Perceived Causes of Marriage Breakdown and Conditions of Life AILSA BURNS Macquarie University A large number of structural and attitudinal variables have been regularly found to be asso- ciated with marital dissolution.Fourteen such variables were located and related to the causes of breakdown as perceived by 335 divorced and separated men and women.Five perceived causes were associated with perceptions of the marriage as beginning to break down at particular time periods,but not with duration of marriage.Sex of respondent, SES,respondent's and spouse's religion,age at marriage,marital status (divorced or sepa- rated),parity of marriage,parental approval,stability of parents'and in-laws'marriages, length of premarital acquaintance,marriage duration and number of children,and stan- dard of living since separation all were associated with one or more of the reasons given by respondents for the marriage breakdown.An exploratory factor analysis gave a 7-factor solution supporting the notion that different types of marriage breakdown can be de- scribed. Why do marriages break down?One can seek to by his sample of divorcees:"These complaints answer this question from the participants'point were not the causes of the divorce-after all every of view or by reference to certain structural and still married wife could very likely make one or demographic variables-such as age of mar- more of these charges against her husband." riage-which regularly have been found to be as- Nevertheless,in seeking to understand marital sociated with likelihood of divorce.A rather well- breakdown,it would be wrong to disregard the established list of such variables now exists.In ad- testimony of those in the best position to know dition,the rapid increase in divorce among what went wrong in their own case.It may be,for younger cohorts indicates that youth itself is now instance,that certain structural conditions work a predictor of marital dissolution (Cherlin,1981). their effects through promoting certain types of Researchers have been inclined to regard these grievance.Goode's work suggests that this may be structural correlates as more worthy of explora- the case.For example,he found an association tion than the reasons given by couples for separat- between type of complaint and duration of mar- ing.This skepticism is perhaps understandable, riage,with a wife married for a short duration given the notoriously frequent disagreement be- likely to complain of her husband's personality tween husband and wife as to the cause of the and a wife married for 15 or more years to com- trouble and the changes that occur over time in plain of his drinking,adultery,and "helling their perceptions of the experience (Kitson and around.'Other studies make it clear that the Raschke,1981).Goode (1956:131)raises another divergence so commonly noted between the ac- point when he comments on the reasons advanced counts of husbands and wives is not random. Fulton (1979)speaks of "his',and "her"divorce and found wives in a large sample more critical of This project was funded by Macquarie University the ex-partners and of the marriages,more likely Research Grant.Thanks to Irene Lovelock (data collec- to report arguments and violence,and less likely tion)and David Cairns (data analysis). to feel that the marriages had been happy for much of the time.Levinger (1966)found wives School of Behavioural Sciences,Macquarie University,more likely to complain of cruelty,drinking, North Ryde,New South Wales 2113,Australia. physical and verbal abuse,neglect and lack of August 1984 JOURNAL OF MARRIAGE AND THE FAMILY 551 This content downloaded from 211.80.94.134 on Mon,19 Dec 2016 05:27:55 UTC All use subject to http://about.jstor.org/terms
Perceived Causes of Marriage Breakdown and Conditions of Life AILSA BURNS Macquarie University A large number of structural and attitudinal variables have been regularly found to be asso- ciated with marital dissolution. Fourteen such variables were located and related to the causes of breakdown as perceived by 335 divorced and separated men and women. Five perceived causes were associated with perceptions of the marriage as beginning to break down at particular time periods, but not with duration of marriage. Sex of respondent, SES, respondent's and spouse's religion, age at marriage, marital status (divorced or sepa- rated), parity of marriage, parental approval, stability of parents' and in-laws' marriages, length of premarital acquaintance, marriage duration and number of children, and stan- dard of living since separation all were associated with one or more of the reasons given by respondents for the marriage breakdown. An exploratory factor analysis gave a 7-factor solution supporting the notion that different types of marriage breakdown can be de- scribed. Why do marriages break down? One can seek to answer this question from the participants' point