Seizing, Freezing, and … Suffering? Looking at Need for Closure in Romantic Relationships

Need for Closure (NFC), our tendency to “seize” upon immediate answers and “freeze” by failing to update (Kruglanski, 1989), was predicted to be negatively associated with relationship satisfaction because, once in a relationship, high NFC individuals would be motivated to stay in the relationship, even if they were unhappy, because of a preference for the familiar versus the unknown.  Two measures of relationship satisfaction were analyzed using linear regressions with the two dimensions of NFC, Decisiveness and Need for Simple Structure (NFSS), as continuous predictors.  NFSS proved to be a weak, inconsistent predictor, while Decisiveness turned out to be a strong, positive predictor of relationship satisfaction.  A possible basement effect with the sample used and alternative conceptions of the NFC construct were discussed in effort to explain the results.

Since the inception of research on romantic relationships, researchers have looked for factors that contribute to relationship satisfaction and stability.Personal factors, such as attachment style (Feeney, Noller, & Roberts, 2000;Shaver & Hanzen, 1993), have been examined, along with relationship characteristics, such as equity (Hatfield & Rapson, 1993) and investment (Rusbult, 1983).While there have been several studies on the role of personality characteristics in relationship satisfaction and stability (Campbell, 1999;Caughlin, Huston, & Houts, 2000), few have examined the impact of general cognitive characteristics.This study examines the possibility that one such multi-dimensional cognitive tendency, Need for Closure, is a factor in relationship satisfaction and stability.

Need for Closure
Need for Closure (NFC) has been defined as "the desire for a definite answer on some topic, any answer as opposed to confusion and ambiguity" (Kruglanski, 1989, emphasis in original).When faced with ambiguous situations, individuals high in NFC tend to "seize" on an apparent solution (often times, this translates into the first plausible solution they encounter) and "freeze" (i.e., not let subsequent information affect their opinion; Kruglanski).
A representative example of the impact of NFC can be seen in a classic study by Webster and Kruglanski (1994), in which the authors asked participants to listen to tape-recorded information about a job interview candidate and rate that candidate's personality and chances of success at the job.All participants received the same mix of positive and negative information, but half the participants first received information that presented the candidate in a positive light, followed by information that presented the candidate in a negative light.For the other half of the participants, the order of presentation was reversed.
Order of presentation did not affect participants low in NFC: their cognitive apparatus was flexible enough to accommodate new information, even if that information was contrary to previous information.For participants high in NFC, however, order of presentation significantly impacted perceptions of the candidate: those who received the negative information first formed a low opinion of the target and retained that opinion despite incoming positive information.After all the information was presented, they rated this candidate as being of significantly lower quality than did the high NFC participants who received positive information first.Thus, high NFC participants seized upon an immediate opinion and were not affected by subsequent, contrary information.Meanwhile, the opinions of low NFC individuals did not differ significantly based on order of presentation.
NFC has been linked to a wide variety of personal and cognitive characteristics.
For instance, it positively correlates with religious fundamentalism (Saroglou, 2002), cultural conservatism (Van Hiel & Mervielde, 2004), and authoritarianism (Chirumbolo, Areni, & Sensales, 2004), and negatively correlates with a desire for cognitive complexity (Webster & Kruglanski, 1994), individual and group creativity, and tolerance of multiculturalism (Chirumbolo et al.).While research has examined the role of NFC in satisfaction with jobs (Kosic, 2002) and job candidates (Webster & Kruglanski), the role of NFC in satisfaction with romantic relationships has yet to be examined.

