Do Cold Feet Warn of Trouble Ahead? Premarital Uncertainty and Four-Year Marital Outcomes

Abstract

Are the doubts that people feel prior to marriage signs of impending difficulties or normative experiences that can be safely ignored? To test these opposing views, we asked 464 recently-married spouses whether they had ever been uncertain about getting married and then compared four-year divorce rates and marital satisfaction trajectories among those partners with and without premarital doubts. Doubts were reported by at least one partner in two-thirds of couples. Women with premarital doubts had significantly higher four-year divorce rates, even when controlling for concurrent marital satisfaction, the difficulty of their engagement, history of parental divorce, premarital cohabitation, and neuroticism. Among intact couples, men’s and women’s doubts predicted less satisfied marital trajectories. Premarital doubts appear to be common but not benign, suggesting that valid precursors of marital distress are evident during couples’ engagements.

From Much Ado about Nothing to Runaway Bride, images of premarital doubts are ubiquitous in Western society. Perhaps as common today are popular articles about what this uncertainty really means and whether newlyweds-to-be should be alarmed by doubts. Should premarital uncertainty be dismissed outright, or is it meaningful? Among couples who do proceed with their marriages, even in the face of uncertainty, are they at heightened risk for negative outcomes later on compared to couples for whom these doubts were not present? The current study is the first to empirically address these questions.

On the one hand, there are good reasons to believe that premarital doubts will not predict future distress. To the extent that confidence commonly diminishes as major events approach (Gilovich, Kerr, & Medvec, 1993), premarital uncertainty might simply be reflective of this general tendency to become more anxious and less optimistic, not of anything meaningful about the relationship itself. Even if the complaints were valid (“Do I really want to spend my life with someone who can be moody?”), they need not be cause for concern: “growth” or “work-it-out” theories of relationships (Franiuk, Cohen, & Pomerantz, 2002; Knee, 1998) maintain that relationships evolve over time such that what was once problematic is no longer so. We might also expect that once couples get married, they should be motivated to downplay any doubts they might have once harbored and focus on what lies ahead, much in the same way that romantic partners frequently engage in a variety of cognitive processes that allow them to see their relationship and partner in a positive light and maintain their commitments to each other (e.g., believing that their relationships have recently improved, Karney & Frye, 2002; believing their partners meet their ideals, Murray, Holmes, & Griffin, 1996; making benign attributions for irritating behaviors, Bradbury & Fincham, 1990).

But what if premarital uncertainty is not simply another normative challenge to overcome, but is instead a true warning sign? When asked to recall the history of their marriages, individuals who have divorced commonly report feeling uncertain prior to their weddings, but ultimately dismissed these doubts as mere anxieties or told themselves that all relationships were difficult in the beginning (e.g., Vaughan, 1986). These underlying concerns remained, however, and in time grew stronger and eventually disrupted the marriage. These firsthand accounts – though limited by their retrospective nature –raise the possibility that doubts capture real concerns about the status quo and foreshadow more trouble ahead. Although no studies to date have empirically tested this claim, there is increasing evidence that (1) marital satisfaction and problems are relatively stable over time (e.g., Kamp Dush & Taylor, 2012) and (2) the roots of relationship distress are apparent from the beginning of couples’ relational trajectories (e.g., Lavner, Bradbury, & Karney, in press; Mattson, Frame, & Johnson, 2011), consistent with the idea that concerns about the relationship during the engagement are a good proxy for how the marriage will unfold.

To test these opposing views, we asked recently married spouses about their premarital uncertainties and subsequently compared four-year marital outcomes of men and women with and without premarital doubt. If cold feet are just that and nothing more, we would expect no differences in four-year divorce rates or marital satisfaction trajectories between individuals with and without premarital doubt. If, however, cold feet are meaningful, we would expect higher divorce rates and poorer marital satisfaction trajectories among individuals who married despite experiencing premarital jitters. To allow for the possibility that the effects of premarital uncertainty might differ by gender, we examined these effects separately for men and women. Given evidence that women (1) have more nuanced views of their relationships (e.g., Rempel, Holmes, & Zanna, 1985) and greater sensitivity to relationship problems than men (e.g., Rubin, Peplau, & Hill, 1981), and (2) are more likely to initiate divorce (e.g., Amato & Previti, 2003), we predicted that women’s doubts should be more diagnostic than men’s doubts. Specifically, we predicted that if doubts did predict negative outcomes (i.e., higher divorce, lower marital satisfaction), they would be more likely to do so among women than men.