of view or by reference to certain structural and demographic variables-such as age of mar- riage-which regularly have been found to be as- sociated with likelihood of divorce. A rather well- established list of such variables now exists. In ad- dition, the rapid increase in divorce among younger cohorts indicates that youth itself is now a predictor of marital dissolution (Cherlin, 1981). Researchers have been inclined to regard these structural correlates as more worthy of explora- tion than the reasons given by couples for separat- ing. This skepticism is perhaps understandable, given the notoriously frequent disagreement be- tween husband and wife as to the cause of the trouble and the changes that occur over time in their perceptions of the experience (Kitson and Raschke, 1981). Goode (1956:131) raises another point when he comments on the reasons advanced This project was funded by Macquarie University Research Grant. Thanks to Irene Lovelock (data collec- tion) and David Cairns (data analysis). School of Behavioural Sciences, Macquarie University, North Ryde, New South Wales 2113, Australia. by his sample of divorcees: "These complaints were not the causes of the divorce-after all every still married wife could very likely make one or more of these charges against her husband." Nevertheless, in seeking to understand marital breakdown, it would be wrong to disregard the testimony of those in the best position to know what went wrong in their own case. It may be, for instance, that certain structural conditions work their effects through promoting certain types of grievance. Goode's work suggests that this may be the case. For example, he found an association between type of complaint and duration of mar- riage, with a wife married for a short duration likely to complain of her husband's personality and a wife married for 15 or more years to com- plain of his drinking, adultery, and "helling around." Other studies make it clear that the divergence so commonly noted between the ac- counts of husbands and wives is not random. Fulton (1979) speaks of "his" and "her" divorce and found wives in a large sample more critical of the ex-partners and of the marriages, more likely to report arguments and violence, and less likely to feel that the marriages had been happy for much of the time. Levinger (1966) found wives more likely to complain of cruelty, drinking, physical and verbal abuse, neglect and lack of August 1984 JOURNAL OF MARRIAGE AND THE FAMILY 551 This content downloaded from 211.80.94.134 on Mon, 19 Dec 2016 05:27:55 UTC All use subject to http://about.jstor.org/terms
love:husbands of in-law problems and sexual in-represented,with 270 of men and 2690 of compatibility. women's husbands in the semiskilled,unskilled, The present study aims to extend this line of and unemployed categories,compared with an research by exploring the associations between the estimated 20.4 semi-and unskilled in the reasons advanced by a sample of men and women general Australian population.The proportion of for the failure of their marriages,and a number of Roman Catholics was lower than that in the structural and demographic variables.Specifical- general population but higher than the average ly,these are sex of respondent,marital status proportion of divorced persons categorized as (divorced or separated),age at marriage,number Catholics in annual divorce statistics.Moreover, of times married,religion,socioeconomic status, these were just as likely as Protestants to be marital status of parents and spouse's parents, divorced (rather than separated),and 910 of still- duration of marriage,and number of children of married Roman Catholic women were intending the marriage.The study also includes a group of divorce.Immigrants from non-English-speaking variables that are not structural or demographic countries,currently a significant minority in the but that have been shown to be associated with Australian population,were underrepresented,no proneness to divorce.These are parental attitude doubt partly due to the fact that the study's pub- to the marriage,length of prior acquaintance, licity was restricted to the English-language point at which the marriage was seen as "begin- media. ning to break down,''and perceived current stan- Altogether 390 of the women and 350 of the dard of living compared with that pertaining dur- men were divorced,the remainder of the sample ing the marriage (Goode,1956:Fulton,1979: being separated only (this distribution cor- Thornes and Collard.1979). responds roughly to that obtaining in the Australian population at the time of the study: DESIGN Burns,1980).The age range was from 21 to 60 The Sample years,with the largest single group of both men The data come from a volunteer sample of 233 and women being aged 31-40.In 7007 of the women and 102 men,divorced or separated,who cases,the divorce or separation was of less