The Proposed Role of NFC in Romantic Relationships
One could argue that dating and forming romantic relationships can be fraught with ambiguity (Drigotas, Rusbult, & Verette, 1999;Tran & Simpson, 2009).Deciding how we feel or how our partner feels about the relationship, interpreting when to escalate a situation or a relationship, interpreting the meaning of our partners' actions and responses… seldom are these aspects of relationship formation and maintenance objective and clear-cut.These matters can become more predictable, however, the more time we spend with a particular individual.Thus, if one is especially uncomfortable at the prospect of dealing with the continuous string of unfamiliar people that characterizes the dating scene, it is reasonable that he or she would be motivated to opt for the first passable relationship opportunity so that circumstances will become a little more stable and predictable.For these same reasons, such a person might also seek to maintain the relationship despite subsequent, unpleasant complications.
There is peripheral support for this idea in the relationship literature, even if the specific phrase "Need for Closure" is not used.Drigotas and Rusbult (1992) have proposed a Dependence Model of breakups that states that some individuals remain in unsatisfying relationships due to dependence on the other person and the relationship as a whole.The authors framed this dependence as a function of the available alternative relationships, but Johnson and Rusbult (1989) also said that individuals high in commitment tend to derogate alternative relationships.Thus, even if alternative options are available, individuals seeking predictability may be motivated to derogate those options.In addition, high NFC individuals are likely to view alternative relationships, no matter how otherwise enticing, typically as having less predictability as compared to established relationships, and this could lead such individuals to place less value on these alternatives.Finkel and Rusbult (2008) also have proposed that, while a desire to make personal sacrifices for the sake of relationship maintenance is often associated with relationship satisfaction, an excessive desire to do so can lead to neglect of one's own well-being.Research has produced support for this idea in cases featuring severe consequences (i.e., physical and psychological abuse) and well as those featuring more general dissatisfaction (Fritz & Helgeson, 1998;Rusbult & Martz, 1995).
Finally, although the impact of NFC on relationship quality has yet to be studied, several subscales of NFC have been correlated with types of Insecure Attachment (Mikulincer, 1997), which in turn, has been associated with relationship instability later in life (Bartholomew & Horowitz, 1991;Karavasilis, Doyle, & Markiewicz, 2003).The current research would examine the link between the construct of NFC and relationship satisfaction and stability more directly.
Thus, in the realm of dating and relationships, it seems possible that the cognitive tendencies described by NFC could combine to create a situation where one seizes upon the first minimally acceptable dating partner and stays with him or her despite subsequent problems.The high NFC individual would be motivated to stay in an existing relationship due to his or her high level of decisiveness and due to a discomfort with the ambiguity of starting a new relationship (as opposed to relative comfort of the familiar one).This is not to say that high NFC individuals are more likely to become involved in unsatisfying relationships, or that their cognitive characteristics lead the relationships to be unsatisfying, merely that once involved in a relationship, they will be motivated to remain involved, even if the relationship is less than optimal.

