Personality, cognitive style and assortative mating

Personality, cognitive style and assortative mating

Personality and Individual Differences 30 (2001) 1199±1209 www.elsevier.com/locate/paid Personality, cognitive style and assortative mating Joseph G...

117KB Sizes 0 Downloads 76 Views

Personality and Individual Differences 30 (2001) 1199±1209

www.elsevier.com/locate/paid

Personality, cognitive style and assortative mating Joseph Glicksohn *, Hilla Golan Department of Criminology, Bar-Ilan University, 52100 Ramat Gan, Israel Received 8 October 1999; received in revised form 2 May 2000; accepted 10 May 2000

Abstract We investigated the degree of assortative mating among married couples, in terms of both personality (the Eysenckian Big Three and sensation seeking) and cognitive style (®eld dependence-independence; FDI). We found sizable assortative mating (0.25r0.29, p<0.05) for three dimensions of sensation seeking, namely experience seeking (ES), disinhibition (Dis) and boredom susceptibility (BS), all implicating an impulsive mode of sensation seeking. This stands in marked contrast to the Eysenckian Big Three, for which there does not seem to be assortative mating. We also found sizable assortative mating for FDI (r=0.38). We analyze and discuss these ®ndings with respect to such factors as age, length of relationship, lie scores, the formation of pseudo couples, the assessment of pro®le similarity and match-mismatch in FDI. # 2001 Elsevier Science Ltd. All rights reserved. Keywords: Assortative mating; Field dependence; Sensation seeking; Personality

1. Introduction Some marriages are made in heaven; some seem to be made in hell. What type of `glue' binds a couple together? One suggestion is that there is no lawfulness here; choosing a marital partner, as opposed to other fateful decisions that one makes in life, is just done on a random basis, though within a well-de®ned pool of people essentially similar to oneself (Lykken & Tellegen, 1993). That is to say, while similarity (however de®ned) restricts the sampling space of potential partners, the particular choice is made at random. A second option, having by far a more popular backing, is to suggest that `two birds of a feather ¯ock together', or in psychological parlance, assortative

* Corresponding author. Fax: +972-3-6350995. E-mail address: [email protected] (J. Glicksohn). 0191-8869/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved. PII: S0191-8869(00)00103-3

1200

J. Glicksohn, H. Golan / Personality and Individual Di€erences 30 (2001) 1199±1209

mating based on similarity (Byrne, 1971; Caspi & Herbener, 1990; Caspi, Herbener & Ozer, 1992; DuFort, Kovess & Boivin, 1994; Feingold, 1988; Lesnik-Oberstein & Cohen, 1984; Melamed, 1994; Murstein, 1970; Russell & Wells, 1991; Schroth, 1991; Thornquist, Zuckerman & Exline, 1991). Presumably, similarity (and dissimilarity) may be assessed along key personality dimensions, though it is far from clear: (1) how similarity should be de®ned (e.g., as a proximity [distance] measure between pro®les, or as the degree of correlation of key traits); and (2) at what level in the hierarchy of personality should this be assessed (e.g., at the [supertrait] level of extraversion, psychoticism and neuroticism, or at the [trait] level of sociability, impulsivity, sensation seeking, etc.). For present purposes, we shall be employing a Pearson correlation as an index of similarity. This might not be the best choice (e.g., Maguire, 1999); this might also not be the wisest choice (e.g., Tversky, 1977; Tversky & Gati, 1978). But, the degree of spousal correlation will have clear meaning for the reader. We shall be looking both at the level of the supertraits and at the level of the traits, to see whether assortative mating based on similarity is more clearly seen at one or the other levels of personality description. We shall also be comparing similarity of pro®les to similarity of traits to see whether we can add to this developing literature (McCrae, 1993; McCrae, Stone, Fagan & Costa, 1998). A third suggestion regarding the type of `glue' binding a couple together, is expressed by the notion of complementarity (Hinde, 1997; Winch, 1963). While `opposites' might be mutually attractive and attracting, this notion is one which is even harder to assess and study than that expressed by assortative mating. One way, however, of operationally de®ning the concept is via the categorization of match-mismatch in cognitive style. To this end, the cognitive style of ®eld dependence-independence (FDI) of each spouse will be assessed; each will then be categorized as being either ®eld dependent (FD) or ®eld independent (FI). Four spousal pairs are thus made possible: the spouses can be matched in cognitive style (e.g., both being FD); they can also be mismatched in cognitive style (i.e., one being FD and the other being FI). The latter combination would be of particular interest, in that an FD±FI spousal relationship exempli®es the notion of complementarity (see below). The Eysenckian Big Three (Eysenck & Eysenck, 1985) has generated research assessing the degree of spousal similarity. Russell and Wells (1991) reported that assortative mating was unlikely for extraversion (E), and more likely for psychoticism (P). However, they suggested that the best factor for spousal matching was the Eysenckian Lie Scale (L), falling outside the realm of discussion of the Big Three. This raises the specter of social desirability (Russell & Wells, 1992), and we shall be returning to this issue from the vantage of our own study. If assortative mating is found for both P and L, then this should also be found for impulsive sensation seeking (Glicksohn & Abula®a, 1998; Zuckerman, 1994). Zuckerman's own study here (Thornquist et al., 1991) highlights the subfactors of experience seeking (ES) and disinhibition (Dis), replicating ®ndings of an earlier study by Lesnik-Oberstein and Cohen (1984). Both ES and Dis load on the P-dominated impulsive mode of sensation seeking (Glicksohn & Abula®a, 1998). Thus, one could gain valuable information by exploring both the level of the supertraits and the level of one trait, sensation seeking, which loads on two of these three dimensions (E and P). We stress, however, that the spousal correlations are all moderately positive, tending to range at best between 0.30 and 0.40 at the p<0.01 level (see Thornquist et al., 1991, with respect to sensation seeking; see Russell & Wells, 1991, with respect to the Eysenckian dimensions). This is a comparable range to that found for attractiveness (Feingold, 1988). Lesnik-Oberstein and Cohen