than three-years standing.Men were somewhat more responded to publicity in metropolitan and sub- urban newspapers in Sydney,Australia.This likely to describe their spouses as the departing party (6000)than were women(5200).This sug- method was chosen because court records in gests that,as might be expected,people who saw Australia are confidential and not available to re- themselves as the injured party were particularly searchers,and because this project lacked the re- sources to seek out a random sample of divorced likely to have joined the study;and there was some evidence that this was the case (Burns, and separated men and women from the general 1980) population.The publicity took the form of short feature pieces describing the study as an investiga- The Questionnaire tion into the experience of divorce and separation The questionnaire comprised five sections: in Australia and calling for volunteers.Due to background material,causes and timing of the lack of interviewing resources,a self-report for- breakdown,experiences with the courts and with mat was used,and the questionnaire was distri- counseling agencies,reaction of children,current buted by mail to all men and women who con- tacted us(N 412).Of these,335 returned com- life situation and attitudes.In the present analysis,only the section on causes and a selection pleted (although not always fully completed) questionnaires.It is not possible to state how of variables from other sections is included.To assess causes respondents were offered a checklist representative the sample is of the divorced and of 18 reasons,based on Goode's list (1956)and separated people in Australia,since the character- istics of this population are not known.In the asked to check all items that were relevant in their own cases.The checklist was followed by an absence of this information,a comparison with open-response item. the general Australian population was made in terms of occupation,religion,and country of origin.The occupational distribution of male re Treatment of Data spondents and of female respondents'husbands Open responses were coded into the data,and proved to be very similar and only slightly higher all causes were coded as dichotomous variables than that of the general population (Burns,1980). (yes 1,no =2).Perceived onset of breakdown Given the usual tendency of volunteer samples to was collapsed for the main analysis into a 4-point be of higher SES (Moser and Kalton,1971),the scale(within the first year of marriage,2-5 years, lower occupational range was unexpectedly well- 6-10 years,and 11 years or later).Duration of 552 JOURNAL OF MARRIAGE AND THE FAMILY August 1984 This content downloaded from 211.80.94.134 on Mon,19 Dec 2016 05:27:55 UTC All use subject to http://about.jstor.org/terms
love; husbands of in-law problems and sexual in- compatibility. The present study aims to extend this line of research by exploring the associations between the reasons advanced by a sample of men and women for the failure of their marriages, and a number of structural and demographic variables. Specifical- ly, these are sex of respondent, marital status (divorced or separated), age at marriage, number of times married, religion, socioeconomic status, marital status of parents and spouse's parents, duration of marriage, and number of children of the marriage. The study also includes a group of variables that are not structural or demographic but that have been shown to be associated with proneness to divorce. These are parental attitude to the marriage, length of prior acquaintance, point at which the marriage was seen as "begin- ning to break down," and perceived current stan- dard of living compared with that pertaining dur- ing the marriage (Goode, 1956; Fulton, 1979; Thornes and Collard, 1979). DESIGN The Sample The data come from a volunteer sample of 233 women and 102 men, divorced or separated, who responded to publicity in metropolitan and sub- urban newspapers in Sydney, Australia. This method was chosen because court records in Australia are confidential and not available to re- searchers, and because this project lacked the re- sources to seek out a random sample of divorced and separated men and women from the general population. The publicity took the form of short feature pieces describing the study as an investiga- tion into the experience of divorce and separation in Australia and calling for volunteers. Due to lack of interviewing resources, a self-report for- mat was used, and the questionnaire was distri- buted by mail to all men and women who con- tacted us (N = 412). Of these, 335 returned com- pleted (although not always fully completed) questionnaires. It is not possible to state how representative the sample is of the divorced and separated people in Australia, since the character- istics of this population are not known. In the absence of this information, a comparison