Measuring NFC in the Current Study
NFC is comprised of five components: Preference for Predictability, Preference for Order, Discomfort with Ambiguity, Decisiveness, and Close-Mindedness (Webster & Kruglanski, 1994).A sample item for each of the components and the reported alphas from the original Webster and Kruglanski study are listed below: Preference for Predictability (α = .82):"I dislike unpredictable situations." Preference for Order (α = .78):"I find that establishing a consistent routine enables me to enjoy life more." Discomfort with Ambiguity (α = .67):"When I am confused about an important issue, I feel very upset." Decisiveness (α = .70):"When faced with a problem, I usually see the best solution very quickly." Close-Mindedness (α = .62):"I do not usually consult many different opinions before forming my own view." Often, these scores are summed to form a total NFC score.However, Decisiveness will be kept separate from the other four scores in this study for several reasons.First, Neuberg, Judice, and West (1997) give a compelling argument that NFC is comprised of two distinct constructs: the preference for decisive answers (i.e., Decisiveness) and the need to create and maintain simple structures (i.e., the other four components).This two-dimensional structure of the original NFC scale was confirmed in a factor analysis conducted by Roets and Van Hiel (2007).Second, when the five components are combined, one of them involves the former construct and four of them involve the latter construct.This creates a total score that is largely determined by our need for simple structure, as it accounts for 35 of the 42 items that comprise the scale.
To provide equal weighting to these two constructs (since it claims to involve both "seizing" and "freezing"), this study examines Decisiveness and a composite Need for Simple Structure (NFSS) measure created by averaging the other dimensions.Finally, in this study's sample, Decisiveness did not significantly correlate with the NFSS items (r = .05,p < .46),so the two dimensions should not be combined into a composite variable.
Hypotheses 1) High scores on both the Decisiveness and NFSS components should show a negative relationship with measures of relationship satisfaction.This is not to say that individuals with these characteristics are more likely to get involved in unsatisfying relationships, but that they will remain involved in unsatisfying relationships longer than individuals low on these dimensions.Thus, at any given time, these individuals, as a group, will be numerically more likely to be involved in unsatisfying relationships.
2) Similarly, individuals in unsatisfying relationships should be more likely to be high in Decisiveness and NFSS.
3) High scores on the Decisiveness and NFSS components should be positively associated with relationship length.
Decisiveness variable indicated a greater tendency to "seize" upon the first available solution, while higher values on the NFSS construct indicated a greater tendency to "freeze" and maintain one's current position in the face of incoming information.
The other scale used was Hendrick's (1988) Relationship Assessment Scale (RAS).If participants were not currently involved in a romantic relationship, they did not complete this scale.The scale consists of seven items pertaining to one's current relationship and is answered using a seven-point Likert scale.The scores were averaged to form a total index of relationship satisfaction.The RAS was selected because it is shorter than the Dyadic Adjustment Scale (Spanier, 1976) and covers more dimensions than the Perceived Relationship Quality Component Inventory (Fletcher, Simpson, & Thomas, 2001).
The other primary measure that participants completed was loosely based on a scale constructed by Sakalli-Ugurlu (2003).Once again, if participants were not currently involved in a romantic relationship, they did not complete this scale.It consists of a list of 18 adjectives that range from positive descriptions (e.g., "happy") to negative descriptions (e.g., "boring").Participants are asked to select nine from the list that were most descriptive of their own relationships and rank the nine items from most to least relevant.A Descriptor score was created by adding one point if a positive characteristic was selected and subtracting one point if a negative characteristic was selected, giving a total range of -9 to +9 on this measure.The purpose for including this measure was to tap an aspect of relationship evaluation that was different than the RAS (i.e., description instead of assigning a numeric value).Scores produced by this scale showed a significant positive relationship with RAS scores, r = .46,p < .01(see Table 1).
Finally, participants were asked how long they have been in their current relationship.All responses on this question were converted into weeks.

Results
First, t-tests were conducted that examined whether there were differences between participants who were in relationships and those who were not in terms of Decisiveness and NFSS.Neither of the t-tests was significant, so one can assume equivalence between the groups on the dimensions of interest.
Next, RAS scores and Descriptor scores provided by participants involved in relationships were analyzed using linear regressions.Decisiveness and NFSS were used as predictors in the regressions, along with an interaction variable.

Relationship Assessment Scores
A regression analysis was conducted using NFSS scores, Decisiveness scores, and the interaction variable as predictors and RAS scores as the dependent variable.
Decisiveness was a significant predictor of RAS scores, t (139) = 2.27, β = 1.15, p < .05,such that high Decisiveness was equated with high relationship satisfaction.The interaction term was a marginally significant predictor of RAS scores, t (139) = -1.77,β = 1.09, p < .08,and will be examined further below.NFSS failed to significantly predict RAS scores.
The interaction effect for RAS scores was further examined using the general procedures outlined by Aiken and West (1991).This essentially requires re-conducting the original analysis, except that one of the predictors acts as a moderator, with the "high" condition representing scores one standard deviation above the variable mean, and the "low" condition representing scores one standard deviation below the mean.
When the regression was analyzed using Decisiveness as the predictor and NFSS as the moderator, the pooled slope for Decisive was significant and positive for participants low in NFSS (see Table 1).In other words, people who had low NFSS scores had higher RAS scores when they were also high in Decisiveness.In contrast, the simple slope for Decisive was not significant for participants with high NFSS scores (see Figure 1).Similar results emerged from the analysis using NFSS as the predictor and Decisiveness as the moderator: the pooled slope for NFSS was significant and negative for participants with high Decisiveness scores (see Table 1).That is, people who were high in Decisiveness had higher RAS scores when they were low in NFSS.In contrast, the simple slope of NFSS was not significant for participants low in Decisiveness.