To ensure that doubts added unique predictive value, we controlled for other well-established predictors of marital outcomes (see Karney & Bradbury, 1995 for review), including characteristics of partners’ backgrounds (parental divorce; e.g., Amato, 1996), personality (neuroticism; e.g., Kelly & Conley, 1987), and relationship history (premarital cohabitation; e.g., Bennett, Blanc, & Bloom, 1988). These controls allowed us to test whether premarital doubts proved problematic above and beyond the effects of other variables that might contribute to such doubts, including people who (a) had more reason to be hesitant about marrying given their family background, (b) were generally characterized by more negative affective traits, or (c) had doubts because they knew more about their partners due to living together. Finally, and perhaps most importantly, we controlled for individuals’ concurrent marital satisfaction and their reports about the global emotional tone of their engagement in order to ensure that their reports of premarital doubts were not biased by their current marital satisfaction or by a broader tendency to remember their premarital period negatively.

Method

Sampling and Participants

Participants were 464 spouses from 232 newlywed couples in two longitudinal studies. Couples in both studies were eligible to participate if this was the first marriage for both spouses, the couple had been married for less than 6 months, neither partner had children, both partners were older than 18 years and wives were younger than 35 years, both spouses spoke English and had received at least a 10th grade education (to ensure comprehension of questionnaires), and the couple had no immediate plans to move from the area.

The first sample comprised 60 newlywed couples recruited from newspaper advertisements in the Los Angeles area between February 1991 and October 1991. More than 350 couples responded to the advertisements; the first 60 couples who met the criteria were invited to participate. The second sample consisted of 172 newlywed couples identified from marriage licenses filed in Los Angeles County between May 1993 and January 1994. Couples who met the initial criteria were sent a letter inviting them to participate in the study. The first 172 couples who met the criteria and arrived at their scheduled laboratory appointment constituted the second sample.

We combined the two samples because all couples met identical selection criteria; the studies used highly similar data collection procedures and intervals; and doing so afforded more power. For the combined sample, at the initial data collection, husbands averaged 27.0 years of age (SD = 3.8) and 15.6 years of education (SD = 2.2), with a median annual income between $21,000 and $30,000. Sixty-nine percent were Caucasian, 14% were Latino-Chicano, 11% were Asian American–Pacific Islanders, and 4% were African American. Wives averaged 25.5 years of age (SD = 3.4) and 16.0 years of education (SD = 1.9), with a median annual income between $11,000 and $20,000. Sixty-five percent were Caucasian, 15% were Latina-Chicana, 13% were Asian American–Pacific Islanders, and 5% were African American.

Procedure

Couples meeting all eligibility criteria were scheduled to attend a three-hour laboratory session, which was conducted within the first six months of their marriage. During the session, spouses independently completed questionnaires and were interviewed individually. Marital satisfaction and divorce status were assessed at approximately 6-month intervals for four years after the initial assessment. Depending on the sample, participants were paid $50–$75 initially and $25–$35 at each follow-up.

Measures

Marital satisfaction

We assessed marital satisfaction every six months (eight times overall) using a version of the Semantic Differential (SMD; Osgood, Suci, & Tannenbaum, 1957), a measure that asks spouses to rate their perceptions of their relationship on 7-point scales between 15 pairs of opposing adjectives (e.g., bad –good, dissatisfied–satisfied). The SMD yields scores from 15 to 105, such that higher scores reflect greater relationship satisfaction.

Premarital experiences

Premarital doubts were assessed at the first session through a single yes-no question: “Were you ever uncertain or hesitant about getting married?” Spouses responded to this question separately in interviews examining the history of couples’ pre-marital relationships. Spouses were also asked in these interviews whether their engagements (or the period of time they decided to marry and their actual wedding) were “smooth” or “difficult and turbulent.”