J. Glicksohn, H. Golan / Personality and Individual Di€erences 30 (2001) 1199±1209

1201

(1984) also looked at the cognitive style of ®eld dependence-independence (FDI) in their research. Evidently, there is also assortative mating here, and within this same range of correlation, but there is also the distinct possibility for complementarity. Witkin and Goodenough (1977, 1981) emphasized that ®eld-independents work more autonomously than do ®eld-dependents, while ®eld-dependents prefer to work with others. Further, ®eld-dependents tend to compromise in con¯ict situations, whereas ®eld-independents are unwilling to budge (Witkin & Goodenough, 1977). Consider, therefore, a married couple comprised of one ®eld-independent partner and a ®eld-dependent spouse. There should be little strife here, because the two styles match like a lock and key. Turning to the case of the marriage of two ®eld dependents, we note that because they are both other-oriented (Dreyer, 1991; Witkin & Goodenough, 1977), they should also provide for a nice match. But now consider the case of the marriage of two ®eld independents. Neither is person-oriented; both of them going o€ in their own, autonomous directions. The room for potential strife here is quite large. By considering the FDI assessment of each spouse, we can study both similarity and complementarity in our sample of married couples. What is the relationship (if any) between FDI and personality? Zuckerman (1979, 1994) has suggested that high sensation seekers should be ®eld independent. Eysenck, on the other hand, had suggested that sensation seeking loads on his dimension of extraversion (Eysenck, 1994), and that extraverts should be ®eld dependent (Eysenck, 1982, 1983). This basic mismatch in prediction provided an additional impetus for conducting the study reported here. We thought it well worthwhile to explore both personality and cognitive style when looking at assortative mating. As we shall show, even with a small sample, we were able to replicate a number of key ®ndings in the literature, while emphasizing the centrality of both the FDI cognitive style and impulsive sensation seeking in this matching. We shall also highlight some diculties involved in assessing spousal similarity. 2. Method 2.1. Participants Sixty-®ve married couples participated on a voluntary basis. All were Israeli and mainly middle class. They were either acquaintances or friends of the second author, or else one of the spouses was studying in our department, and signed up for the study as part of the 1st-year requirement. Their ages ranged between 26 and 66 years (Mwife=37.1, Mhusband=39.4), and the length of their relationship ranged between 4 and 54 years (M=16.3 years), which is a much larger range than that commonly reported (if at all) in the literature. The majority of these couples (n=52) had children living at home (i.e., at least one child under the age of 18 years). We asked each spouse to rate his or her degree of happiness in the relationship. As Russell and Wells (1991) have suggested, even a single item tapping marital satisfaction is probably adequate for research purposes. For 20 couples, this was on a 5-point scale (1 [not at all happy] to 5 [perfect]; Mwife=4.22, Mhusband=4.53; average spousal [wife/husband] ratio=1.01); for 45 couples, this was assessed as a percentage (0 [not at all happy] to 100 [perfect]; Mwife=85.78, Mhusband=85.11; average spousal ratio=1.02). These couples were therefore: (1) quite happy in their relationship; and (2) happy to the same degree, and irrespective of the format of the question.