Scores
Low NFSS High NFSS

Descriptor Scores
A regression analysis was conducted using NFSS, Decisiveness, and the interaction variable as the predictors and Descriptor scores as the dependent variable.
Decisiveness was a significant predictor of Descriptor scores, t (126) = 2.24, β = 1.21, p < .05,such that high Decisiveness was equated with more positive relationship descriptions.NFSS was a marginally significant predictor of Descriptor scores, t (126) = 1.82, β = .62,p < .08,such that higher levels of NFSS were associated with more positive relationship descriptions.The interaction term also was a significant predictor of Descriptor scores, t (126) = -2.00,β = -1.30,p < .05.This interaction effect was further examined using the method highlighted above.
When the regression was analyzed using Decisiveness as the predictor and NFSS as the moderator, the pooled slope for Decisive was significant and positive for participants low in NFSS (see Table 1).In other words, people who had low NFSS scores provided more positive Descriptors when they were also high in Decisiveness.
In contrast, the simple slope for Decisive was not significant for participants with high NFSS scores (see Figure 2).Similar results emerged from the analysis using NFSS as the predictor and Decisiveness as the moderator: the pooled slope for NFSS was marginally significant and negative for participants with high Decisiveness scores (see Table 1).That is, people who were high in Decisiveness provided more positive Descriptors when they were low in NFSS.In contrast, the simple slope of NFSS was not significant for participants low in Decisiveness.

Current Relationship Length
A regression analysis was conducted using the continuous versions of NFSS, Decisiveness, and the interaction variable as the predictors and the response to the question "What is the length of your current relationship?"as the dependent variable.A main effect was predicted for the NFSS components, such that participants high in Need for Simple Structure would have significantly longer relationships than those low in the construct.Although both Decisiveness and NFSS were significantly correlated with relationship length (see Table 1), neither variable nor the interaction significantly predicted relationship length in the regression.

Low Quality Relationships
This study seeks to examine the role of Need for Closure on relationship satisfaction, but one of its primary hypotheses involves the idea that high Decisiveness and Need for Simple Structure are associated with low-quality relationships.This does not mean, however, that extremely low levels of Decisiveness and NFSS were expected to equate to extremely satisfying relationships, which is the sort of assumption made by the previous analyses.Thus, an analysis was conducted comparing low quality relationships against the rest of the sample.In this case, "low-quality relationships" were arbitrarily defined as those with mean values 5.00 or lower on the RAS scores.
Although a value of five on a seven-point scale may seam high, 105 of the 144 participants completing the RAS had scores over 5.00.
T-tests were conducted comparing these two groups ("low-quality relationships" versus "high-quality relationships") in terms of Decisiveness, NFSS, and Relationship Length.There was a significant difference in terms of Decisiveness, t (141) = -2.79,p < .01,such that low-quality relationships (M = 3.29, SD = .89)were lower in Decisiveness than high-quality relationships (M = 3.76, SD = .90).Conversely, there was a marginally significant deference in terms of NFSS, t (141) = 1.87, p < .07,such that low-quality relationships (M = 4.09, SD = .61)were higher in Need for Simple Structure than high-quality relationships (M = 3.86, SD = .66).The low-quality relationship (M = 29.51,SD = 26.23)were actually longer (in terms of weeks) on average than the high-quality relationships (M = 25.12,SD = 19.36),but this difference did not reach significance (p = .28).