Background risk

As an additional discriminative test, we included three factors known to increase risk for relationship distress (see Karney & Bradbury, 1995): (1) parental divorce before age 16; (2) neuroticism, assessed using the 23-item Neuroticism scale of the Eysenck Personality Questionnaire (Eysenck & Eysenck, 1978); and (3) cohabitation with the spouse prior to marriage. Parental divorce and premarital cohabitation were assessed in an individual interview in Study 1 and via questionnaire in Study 2.

Results

Descriptive statistics

Forty-seven percent of husbands (n = 108) and 38% of wives (n = 87) reported being uncertain about getting married. Follow-up chi-square analyses indicated that premarital doubts were more frequent among men than women, χ²(1, N = 456) = 4.28, p < .05.

We conducted a series of analyses to compare the characteristics of individuals with and without premarital doubts. For husbands and wives, no significant differences were found between individuals with and without premarital doubts for age, income, education, race/ethnicity, months dated before marriage, premarital cohabitation, or parental divorce (all p > .10). Husbands with premarital doubts had significantly higher neuroticism (M = 8.11, SD = 4.78) compared to husbands without premarital doubts (M = 6.17, SD = 4.78), t(222) = 2.95, p < .01. Wives with and without premarital doubts did not differ significantly in neuroticism (p > .10). There were also non-significant trends such that husbands and wives with premarital doubts were more likely to report that their engagement was “difficult and turbulent” (41% and 41%, respectively) than husbands and wives without premarital doubts (30% and 29%, respectively), χ²(1, N = 217) = 2.91, p = .09 (husbands) and χ²(1, N = 228) = 3.37, p = .07 (wives). Finally, we examined whether there were differences in the rate of childrearing over the following four years among individuals with and without doubts. Although no differences were found among wives, there was a non-significant trend such that husbands with doubts were slightly less likely to have children in the following four years (28%) compared to husbands without doubts (38%), χ²(1, N = 224) = 2.73, p = .10.

Predicting divorce

We then examined four-year divorce rates among husbands and wives with and without premarital doubts. Divorce status was known for 228 couples (98% of the total sample). Overall, 12% of couples (n = 27) dissolved their relationship by four years. For husbands, 9% of those who reported not having premarital doubts divorced by four years (n = 10 out of 117) compared to 14% of those who did report premarital doubts (n = 15 out of 106); these groups did not differ significantly, χ²(1, N = 223) = 1.76, p > .10. Among wives, 8% of those who reported not having premarital doubts divorced by four years (n = 11 out of 141) compared to 19% of those who did report premarital doubts (n = 16 out of 84). Chi-square analyses indicated that these rates differed significantly, χ²(1, N = 225) = 6.31, p < .05.

Given that the measure of premarital doubts was retrospective, we tested the possibility that the reports of doubts were biased using two discriminative tests. First, we tested the possibility that wives with lower levels of satisfaction were more likely to report doubts and that it was this lower initial satisfaction that explained the divorce effect. Although concurrent satisfaction did significantly predict four-year divorce rates in the expected direction (p < .05), premarital doubts remained significant when controlling for initial satisfaction in a binary logistic regression, indicating that retrospective bias based on current marital satisfaction alone did not account for the doubt-divorce association. Second, we tested whether a global tendency toward recalling the premarital period in a negative light explained these effects using wives’ reports about whether their engagement was “difficult and turbulent.” Four-year divorce rates did not differ between wives who reported their engagements were difficult and those who did not, χ²(1, N = 226) = 0.31, p > .10. Further, wives’ doubts remained a significant predictor of dissolution (p < .05) when controlling for turbulent engagement, indicating that doubts were themselves uniquely predictive of negative marital outcomes.

We then conducted a series of binary logistic regressions in which we entered wives’ premarital doubts simultaneously with other well-established risk factors, including premarital cohabitation, parental divorce, and neuroticism. Again, premarital doubts remained significant (p < .05) in each of these models. Premarital cohabitation and parental divorce also predicted subsequent divorce (p < .05). Wives’ premarital doubts also remained significant (p < .05) when controlling for husbands’ premarital doubts, which were not a significant predictor of divorce.