1202

J. Glicksohn, H. Golan / Personality and Individual Di€erences 30 (2001) 1199±1209

2.2. Personality measures 2.2.1. The Sensation Seeking Scale (SSS), form V (Zuckerman, 1979) This is a 40-item forced-choice questionnaire comprised of four subfactors: thrill and adventure seeking (TAS), disinhibition (Dis), experience seeking (ES) and boredom susceptibility (BS). Psychometric data for the Israeli population using our Hebrew version have recently been reported (Glicksohn & Abula®a, 1998). 2.2.2. The Eysenck Personality Questionnaire Ð Revised, short version (EPQ-R-S; Eysenck, Eysenck & Barrett, 1985) This is a shortened and revised version of the EPQ, consisting of 48 items having a forcedchoice format, tapping extraversion (E), neuroticism (N) and psychoticism (P), as well as a lie scale (L). Psychometric data for the Israeli population using our Hebrew version have recently been reported (Glicksohn & Abula®a, 1998). The questionnaire data were coded using standard keys provided by Zuckerman (e.g., 1994) and Eysenck et al. (1985). We therefore worked with raw scores on eight variables (the four subfactors of SSS), the three Eysenckian dimensions of E, N and P, and a score on the lie scale (L), incorporated within the EPQ-R-S. 2.3. Cognitive style The Group Embedded Figures Task (GEFT), comprising seven practice ®gures and 18 test ®gures (Witkin, Oltman, Raskin & Karp, 1971), requires the subject to locate a simple ®gure (any of eight, appearing on a separate sheet at the end of the booklet) embedded within each complex ®gure. The subject attempts to locate the simple ®gures as quickly as possible, within the time allocated, which is partitioned into three sections: the practice stage, in Part A of the task (seven ®gures), and then sections B (nine ®gures) and C (nine ®gures). The total number of correct answers on sections B and C (ranging between 0 and 18) serves as the performance measure. 2.4. Procedure The study was conducted at each couple's home, at a convenient time. The various instruments of the study were administered by the second author, and completed in varying orders. The two spouses sat apart from each other and independently completed the forms. The second author was present throughout the session, and answered their questions if required. Their anonymity was assured. 3. Results 3.1. Personality traits Table 1 presents summary data for the various personality traits and the cognitive style of FDI, together with paired t-tests for respective di€erences in scores, and spousal correlations (Pearson

J. Glicksohn, H. Golan / Personality and Individual Di€erences 30 (2001) 1199±1209

1203

Table 1 Summary data and inferential tests (*p<0.05) for the personality traits and the cognitive style of FDI (®eld dependence±independence) P Wife (meanSE) Husband (meanSE) t r

E

N

L

TAS

ES

Dis

BS

GEFT

2.110.20

8.160.38 5.920.37 6.090.33

3.830.35

5.220.23

3.770.21

2.480.20

2.800.22

8.720.34 3.290.33 5.400.35

6.110.35

5.170.23

4.480.23

2.850.23 11.170.63

2.51* 0.11

1.21 0.04

4.80* 0.07

2.65* 0.28*

1.39 0.25*

5.93* 0.21

1.64 0.10

<1 0.29*

9.510.63

2.38* 0.38*

correlation coecients) for these. As can be seen, the husbands scored relatively higher on P, TAS, Dis and on the GEFT, while the wives scored relatively higher on N (neuroticism), all conforming to general expectations regarding sex di€erences in these measures (Glicksohn & Abula®a, 1998; Witkin et al., 1971). There was assortative mating for ES, Dis and BS (0.25r0.29, p<0.05) as reported in the literature (Lesnik-Oberstein & Cohen, 1984), all of these previously found to load on a dimension tapping impulsive sensation seeking (Glicksohn & Abula®a, 1998). Note in particular the spousal correlation for FDI (r=0.38), also reported in the literature (Lesnik-Oberstein & Cohen, 1984). In contrast, there was no assortative mating for any of the Eysenckian Big Three (P-E-N). One option for assessing just how large these spousal correlations were, was to employ what Kenny and Acitelli (1994) refer to as a `pseudo-couple' analysis, that is ®rst pairing each wife to a husband on a pseudo-random basis, and then computing `spousal' correlations. As Kenny and Acitelli (1994) argue: A couple may appear to be very similar in their responses, yet they may be no more similar than a man and a woman who are totally unacquainted. Consider the case of a study of food preferences in which two of the foods are liver and ice cream. The ®nding that couples share similar food preferences would not be surprising in that it is known that most people do not like liver and most people do like ice cream. We would want to know if a couple were more similar to one another than they were to a randomly paired member to whom they were not married. (p. 419) This was accomplished by generating a random uniform distribution between integers 1 and 65, and then based on this, pairing husbands to wives. For ES, Dis and BS, the correlations were now negligible (0.006r0.119, ns). Thus, assortative mating for (impulsive) sensation seeking, presented above, is clearly signi®cant. For GEFT, in contrast and much to our surprise, the correlation was still substantial (r=0.32), leaving the previous ®nding problematic. Kenny and Acitelli (1994, p. 420) discuss the case where a measure has an associated large sex di€erence (as is the case here) resulting in a stereotype-based similarity score that would not be removed by such a pseudo-couple analysis. We do not feel that the analogy holds for the case of GEFT, but we are at a loss to come up with a reasonable explanation for this artifact in our study. We shall be reporting on our analysis of the GEFT match-mismatch later on in the paper.