with the general Australian population was made in terms of occupation, religion, and country of origin. The occupational distribution of male re- spondents and of female respondents' husbands proved to be very similar and only slightly higher than that of the general population (Burns, 1980). Given the usual tendency of volunteer samples to be of higher SES (Moser and Kalton, 1971), the lower occupational range was unexpectedly well- represented, with 27% of men and 26% of women's husbands in the semiskilled, unskilled, and unemployed categories, compared with an estimated 20.4% semi- and unskilled in the general Australian population. The proportion of Roman Catholics was lower than that in the general population but higher than the average proportion of divorced persons categorized as Catholics in annual divorce statistics. Moreover, these were just as likely as Protestants to be divorced (rather than separated), and 91% of still- married Roman Catholic women were intending divorce. Immigrants from non-English-speaking countries, currently a significant minority in the Australian population, were underrepresented, no doubt partly due to the fact that the study's pub- licity was restricted to the English-language media. Altogether 39% of the women and 35% of the men were divorced, the remainder of the sample being separated only (this distribution cor- responds roughly to that obtaining in the Australian population at the time of the study: Burns, 1980). The age range was from 21 to 60 years, with the largest single group of both men and women being aged 31-40. In 70% of the cases, the divorce or separation was of less than three-years standing. Men were somewhat more likely to describe their spouses as the departing party (60%) than were women (52%). This sug- gests that, as might be expected, people who saw themselves as the injured party were particularly likely to have joined the study; and there was some evidence that this was the case (Burns, 1980). The Questionnaire The questionnaire comprised five sections: background material, causes and timing of the breakdown, experiences with the courts and with counseling agencies, reaction of children, current life situation and attitudes. In the present analysis, only the section on causes and a selection of variables from other sections is included. To assess causes respondents were offered a checklist of 18 reasons, based on Goode's list (1956) and asked to check all items that were relevant in their own cases. The checklist was followed by an open-response item. Treatment of Data Open responses were coded into the data, and all causes were coded as dichotomous variables (yes = 1, no = 2). Perceived onset of breakdown was collapsed for the main analysis into a 4-point scale (within the first year of marriage, 2-5 years, 6-10 years, and 11 years or later). Duration of 552 JOURNAL OF MARRIAGE AND THE FAMILY August 1984 This content downloaded from 211.80.94.134 on Mon, 19 Dec 2016 05:27:55 UTC All use subject to http://about.jstor.org/terms
TABLE 1.PERCEIVED CAUSES OF MARRIAGE BREAKDOWN 0%of All Respondents 0%Men 0%Women Causesa (n=335) (n▣102) (n=233) Sexual incompatibility 45 56 40 Lack of communicationb 40 40 Husband's lack of time at home 40 28 46 Financial 2 Husband's association with another womanc 31 sandru 30 26 39 Wife's lack of interestt 25 Friction with relatives 23 21 Disagreements over children 0 Wife's association with angther manc 19 Husband's lack of interest 13 Wife's ill health 13 22 Inadequate housing 9 3 Religious differences Husband's gambling 5 3 Note:Percentages do not add up to 100.Multiple complaints are included. aCauses mentioned by less than 5 are not included. bIncludes lack of common interests. and mentioncd. eIncludes wife's perception of husband as having sadistic, cruel,or brutal personality. fincludes statement of resentment about the lack of stimulation. marriage was coded as a 6-point scale (2,2-5, ate analyses,but probability values between .01 6-10,11-15,16-20 and 20+years).Other vari- and,05 are reported in Table 2 as indicating inter- ables included in the analysis were sex of respon- esting trends.Subsidiary analyses include correla- dent;marital status of respondent (divorced or tions and an exploratory factor analysis (BMD separated);whether the marriage and separation P4M:Principal Components DQUART,non- being described was the first or not;number of orthogonal rotation). children;parental attitudes to the marriage (strongly approve,approve,neutral [including don't know],disapprove,or strongly disapprove); RESULTS length of acquaintance prior to marriage (7-point scale:0-3,3-6,6-12 months,12 months to 2 Sex Differences in Perceived Causes years,2,3-5,and 5+years).Respondents'evalu- of the Breakdown ations of their present standards of living was Most respondents indicated that the breakdown coded as higher,similar to,somewhat lower,or