Last, we examined divorce rates at a couple-level by classifying each couple into one of four groups. Thirty-six percent of couples were characterized as having no doubts (n = 84), 26% of couples were characterized with only husband doubts (n = 61), 17% of couples were characterized with only wife doubts (n = 40), and 20% of couples were characterized with both doubts (n = 47). We then conducted a series of contrasts in which we compared divorce rates in each group to the no doubt group (6% divorce). The husband only doubt group (10% divorce) did not differ significantly from the no doubt group, χ²(1, N = 144) = 0.81, p > .10, but the wife only doubt (18% divorce) and the both doubt groups (20% divorce) did, χ²(1, N = 122) = 4.59, p < .05 and χ²(1, N = 130) = 5.73, p < .05, respectively. Divorce rates did not differ significantly between the wife doubt and the both doubt groups, χ²(1, N = 84) = 0.02, p > .10.

Predicting marital satisfaction among intact couples

We next examined marital satisfaction trajectories among the couples who remained married at four years using growth curve analytic techniques and the HLM 7.0 computer program (Raudenbush, Bryk, & Congdon, 2010). Husbands’ and wives’ data were estimated simultaneously within the same equations. Time was estimated as number of months since the couple’s wedding date and was uncentered so that the intercept terms (B_f00 and B_m00) could be interpreted as the initial value six months into marriage. To test for differences in intercepts and linear slopes among individuals with and without premarital doubts, we used the following equations:

$$\begin{array}{l}\text{Level}1:Y_{\text{ti}}={(\mathit{female})}_{\text{ti}}[{\pi }_{\text{f}0\text{i}}+{\pi }_{\text{f}1\text{i}}{(\mathit{Time})}_{\text{ti}}]+{(\mathit{male})}_{\text{ti}}[{\pi }_{\text{m}0\text{i}}+{\pi }_{\text{m1i}}{(\mathit{Time})}_{\text{ti}}]+{\text{e}}_{\text{tij}}\\ \text{Level}2:{\pi }_{\text{f}0\text{i}}(\text{wife}\text{intercept})={\beta }_{\text{f}00}+{\beta }_{\text{f}01}({\text{doubt}}_{\text{f}})+{\mu }_{\text{f}0\text{i}}\\ {\pi }_{\text{f}1\text{i}}(\text{wife}\text{slope})={\beta }_{\text{f}10}+{\beta }_{\text{f}11}({\text{doubt}}_{\text{f}})+{\mu }_{\text{f}1\text{i}}\\ {\pi }_{\text{m}0\text{i}}(\text{husband}\text{intercept})={\beta }_{\text{m}00}+{\beta }_{\text{m}01}({\text{doubt}}_{\text{m}})+{\mu }_{\text{m}0\text{i}}\\ {\pi }_{\text{m}1\text{i}}(\text{husband}\text{slope})={\beta }_{\text{m}10}+{\beta }_{\text{m}11}({\text{doubt}}_{\text{m}})+{\mu }_{\text{m}1\text{i}}\end{array}$$

These equations include separate intercepts and slopes for men and women, and sex-specific variance components at Level 2. Sex-specific premarital doubts were included at Level 2 as a predictor of intercepts and slopes (e.g., husbands’ premarital doubts predicted their own intercepts and slopes), and were coded such that the reference group (coded as 0) was no doubts and doubts were coded as 1.

Results, shown in Table 1, indicated that men and women with premarital doubts had significantly lower levels of initial satisfaction than individuals without premarital doubts (p < .05). Individuals in both groups subsequently underwent similar significant linear declines in satisfaction over time (p < .001). Thus, compared to individuals without premarital doubts, the marital satisfaction trajectories of individuals with premarital doubts started lower and remained less satisfied over time.

Table 1.

Summary of Multilevel Models Comparing Marital Satisfaction Trajectories Among Intact Couples With and Without Premarital Doubts (N = 183 couples)

Fixed effect	Coefficient (SE)	t test	df	Effect size r
Husbands
Intercept	95.17 (0.92)
Intercept x doubt	−2.60 (1.27)	−2.05*	181	0.15
Linear slope	−0.74 (0.17)	−4.25***	181	0.30
Linear slope x doubt	0.04 (0.26)	0.15	181	0.01
Wives
Intercept	97.78 (0.74)
Intercept x doubt	−2.91 (1.23)	−2.36*	181	0.17
Linear slope	−1.11 (0.20)	−5.50***	181	0.38
Linear slope x doubt	−0.17 (0.32)	−0.51	181	0.04

^*p < .05.