1204

J. Glicksohn, H. Golan / Personality and Individual Di€erences 30 (2001) 1199±1209

We checked whether the spousal correlations could be attributed to such confounding factors as age, length of relationship and each spouse's lie (L) score. ES and Dis (but not BS) scores of both wives and husbands were all found to be negatively correlated with their age and with the length of their relationship ( 0.47r 0.26, p<0.05); GEFT scores of both wives and husbands were also found to be negatively correlated with their age and with the length of their relationship ( 0.60r 0.27, p<0.05); L scores of both wives and husbands were found to be positively correlated with their age and with the length of their relationship (0.27r0.41, p<0.05). In addition, the L score of the wives was negatively correlated with their scores on ES, BS and GEFT ( 0.40r 0.32, p<0.05). Given this, we partialed out the contribution of length of the relationship, age and the L score from the data of each spouse (see Kenny & Acitelli, 1994, for a discussion of the need for such an adjustment when assessing spousal similarity). Assortative mating for ES and for Dis was now lost (r=0.19 and 0.18, respectively, ns); the spousal correlation for BS was preserved (r=0.23, p<0.05, one-tailed); the spousal correlation for GEFT (r=0.24, p<0.05, one-tailed), while diminished, was still preserved. 3.2. Spousal similarity Given the degree of assortative mating found for ES, Dis and BS separately and locally, we could continue in our analysis by looking at the degree of con®gural spousal similarity (TAS, BS, Dis and ES), computed as the degree of agreement between the sensation-seeking pro®les of the spouses. To this end, we employed McCrae's (1993) rpa coecient (pa=pro®le agreement), which is analogous to the Pearson correlation coecients computed above. This was accomplished in three steps. First, each trait score was standardized (relative to the sample of 65 husbands or wives, respectively); second, McCrae's (1993) Index of Pro®le Agreement (Ipa) was computed, using these standardized scores for the husbands (h) and wives (w), Ipa ˆ ‰2k ‡ p p6hw h2 w2 Š= 40k, where k=4 trait scores; third, rpa was computed (rp ˆ ‰Ipa = ‰…k 2†‡ …Ipa †2 Š). The couples varied, of course, in their degree of similarity ( 0.35rpa0.35), though tended to exhibit a low degree of pro®le similarity (M=0.10). On restricting the analysis to ES, Dis and BS, for each of which local similarity had been established, the degree of pro®le similarity was slightly enhanced ( 0.50rpa0.50; M=0.17). We could compare the latter to con®gural spousal similarity for P, E and N ( 0.50rpa0.50; M=0.16). Note that pro®le similarity was practically the same. However, while ES, Dis and BS had been found, in separate, to exhibit assortative mating, this had not been the case for P, E and N. And neither measure was found to be related to marital satisfaction, computed as an average spousal ratio (r= 0.10 and 0.12, respectively, ns). Furthermore, while rpa should increase, perhaps, with length of the relationship (McCrae et al., 1998), this was far from being the case for either the Big Three, impulsive sensation seeking (ES, Dis and BS), or general sensation seeking (0.04r0.10). We computed similar indices using our pseudo-couples. For general sensation seeking, the degree of pro®le similarity was quite similar to that found for the actual couples, though with a smaller mean value ( 0.35rpa0.35; M=0.06). For the Big Three, this was also similar ( 0.50rpa0.50; M=0.07). Thus, the use of such a similarity index, in the present context of spousal similarity, becomes questionable.