Discussion
This study examined the effect of components of Need for Closure on relationship satisfaction using two different dependent measures for relationship satisfaction and produced several telling results.One of the central hypotheses for this study was that being high in Decisiveness and high in Need for Simple Structure would lead to an unwillingness to leave unsatisfying relationships.There were few results to confirm this hypothesis.Although low-quality relationships were associated with higher levels of NFSS (to a marginally significant degree), they were also associated with significantly lower levels of Decisiveness.In fact, Decisiveness was, by far, the most consistent predictor of both measures of relationship satisfaction, and the correlation was always positive.
One possible problem in testing this hypothesis in the manner used has to do with the sample.It required a critical mass of "low-quality relationships," which the sample did not provide: an average RAS score of 5 out of a possible 7 doesn't exactly imply that these individuals are miserable.There probably was also a high basement effect for these scores: out of 144 participants, the lowest RAS score was 2.14 (recorded by one participant), and only 19 participants had RAS scores lower than "4."Perhaps a more sensitive or less obvious measure of relationship satisfaction (or dissatisfaction) is needed, or a way to find a large number of people in bad relationships who are willing to fill out a questionnaire.
Another issue has to do with the Need for Closure construct.A recent article by Roets and Van Hiel (2007) re-examined the two-dimensional approach to the construct offered by Neuberg and others (1997), but also provided a more elegant solution.Mannetti and others (2002) pointed out that the Decisiveness subscale dealt with questions about actual abilities (e.g., "When faced with a problem, I usually see the one best solution very quickly.")rather than merely preferences, which are assessed in the other four scales.Roets and Van Hiel constructed an alternative Decisiveness subscale that taps preferences, rather than abilities, and conducted several factor analyses, determining that the best solution for a unidimensional NFC construct was to replace the original Decisiveness subscale with their subscale and to drop the Closemindedness subscale.
Of course, explaining the role of decisiveness, as an ability rather than a preference, does not become markedly easier.However, there may be an indirect link if we view conceptually related concepts.For example, while the relationship between the concept of decisiveness, as operationalized by the Need for Closure subscale, and relationship satisfaction had never been directly examined, it is conceptually tied to the personality construct of extroversion (Trapnell & Wiggins, 1990).Extroversion, in turn, has shown a significant positive correlation with relationship satisfaction on multiple occasions (Kwan, Bond, & Singelis, 1997;Roisman et al., 2007).If behavioral decisiveness is viewed as a facet of extroversion, it may turn out to be an especially potent facet in predicting relationship satisfaction.
An alternative way to interpret the interaction of Decisiveness and NFSS in this study could be to say that participants who are high in Decisiveness and low in NFSS seem to have exceptionally fulfilling relationships.Using separate dependent measures, similar trends were revealed (see Figure 1 and 2).The reason for the high level of satisfaction is currently unknown, but the combination of characteristics describes "seizing" without "freezing."Perhaps these individuals are low in behavioral inhibition, but do not have the aversion to ambiguity that (might) lead to staying in unfulfilling relationships.
The results of this study provide minimal support for the hypothesis that High NFC individuals are more apt to stay in unsatisfying relationships, and we still don't know whether High NFC individuals view relationships is a particular way (e.g., perhaps with an overemphasis on predictability).Further research would have to address these issues before specific conclusions can be reached.Before abandoning the examination of the link between NFC and relationship satisfaction, however, it would be advised to attain a larger sample of low-quality relationships and to closely follow the recommendations provided by Roets and Van Hiel (2007) as to use of the NFC construct.Otherwise, the ethnic composition of the sample was quite diverse (reflecting the population of University of Hawaii students) and an effort was taken to get a decent age range of participants without having outliers impact the "length of relationship" variable.
Perhaps a more promising research avenue would be to examine the finding that individuals high in behavioral Decisiveness and low in Seizing are especially likely to become involved with satisfying relationships.If this finding could be replicated with other published measures of relationship satisfaction, then researchers could further explore the aspects of these cognitive tendencies that lead to fulfilling, stable relationships.

Figure 2 .
Figure 2. Interaction effect between Need for Simple Structure and Decisiveness on Descriptor Scores

Table 1
Correlations between dependent and independent variables used in the studyInteraction effect between Need for Simple Structure and Decisiveness on RAS scores Note: Standard errors are shown in parentheses.Low and high refer to one standard deviation above and below the sample means.* p < .10. ** p < .05. *** p < .01.