^**p < .01.

^***p < .001

Notes: All intercepts were significant p < .001 because the lowest possible score was higher than zero, so these statistics are not reported. Doubts were coded 0 for no doubts and 1 for doubts, so the interaction term represents the difference for individuals with premarital doubts. Effect size r = sqrt [t²/(t² + df)].

We repeated these analyses controlling (separately) for parental divorce, premarital cohabitation, and neuroticism (all entered as sex-specific variables at Level 2). As with the divorce analyses, premarital doubts remained a significant predictor of initial levels of satisfaction (p < .05) even with these controls. The one exception to this general pattern was that husbands’ doubts no longer predicted their initial satisfaction when controlling for their neuroticism.

Discussion

Premarital doubt and uncertainty have long been thought to be common among spouses-to-be, but different theoretical vantage points make different predictions regarding whether this uncertainty is meaningful or whether it can be safely ignored. Using data from 232 newlywed couples, we first confirmed that premarital doubts are indeed common, even among couples who married: in approximately two-thirds of newlywed couples, one or both partners reported having experienced some premarital uncertainty about their decision to wed. Men more commonly had doubts than women (~50% versus ~40%), though more than half of partners were similar to each other with regard to whether or not they had doubts.

Comparisons among spouses with and without doubts showed that doubts did predict poorer marital outcomes after four years, especially among women. Women with premarital doubts went on to divorce at rates that were approximately 2.5 times higher than women without premarital doubts, and these women had less satisfied marital trajectories if they did remain married. These results remained significant even when controlling for aspects of wives’ backgrounds (premarital cohabitation, parental divorce), their personality (neuroticism), and their concurrent marital satisfaction, increasing confidence that the doubts reflected aspects of the relationship itself and indicating that doubts provided unique predictive value above and beyond these other well-established predictors. In contrast, men’s doubts did not predict divorce, either independently or in combination with women’s doubts. Although men’s doubts did predict less satisfied marital trajectories among those spouses who remained married, these results were no longer significant when controlling for neuroticism, suggesting that men’s doubts proved problematic because they reflected broader affective traits.

Before discussing the implications of these results, it is important to acknowledge several methodological limitations. First, our measure of premarital doubts consisted only of a single yes-no item, raising concerns about the validity of responses. Future work should assess premarital doubts with multiple items that assess different types and degrees of doubt. Second, our data were collected approximately twenty years ago. To the extent that marriage continues to change over time, including the age at which people get married (U.S. Census Bureau, 2011) and the percent of the population that is married (e.g., Pew Research Center, 2011), more research is needed to determine whether premarital doubts are as diagnostic of subsequent marital difficulties in the current social climate. Third, and most notably, our measure of premarital doubts was retrospective, thus raising alternative explanations for any obtained results. Although we asked spouses about their premarital doubts very early in their marriages, it is possible that how the relationship was going just a few months after the wedding biased spouses’ memories for how they felt before the wedding, such that spouses who were feeling discontented were more likely to recall doubts and those who were feeling very satisfied may have failed to recall their earlier misgivings (see Ross, 1989, for a discussion of the construction of personal histories). Nonetheless, we took several steps to rule out this alternative explanation, including controlling for concurrent satisfaction and showing that the effects were unique to doubts and did not appear for other retrospective memories of premarital relationship quality that did not involve doubts. In both cases, wives’ premarital doubts remained a significant predictor of divorce, giving greater confidence to the findings reported here. Future research should expand on these findings using a sample of engaged couples followed longitudinally into the early years of marriage to further understand this phenomenon. This work should also examine whether doubts differentially affect the timing of divorce such that individuals with doubts dissolve their marriages earlier than individuals without doubts.