J. Glicksohn, H. Golan / Personality and Individual Di€erences 30 (2001) 1199±1209

1205

3.3. Grouping by ®eld dependence±independence Using a cut-o€ value of 12 (Witkin et al., 1971), and making this common to both males and females, we could de®ne for each spouse whether he or she was ®eld dependent (FD; GEFT<12) or ®eld independent (FI). Four combinations were thus both possible and actualized in our data: Both spouses being FD (n=21; Mwife=5.6, Mhusband=6.1); both spouses being FI (n=19; Mwife=14.6, Mhusband=15.0); the husband being FI coupled with the wife being FD (n=17; Mwife=6.5, Mhusband=15.3); the wife being FI coupled with the husband being FD (n=8; Mwife=14.1, Mhusband=6.6). Thus of the 65 couples, 40 were matched for FDI (62%). For matched couples, the spousal correlation for GEFT was positive (r=0.88, p<0.0001); for mismatched couples, this was negative (r= 0.70, p<0.0001). On partialing out the contribution of length of the relationship, age and the L score from the GEFT score of each spouse, these correlations were reduced but still substantial (r=0.72 and 0.40, respectively). Clearly, both similarity and complementarity in cognitive style can provide for di€erent types of marriage. We therefore checked to see whether this FDI grouping would have bearing for two key dyadic variables: (1) length of the relationship and (2) spousal ratio for marital satisfaction. To this e€ect, we ran separate 2 (FDI-wife)2 (FDI-husband) analyses of variance (ANOVA). When we looked at the length of the relationship, we found a signi®cant main e€ect for FDI grouping of the wife [F(1,61)=7.39, MSE=134.96, p<0.01], no such e€ect for FDI grouping of the husband [F(1,61)=1.12, ns] and no signi®cant interaction [F(1,61)=1, ns]: FD wives tended to be married longer (M=20.21 years) than FI wives (M=10.89 years). They could, however, have also been older. A 22 ANOVA con®rmed this [F(1, 61)=10.33, MSE=85.42, p<0.01], with no other e€ects being signi®cant: M=40.79 vs 31.93 years. Any study looking at assortative mating over a wide age range, has to consider that age comes with a particular form of baggage: the cohort e€ect (e.g., Cook & Campbell, 1979). It is quite plausible that the older group of participants, now in their late ®fties/early sixties, conform to the more traditional, pre-seventies type of marriage (e.g., Eshelman, 1994), wherein there is a match of an FI husband (`breadwinner') to an FD wife (`housewife'). Indeed, the FI wives ranged in age between 26 and 41 years; the FD wives ranged in age between 27 and 65 years. The correlation between age and GEFT score for the wives was signi®cantly negative (r= 0.60, p<0.001), much stronger than that found for the husbands (r= 0.41, p<0.001). When we looked at the spousal ratio for marital satisfaction, we found a signi®cant main e€ect for FDI grouping of the wife [F(1,61)=4.07, MSE=0.029, p<0.05], a signi®cant main e€ect for FDI grouping of the husband [F(1,61)=5.59, p<0.05] and, more importantly, a signi®cant interaction [F(1,61)=6.05, p<0.05]: the spousal ratio for FI wives married to FD husbands (M=1.20) was the highest (i.e., wife>husband); for all other combinations, the ratio was much smaller and closer to unity (M=0.992, 0.976 and 0.997, respectively). 3.4. Di€erences between spouses on personality traits and marital satisfaction One implication of studies assessing assortative mating based on similarity is that the degree of spousal mismatch along the critical personality dimensions should have predictive faculty for marital satisfaction (Lester, Haig & Monello, 1989; Russell & Wells, 1991; Schroth, 1991; Thornquist et al., 1991). As Zuckerman (1994, p. 194) has concluded, ``there is evidence that a

