19 Negative Contact: Opposites, Additives, or Interactives?
Positive and Negative Contact: Related or Separate Constructs
Indeed, this positivity bias highlights a glaring omission in contact theory. It is naive to think that individuals only experience positive intergroup contact. Moreover, the way positive contact has been measured over the past 60 years has placed positive and negative contact at opposite ends of the same spectrum. For instance, a typical survey item used to measure contact quality looks as follows:
To what extent did you experience the contact with [outgroup] as competitive or cooperative (1 = very competitive; 7 = very cooperative; see Islam & Hewstone, 1993; Lolliot et al., 2015)
Placing positive and negative contact at opposite ends of the same spectrum seems short-sighted. Just as a thought exercise, it is possible to have both high levels of positive contact AND negative contact with people. If you think to experiences with a best friend or a sibling, you are likely able to recall happy memories as well as memories where you were fighting.
We can statistically evaluate these claims that positive and negative contact are (not) at opposite ends of the contact quality spectrum. If positive and negative contact are indeed polar opposites to each other, we would expect separate measures of them to correlate strongly and negatively with each other (as the number of positive contact experiences increases, the number of negative contact experiences decreases). Early studies that measured positive and negative intergroup contact on separate measures found that these two measures of valenced contact correlated weakly with each other. For example, Aberson and Gaffney (2008) found that their separate measure of positive and negative contact correlated weakly with each other (r = -.05).
Other more recent research has corroborated this weak correlation between positive and negative contact
| Study | Correlation between Positive and Negative Contact |
| Kauf et al. (2017) | r = .01 |
| Pettigrew (2008) | r = -.18 |
| Schafer (2020) | r < -.18 |
| Bagci et al. (2023) | -.09 < r < .07 |
These weak correlations between positive and negative intergroup contact suggest that they are two separate constructs and that it is, indeed, fairly common to find instances where people have lots (or little) of both types of valenced contact.
Does Negative Contact Increase Prejudice More than Positive Contact Decreases Prejudice?
It was a study by Barlow and colleagues (2012) that brought this positivity bias into the spotlight. They did so with a rather provocative study titled, “The Contact Caveat: Negative Contact Predicts Increased Prejudice More Than Positive Contact Predicts Reduced Prejdice”. Such a finding is has significant ramifications for research angendas, policy, and conflict resolution. Is such a bold claim substantiated? Lets take a look.
In Study 2, Barlow and colleagues asked 441 White American participants from an online scientific survey pool. They asked about their positive and negative contact experiences with Black people. They also asked questions pertaining to modern racism, old-fashioned (blatant) racism, issue avoidance (avoiding talking about race), active avoidance (avoiding talking to Black people), and skepticism about Obama’s birthplace. Their results are as follows:
| Variables | Positive Contact Correlation | Negative Contact Correlation |
| Modern Racism | -.15 | .27 |
| Old-fashioned Racism | .13 | .24 |
| Issue Avoidance | -.16 | .25 |
| Active Avoidance | -.16 | .20 |
| Skepticism | -.09 | .11 |
By taking the absolute value of these correlations, Barlow and colleagues argue that negative contact has a stronger effect than positive contact does. These findings are consistent with broader valence-asymmetry effects found in psychology (see Baumeister et al., 2001). Valence-asymmetry effects refer to the finding that, in general, negative phenomena have a stronger impact on our attitudes, beliefs, and behaviours, than do positive phenomena. For instance, bad characteristics of a relationship tend to affect the quality of the relationship more so than positive characteristics do (see, for example, Gottman, 1994).
At surface value, Barlow and colleagues findings match up well with these well-established and pervasive valence-asymmetry effects. Barlow et al., however, provided no statistical support for this claim. I statistically compared the strength of the effects of positive and negative contact on the outcomes variables and found that negative contact was more strongly correlated with old-fashioned racism (p = .038). None of the other correlations were statistically different from each other. Such an analysis provides support against their contention that negative contact has a stronger prejudice-increasing effect than positive contact has a prejudice-reducing effect.