Notwithstanding these limitations, we learn from this analysis that premarital doubts are common but not benign and that for women in particular, feeling uncertain about getting married foreshadowed marital distress and dissolution. To our knowledge, this study is the first to use quantitative data to directly and rigorously address the issue of uncertainty in the formation of marital unions. Taken together, the results indicate that premarital doubts are not simply an instance of feeling anxious before a major event (Gilovich et al., 1993) or something to be worked through (e.g., Knee, 1998), but a sign of possible trouble ahead. This appeared to be less true for men, consistent with our prediction that women’s greater attunement toward relationship problems (e.g., Rubin et al., 1981) would render their doubts more diagnostic.

These findings call for more in-depth and prospective research on the uncertainties people harbor about getting married. It would be interesting to examine the content of doubts, which could include specific concerns about the relationship (e.g., “I’m not sure if we’re aligned on having children”) or the partner (e.g., “Does he work too much?”) or may represent anxiety about marriage more generally (e.g., “Am I ready for this commitment?”). These different dimensions of doubt may prove more or less predictive of subsequent relationship distress, and how this distress eventually unfolds. A more nuanced view of premarital uncertainty also calls for greater consideration of how doubts reflect and affect the premarital period more generally. For example, although we did not find significant differences in rates of premarital cohabitation among individuals with and without premarital doubt, it is possible that the processes leading to premarital cohabitation could differ depending on whether people have doubts. Specifically, might individuals without doubt decide to live together because they are certain about the future of their relationship and plan to marry, whereas individuals with doubts choose to live together as a test of those doubts? Additional research is needed to address these types of questions.

Future research should also examine why wives’ doubts were more predictive of marital distress than husbands’ doubts. One useful theoretical framework for interpreting these findings is error management theory, which argues that women need to be particularly skeptical of commitments and err on the side of the caution in order to avoid the significant cost of choosing the wrong mate (e.g., underestimating men’s commitment; Haselton & Buss, 2000). As such, young women without children (as was the case here at the initial assessment) may be especially attuned to relational difficulties before committing to one partner in order to preserve the chance to bear and raise children in the most satisfying partnership, and upon choosing the wrong partner, should be especially motivated to exit the marriage quickly. This framework may also explain why premarital doubt was less common among wives than husbands, in that women with significant doubts should be more likely than men to avoid getting married. Another possibility is that men with doubts may simply not be particularly inclined to act on them, either because men generally take less initiative than wives in the context of relational difficulties (e.g., Amato & Previti, 2003; Doss, Atkins, & Christensen, 2003) or because they receive more benefits from relationships than do women (e.g., Kiecolt-Glaser & Newton, 2001). Further work is needed to disentangle these possibilities.

More generally, these findings raise several practical implications for prevention and intervention. Most notably, by highlighting the predictive value of premarital uncertainty for later marital outcomes above and beyond other risk factors, these analyses point to doubts as another useful, unique diagnostic indicator of couples’ functioning. Doubts should not simply be dismissed as a normative experience or viewed as something that will go away once partners make a commitment to each other. Rather, feelings of premarital uncertainty should be validated, taken seriously, and used as an opportunity for exploration; explicitly encouraging the disclosure of feelings of uncertainty may be necessary given the reluctance many partners feel to share their doubts (e.g., Vaughan, 1986). For example, in the context of premarital counseling, framing discussions around these common feelings might allow couples to safely disclose unresolved issues or lingering questions. In turn, these conversations could be used to reach consensus around difficult topics (e.g., opposing views on whether to have children) or for skill-building around specific problem areas (e.g., learning how to cope better with stress; cf. Bodenmann & Randall, 2012). We caution that the causal status of doubts remains unclear and that uncertainty does not necessarily predict distress for any particular couple, but given that doubts predict poorer outcomes within just the first four years of marriage on average, therapeutic consideration and attention to these feelings are warranted.

In conclusion, the data reported here are the first to demonstrate that premarital uncertainty predicts poorer marital outcomes years later. These findings are notable given the relatively short follow-up period, indicating that this process unfolds early in marriage, and the conservative methodological approach: this was a prospective examination of the relationship between premarital doubts and marital outcomes that excluded couples who, probably as a result of their doubts, decided to end their relationships before marriage. Together, they add to the view that the antecedents of marital distress are evident from the beginning of couples’ relational trajectories (e.g., Lavner et al., in press; Mattson et al., 2011) and that the partners themselves may have access to this information.