1206

J. Glicksohn, H. Golan / Personality and Individual Di€erences 30 (2001) 1199±1209

lack of congruence on sensation seeking can be a source of marital dissatisfaction''. Furthermore, ``the fact that divorced persons are higher on sensation seeking than married persons, and singles as well as in the case of men, suggests that sensation seeking is a risky trait for marriage even among those couples where both partners are high sensation seekers . . .'' (p. 196). Following on from ®ndings in the literature suggesting that absolute di€erences between spouses along the personality traits measured in this study would be negatively correlated with marital satisfaction (for the EPQ-R-S, see Russell & Wells, 1991; for sensation seeking, see Schroth, 1991), we turned our attention to this possibility. However, we found no appreciable correlations here in general, save for a single negative correlation between spousal ratio for marital satisfaction and the absolute di€erence in TAS (r= 0.42, p<0.0001). We shall refrain from discussing this correlation further. We do acknowledge, however, that given that our couples expressed a high degree of marital satisfaction (see Participants), this restriction of range was detrimental to studying such an hypothesis. 4. Discussion We found sizable assortative mating for three dimensions of sensation seeking, namely experience seeking (ES), disinhibition (Dis) and boredom susceptibility (BS), all implicating an impulsive mode of sensation seeking. This stands in marked contrast to the Eysenckian Big Three, for which there does not seem to be assortative mating. We also found sizable assortative mating for FDI. Nevertheless, these ®ndings have to be considered in relationship to such factors as age, length of relationship and spousal lie scores, possibly mediating or moderating degree of spousal matching. ES, Dis and GEFT scores of both wives and husbands were all found to be negatively correlated with their age and with the length of their relationship. What conclusions can be drawn from these correlations? Assortative mating could have had an in¯uence on the length of the couple's relationship, in that better-matched couples manage to last longer together. This is certainly one option. One could also argue for the reverse relationship: length of relationship results in a synergistic convergence of cognitive style and personality. This latter option, however, seems to be implausible, on two counts: First, due to previous ®ndings indicating that personality does not change markedly over the years (Costa & McCrae, 1992, 1994; though, and of importance here, not with reference to sensation seeking Ð see Zuckerman, 1994 Ð nor with reference to cognitive style Ð see Witkin et al., 1971); and secondly, because there is no apparent convergence of spousal trajectories of either marital satisfaction or personality (Caspi et al., 1992; Karney & Bradbury, 1997; Melamed, 1994). But, there is always a third option: the possibility that a third factor, related to both length of the relationship and cognitive style/personality, could create their correlation, and that is age itself. But when age, length of relationship and each spouse's L score were partialed out, these spousal correlations were still apparent for both BS and GEFT. These factors are therefore worthy of further study. Needless to say, a clear limitation of the present study is its cross-sectional nature. We can, however, compare present ®ndings with those of others employing somewhat similar designs. We were able to replicate results reported by both Thornquist et al. (1991) and Lesnik-Oberstein and Cohen (1984) with respect to assortative mating for both the GEFT and sensation seeking. We further note no sizable assortative mating for TAS in this study, and in previous ones (cf. Lesnik-Oberstein, 1984;

J. Glicksohn, H. Golan / Personality and Individual Di€erences 30 (2001) 1199±1209

1207

Thornquist et al., 1991), thus marking the assortative mating to be particular to the impulsive mode of sensation seeking. We further note that this is in marked contrast to the Eysenckian Big Three, for which there does not seem to be assortative mating. This ®nding (or lack of ®nding) lends additional support to the conclusion drawn by Zuckerman (1994, p. 198), highlighting the assortative mating found speci®cally for sensation seeking, in comparison with other personality traits. It is also clear that the FDI construct, and speci®cally match-mismatch in FDI, provides a good way of operationalizing the complementarity hypothesis. This also brings to mind the gender-asymmetry hypothesis (Gilbert, 1986; Wilson, 1986): The idea being that marital satisfaction is a function of gender-related di€erences within couples (see also Kenny & Acitelli, 1994). Thus, if the husband is FI and the wife is FD, this, presumably, would be a `good' match in terms of both cognitive style and the traditional division of household labour. At least for the husband, marital satisfaction should be high. Hinde (1997, p. 72) makes reference to a 1992 study, delineating four major types of marriage, including the case where the husband is the `breadwinner' and household chores are traditionally sex-typed. This type of marriage is reputed to hold for 27% of the cases sampled. Assuming that a traditional marriage is associated with the husband being FI and the wife being FD, marital satisfaction should be high. In our study, 17 of the 65 couples (26%) conformed to the latter pattern. The problematic coupling (at least for the husband, seen in degree of marital satisfaction) is that where the wife is FI and the husband is FD (see also Dreyer, 1991). In our study, eight couples (12%) conformed to the latter pattern. Hinde (1997) notes a similar type of marriage, wherein roles are reversed, reputed to hold for 10% of the cases sampled. It would be of interest, looking at the family as a system (Cox & Paley, 1997), to check the relationship between the FDI interaction and the di€erent types of marriage in future research. Given that the spousal FDI interaction has rarely been examined previously, and that such FDI grouping seems to be productive, it is well worth the e€ort to continue exploring along these lines, especially in a longitudinal design. Sabatelli, Dreyer and Buck (1983) had previously looked at such FDI matching of recently married couples. They reported that husbands, regardless of their own cognitive style (i.e., FD or FI) had more complaints when coupled with FD wives, than when coupled with FI ones. In the present study, in contrast, we found that when the wife was FI, degree of marital satisfaction reported by the husband was lower. Apart from the di€erence in measure used to assess FDI (which does not seem to be that critical, though is acknowledged) and the di€erence in measure used to assess marital satisfaction (again, this does not seem to be that critical), the major di€erence between these two studies seems to be in the choice of sample. In the present study, our participants were 65 married couples having a mean length of relationship spanning 16 years; in the Sabatelli et al. (1983) study, 48 recently married couples participated (mean length of relationship being under 3 years). In addition, while slightly under 25% of their sample had children (speci®cally, only one child), 52 couples (80%) in the present sample had children living at home. Both the presence of children at home (Grote, Frieze & Stone, 1996), and the length of the marital relationship (Hinde, 1997) are two key factors in¯uencing the degree of marital satisfaction. Conceivably, they serve as a suitable ground within which to explore the FDI interaction of the spouses, in a longitudinal study. We have been able to replicate a number of key ®ndings appearing in the literature looking at spousal similarity, while emphasizing the centrality of both the FDI cognitive style and impulsive