A review of the (burgeoning) literature into the effects of positive and negative contact on prejudice finds a wide range of effects.
| Studies finding
Negative > Positive |
Studies finding
Negative = Positive |
Studies finding
Negative < Positive |
| Alperin et al. (2014)
Barlow et al. (2012) Dhont & Van Hiel (2009) Graf et al. (2014) Hayward et al. (2018) Kauf et al. (2017) Labianca et al., (1998) Paolini et al. (2010, 2014) Techakersari et al. (2015) |
Aberson & Gaffney (2008)
Árnadóttir et al. (2018) Mazziotta et al. (2015) Reanalysis of Barlow et al. (2012) |
Aberson & Gaffney (2008)
Brylka et al. (2016) Hayward et al. (2018) Kauf et al. (2017) Mähönen & Jasinskaja-Lahti (2016) Reimer et al. (2017) |
Part of the emergent asymmetry effects might come from the wording of the contact items used. For instance, Aberson and Gaffney (2008), in measuring postive intergroup contact, asked their participants how often they had pleasant / equal / close / cooperative contact with the outgroup. When measuring negative intergroup contact, they asked their participants how frequently they had been insulted / harassed / ridiculed / intimidated by the outgroup. Arguably, one would expect negative contact to have stronger negative effect than positive contact having a positive effect given the degree of negativity / positivity that these items tap. For instance, being harassed is more negative than what pleasant is positive. If we are going to make assertions about the relative strength of positive versus negative intergroup contact, we need to make sure that what we are comparing is, indeed, comparable.
VALENCE-ASYMMETRY EFFECTS – AN UPDATE
What is quite interesting is that some studies in the table above appear in two columns. For instance, Kauf et al. (2017) evidently found instances where negative contact had stronger effects on prejudice than positive contact did. They also found instances where positive contact had stronger effects on prejdice than negative contact. What is going on here? Well…lets take a look.
Study 1 used data from 1383 adult Germans. They measured negative intergroup contact (“How often has a foreigner [bothered / intimidated] you?”) and positive intergroup contact (“How often has a foreigner [helped] you?”). They measured various attitudes towards foreigners, including angry, fear, and happiness. They found that negative contact was more strongly associated with anger and fear whereas positive contact was more strongly related with happiness. In short, this study provides evidence for a specific type of valence-assymetry effect, specifically the affect-matching hypothesis: negative contact better predicts negative outcomes and positive contact better predicts positive outcomes. Several studies investigating the affect-matching hypothesis have returned consistent results (e.g., Barlow et al., 2019; Hayward et al., 2018).
Do Positive and Negative Contact Interact?
Considering positive and negative contact as separate constructs allows for another intriguing investigation into their combined effects. While the valence of the outcome may seem to be important in determining which variable has the stronger effect, it may be that
- prior negative contact could stop future positive contact from reducing prejudice.
- prior positive contact may make future negative contact especially effective at increasing prejudice.
The other pattern of relationships is also possible
- prior negative contact could make future positive contact more effective at reducing prejudice.
- prior positive contact could reduce the effect of future negative contact.
In a recent study using an intriguing sample, Árnadóttir and colleagues (2018) asked 357 Icelandic participants about their positive and negative contact with Polish immigrants. They measured attitudes towards Polish immigrants and Trust in Polish immigrants. They also measured the salience of group identities during positive and negative contact scenarios.
Árnadóttir and colleagues found a weaker correlation between negative contact and (worse) outgroup attitudes and outgroup trust for those participants who reported having more positive intergroup contact with Polish immigrants. As such, positive contact seemed to buffer the respondents against the deleterious effects of negative intergroup contact.
Conversely, they found a stronger correlation between positive contact and (more favourable) outgroup attitudes and outgroup trust for those participants who reported having more negative intergroup contact experiences with Polish immigrants. As such, negative contact seemed to augment the positive effects of positive contact.
Bagi and colleagues (2023) found supporting longitudinal evidence for the buffering hypothesis. In a sample of 687 Turkish adolescents (mean age = 11.11 years old) across 40 classrooms, they found that positive contact with Syrian children reduced the subsequent longitudinal effects of negative contact with Syrian children on avoiding future contact with Syrian children. Additional evidence for the buggering hypothesis can be found in Árnadóttir et al. (2022).
Why did they find this pattern or relationships? Come to class to find out!
