the outcome of elections. The effects of name order have been known for a number of years: candidates often receive more votes when their names are listed first then when their names are listed after the names of their competitors. In fact, studies of real world elections in the US have estimated that the advantage of being listed in first position on the ballot (the primacy effect) is in the range of 2% to 3%.
This focal study sought to extend this finding by focusing in the moderators of this primacy effect. In particular, these researchers looked at some potential moderators, including:
- Amount of information provided
- Ambivalence: i.e., how conflicted the participants felt when making their votes
- Cognitive skills: education was used as a proxy variable
- Cognitive effort: i.e., how hard the participants tried during the candidate evaluation period
- Handedness: right handers v. left handers
Participants from a national sample of Americans in 2009 were used. Of the 1.4 million in the national survey panel, 83,986 were selected for participation. 2,069 chose to participate in the broader survey (2.5% participation rate), and 572 completed the questions for the study, which comprised the study sample.
Participants completed computer-based survey items which asked about two hypothetical candidates for the US Congress: “Alan Mitchell” and “Robert Swanson.” They were shown varying levels of information about the candidates, and then participants “voted” for one of them. Additional questions gauged the information needed for the moderators.
Results showed that there was an overall effect of being listed first on the ballot of 15.3%. Strong moderation was shown for amount of information; the primacy effect when little information about the candidates was presented was 28.3%, which the primacy effect when more information was presented was close to zero (1.7%).
Primacy was also much larger among participants who were more ambivalent about the candidate choices. Also, more educated individuals tended to have much less of a primacy effect, and individuals who reported putting in more cognitive effort had a smaller primacy effect.
Horizontal display of candidate choices led to an increase in the primacy when compared to a vertical display. Interestingly, right handers v. left handers also had a moderating effect on the primacy effect, but only when candidate names were displayed horizontally rather than vertically. Left handers showed a greater tendency toward the primacy effect than right handers when comparing horizontally.
Thoughts From an I-O Perspective
In our opinion, there are a couple of main takeaways from this research for our field. The first involves the design and mechanics of surveys/ballots. Even without considering the moderators of the primacy effect, this study points to the conclusion that surveys should be designed to rotate the presentation of alternative choices. Although this would have been difficult when asking questions via paper surveys, such a rotation is relatively easy for many online survey platforms. Also, it is best to display choices vertically, rather than horizontally in order to help diminish the primacy effect.
Further, we should consider finding ways to reduce the primacy effects by focusing on some of the moderators in this study. Lack of information, ambivalence, lack of cognitive skills, and lack of cognitive effort all appear to lead to an increase in the primacy effect. Although we generally we cognitive skill as a fixed trait and ambivalence might be hard to change, we should be sure to provide enough information to survey participants in a way that is engaging.
As interesting as these findings might be, it is important to note that the effects of ballot presentation in real world conditions are rather small, just few percentage points. These results must be understood with these rather minute effect sizes in mind. After all, how often do important elections turn on just a few percent of the vote?
Franks, A. & Scherr, K. (2015). Using Moral Foundations to Predict Voting Behavior: Regression Models From the 2012 U.S. Presidential Election. Analyses of Social Issues and Public Policy, 15(1), 213-232. doi: 10.1111/asap.12074
Debates about the value of predictive analytics rage beyond the field of I-O psychology, and there is no more salient example of this than trying to predict voting behavior. Moreover, this debate proves itself valuable for other social sciences largely because it provides large-sample heuristic data with the ability to assess temporal changes and swinging motivations over one act. It provides insights into factors associated with decision-making processes, looks at core traits versus states held over extended periods, and examines the impact of antecedent components that lead to conditional waiving of held values. In our world, it is tantamount to looking at Project A data to examine noncognitive factors associated with choosing an employer all the while using a data set one million times the size.
Method and Findings
In this series of three studies, Franks and Scherr explore the concept of “moral foundations” as a predictor of voting behavior as it relates to the 2012 presidential election. Of particular interest to I-Os are the treatment of moral foundations as mutable states versus intransigent traits. The literature behind moral foundation theory has consistently treated these as traits that sway when context is added to a given situation. Further, it has been long hypothesized that these moral foundations are potentially hereditable.
In this research, participants responded to scales from multiple constructs ranging from political leanings to social attitudes about moral situations to basic demographic information. In the first study, researchers used moral foundations theory to identify potential underpinnings of a relationship between morality and behavior using a sample of 144 voters. In the second study, a representative sample from the American National Election Survey (ANES; n = 1107) was used to identify a relationship between key factors like type of moral foundation and voting selection via a hierarchical logistic regression technique. In the final study, a sample of 200 participants completed measures aimed at linking moral foundations of the individual with subjective assessments of moral foundations of others.
Results from these three studies were used to examine the changing nature of moral foundations and also to identify a two-dimensional structure of moral foundations—individualizing foundations versus binding foundations. These foundations ended up predicting actual voting behaviors in the Presidential election above and beyond demographic variables. Upon further investigation, these concepts sound an awful lot like individualistic versus collectivistic orientations—something all I-Os know a lot about.
Thoughts From an I-O Perspective
So how does this research affect the way I-Os should operate? Well, three things are abundantly clear:
- Public policy research could use a refresher on social psychology and applied psychology (ask anyone who voted in the 2016 election about cognitive dissonance).
- Franks and Scherr provide a novel way of reframing state versus trait debates but looking a context from a hereditary perspective, something we in I-O have seemingly ignored.
- The researchers have also provided a series of implications for applying moral foundations to the prediction of counterproductive or self-destructive behaviors.
This last one has the most fertile ground for application in I-O as they unwittingly lay out a new way for predicting one-time bad behaviors in the workplace.