1208

J. Glicksohn, H. Golan / Personality and Individual Di€erences 30 (2001) 1199±1209

sensation seeking in this matching. We have also been able to demonstrate that match-mismatch in FDI provides a good way of operationalizing the complementarity hypothesis. Even so, our study is necessarily limited by the fact that our couples seem to be quite happy in their relationship (as are, indeed, most samples reported in the literature). As Russell and Wells (1992, p. 790) conclude, ``it is quite feasible that liars (or diplomats) enjoy good marriages''. We hope to be able to continue this line of investigation, by looking at dysfunctional couples. If the present data are anything to go on, then distraught marriages and dysfunctional relationships might very well be related to spousal dissimilarity. Acknowledgements This paper is based on an MA thesis submitted by the second author, and supervised by the ®rst. We thank Robin Russell and an anonymous reviewer for their helpful comments. References Byrne, D. (1971). The attraction paradigm. New York: Academic Press. Caspi, A., & Herbener, E. S. (1990). Continuity and change: Assortative marriage and the consistency of personality in adulthood. Journal of Personality and Social Psychology, 58, 250±258. Caspi, A., Herbener, E. S., & Ozer, D. J. (1992). Shared experiences and the similarity of personalities: A longitudinal study of married couples. Journal of Personality and Social Psychology, 62, 281±291. Cook, T. D., & Campbell, D. T. (1979). Quasi-experimentation: Design and analysis issues for ®eld settings. Boston: Houghton Mi‚in Co.. Costa, P. T. Jr, & McCrae, R. R. (1992). Trait psychology comes of age. In T. B. Sonderegger, Nebraska symposium on motivation: Psychology and aging (pp. 169±204). Lincoln, NE: University of Nebraska Press. Costa, P. T. Jr, & McCrae, R. R. (1994). Set like plaster? Evidence for the stability of adult personality. In T. Heatherton, & J. Weinberger, Can personality change? (pp. 21±40). Washington, DC: American Psychological Association. Cox, M. J., & Paley, B. (1997). Families as systems. Annual Review of Psychology, 48, 243±267. Dreyer, A. S. (1991). The role of ®eld dependence±independence in close relationships. In S. Wapner, & J. Demick, Field dependence±independence: Cognitive style across the life span (pp. 308±324). Hillsdale, NJ: Erlbaum. DuFort, G. G., Kovess, V., & Boivin, J. F. (1994). Spouse similarity for psychological distress and well being: A population study. Psychological Medicine, 24, 431±447. Eshelman, J. R. (1994). The family: An introduction (7th ed.). Boston: Allyn and Bacon. Eysenck, H. J. (1982). Book review of H. A. Witkin and D. R. Goodenough (1981). Cognitive styles: Essence and origins. Personality and Individual Di€erences, 3, 103. Eysenck, H. J. (1983). A reply to Fine. Personality and Individual Di€erences, 4, 361. Eysenck, H. J. (1994). Personality: Biological foundations. In P. A. Vernon, The neuropsychology of individual di€erences (pp. 151±207). New York: Academic Press. Eysenck, H. J., & Eysenck, M. W. (1985). Personality and individual di€erences: A natural science approach. New York: Plenum. Eysenck, S. B. G., Eysenck, H. J., & Barrett, P. (1985). A revised version of the psychoticism scale. Personality and Individual Di€erences, 6, 21±29. Feingold, A. (1988). Matching for attractiveness in romantic partners and same-sex friends: A meta-analysis and theoretical critique. Psychological Bulletin, 104, 226±235. Gilbert, D. G. (1986). Marriage and sex: Moving from correlations to dynamic personality interactions Ð limits of monocular vision. In S. Modgil, & C. Modgil, Hans Eysenck: Consensus and controversy (pp. 287±299). Philadelphia: Falmer Press.

J. Glicksohn, H. Golan / Personality and Individual Di€erences 30 (2001) 1199±1209

1209

Glicksohn, J., & Abula®a, J. (1998). Embedding sensation seeking within the big three. Personality and Individual Di€erences, 25, 1085±1099. Grote, N. K., Frieze, I. H., & Stone, C. A. (1996). Children, traditionalism in the division of family work, and marital satisfaction: ``What's love got to do with it?''. Personal Relationships, 3, 211±228. Hinde, R. A. (1997). Relationships: A dialectical perspective. Sussex: Psychology Press. Karney, B. R., & Bradbury, T. N. (1997). Neuroticism, marital interaction, and the trajectory of marital satisfaction. Journal of Personality and Social Psychology, 72, 1075±1092. Kenny, D. A., & Acitelli, L. K. (1994). Measuring similarity in couples. Journal of Family Psychology, 8, 417±431. Lesnik-Oberstein, M., & Cohen, L. (1984). Cognitive style, sensation seeking and assortative mating. Journal of Personality and Social Psychology, 46, 112±117. Lester, D., Haig, C., & Monello, R. (1989). Spouses and marital satisfaction. Personality and Individual Di€erences, 10, 253±254. Lykken, D. T., & Tellegen, A. (1993). Is human mating adventitious or the result of lawful choice? A twin study of mate selection. Journal of Personality and Social Psychology, 65, 56±68. Maguire, M. C. (1999). Treating the dyad as the unit of analysis: A primer on three analytic approaches. Journal of Marriage and the Family, 61, 213±223. McCrae, R. R. (1993). Agreement of personality pro®les across observers. Multivariate Behavioral Research, 28, 25±40. McCrae, R. R., Stone, S. V., Fagan, P. J., & Costa, P. T. Jr (1998). Identifying causes of disagreement between selfreports and spouse ratings of personality. Journal of Personality, 66, 285±313. Melamed, T. (1994). Partner's similarity over time: The moderating e€ect of relationship status. Personality and Individual Di€erences, 16, 641±644. Murstein, B. I.Theories of attraction and love. (1970). New York: Springer. Russell, R. J. H., & Wells, P. A. (1991). Personality similarity and quality of marriage. Personality and Individual Differences, 12, 407±412. Russell, R. J. H., & Wells, P. A. (1992). Social desirability and quality of marriage. Personality and Individual Di€erences, 13, 787±791. Sabatelli, R. M., Dreyer, A., & Buck, R. (1983). Cognitive style and relationship quality in married dyads. Journal of Personality, 51, 192±201. Schroth, M. L. (1991). Dyadic adjustment and sensation seeking compatibility. Personality and Individual Di€erences, 12, 467±471. Thornquist, M. H., Zuckerman, M., & Exline, R. U. (1991). Loving, liking, looking and sensation seeking in unmarried college couples. Personality and Individual Di€erences, 12, 1283±1292. Tversky, A. (1977). Features of similarity. Psychological Review, 84, 327±352. Tversky, A., & Gati, I. (1978). Studies of similarity. In E. Rosch, & B. B. Lloyd, Cognition and categorization (pp. 79± 98). Hillsdale, NJ: Erlbaum. Wilson, G. (1986). Personality, sexual behaviour and marital satisfaction. In S. Modgil, & C. Modgil, Hans Eysenck: Consensus and controversy (pp. 263±285). Philadelphia: Falmer Press. Winch, R. F. (1963). The modern family. New York: Holt, Rinehart & Winston. Witkin, H. A., & Goodenough, D. R. (1977). Field dependence and interpersonal behavior. Psychological Bulletin, 84, 661±689. Witkin, H. A., & Goodenough, D. H. (1981). Cognitive styles: Essence and origins. New York: International Universities Press. Witkin, H. A., Oltman, P. K., Raskin, E., & Karp, S. A. (1971). Manual for embedded ®gures test, children's embedded ®gures test, and group embedded ®gures test. Palo Alto, CA: Consulting Psychologists Press. Zuckerman, M. (1979). Sensation seeking: Beyond the optimal level of arousal. Hillsdale, NJ: Erlbaum. Zuckerman, M. (1994). Behavioral expressions and biosocial bases of sensation seeking. Cambridge: Cambridge University Press.