had multiple causes.The most commonly nomi- much lower than that obtaining during the mar- nated causes are presented in Table 1.The most riage.Socioeconomic status was rated on a frequent complaint is sexual incompatibility, 7-point scale (Congalton,1969)using (ex-)hus- which was mentioned by 56%of the men and band's occupation,with an extra category in- 40%of the women.The next most common com- cluded for unemployed/retired husbands.Reli- plaint is lack of common interests/lack of com- gion was collapsed into Protestant,Roman munication.With respect to other complaints, Catholic,and nonbeliever,with the small number there is somewhat less agreement between of persons of other faiths excluded from these husbands and wives,with wives more likely to analyses.Parental marriages were coded as intact, mention the husband's lack of time at home and divorced,separated,or widowed. his adultery,cruelty,and drinking,while Analysis husbands focused on the wife's adultery and fric- tion with relatives.It is worth noting,however, A series of SPSS MANOVA were run,with the that over one-fourth of the men mention their 15 most commonly nominated causes (plus an own lack of time at home,the top complaint "other",category)treated as the outcome set. among the wives.The sex difference is highly sig- Where data were missing or could not be fitted in- nificant (F=7.78,df=16,311,p<.000:Table to the coding scheme,these cases were dropped 2,column 1).The univariate analyses show that from that particular analysis.As a result,degrees the significant differences occur on complaints of of freedom vary between analyses.In view of the husband's adultery,drinking,cruelty,and lack of large number of variables,a conservative proba- time at home;housing and financial problems bility figure (p =.01)is adopted for the univari- (wives nominated these more frequently);and August 1984 JOURNAL OF MARRIAGE AND THE FAMILY 553 This content downloaded from 211.80.94.134 on Mon,19 Dec 2016 05:27:55 UTC All use subject to http://about.jstor.org/terms
I -DLL 1. rt-Lh\.I v AL/UJ ` UL r IVIALIAII/-\jI DI__lAiLWJJV VV IN % of All Respondents % Men % Women Causesa (n = 335) (n = 102) (n = 233) Sexual incompatibility 45 56 40 Lack of communicationb 40 41 40 Husband's lack of time at home 40 28 46 Financial 32 24 36 Husband's association with another womanc 31 17 37 Husband's drinkingd 30 17 36 Husband's crueltye 26 37 4 Wife's lack of interestf 26 25 26 Friction with relatives 23 29 21 Disagreements over children 20 22 19 Wife's association with anQther manC 19 35 12 Husband's lack of interest1 13 15 12 Wife's ill health 13 13 13 Inadequate housing 9 4 13 Religious differences 5 5 5 Husband's gambling 5 3 7 Note: Percentages do not add up to 100. Multiple complaints are included. aCauses mentioned by less than 5% are not included. bIncludes lack of common interests. CNo homosexual relationships were mentioned. dlncludes husband's alcoholism. eIncludes wife's perception of husband as having sadistic, cruel, or brutal personality. fIncludes statement of resentment about the lack of stimulation. marriage was coded as a 6-point scale (< 2, 2-5, 6-10, 11-15, 16-20 and 20+ years). Other vari- ables included in the analysis were sex of respon- dent; marital status of respondent (divorced or separated); whether the marriage and separation being described was the first or not; number of children; parental attitudes to the marriage (strongly approve, approve, neutral [including don't know], disapprove, or strongly disapprove); length of acquaintance prior to marriage (7-point scale: 0-3, 3-6, 6-12 months, 12 months to 2 years, 2, 3-5, and 5 + years). Respondents' evalu- ations of their present standards of living was coded as higher, similar to, somewhat lower, or much lower than that obtaining during the mar- riage. Socioeconomic status was rated on a 7-point scale (Congalton, 1969) using (ex-) hus- band's occupation, with an extra category in- cluded for unemployed/retired husbands. Reli- gion was collapsed into Protestant, Roman Catholic, and nonbeliever, with the small number of persons of other faiths excluded from these analyses. Parental marriages were coded as intact, divorced, separated, or widowed. Analysis A series of SPSS MANOVA were run, with the 15 most commonly nominated causes (plus an "other" category) treated as the outcome set. Where data were missing or could not be fitted in- to the coding scheme, these cases were dropped from that particular analysis. As a result, degrees of freedom vary between analyses. In view of the large number of variables, a conservative proba- bility figure (p = .01) is adopted for the univari- ate analyses, but probability values between .01 and .05 are reported in Table 2 as indicating inter- esting trends. Subsidiary analyses include correla- tions and an exploratory factor analysis (BMD P4M: Principal Components DQUART, non- orthogonal rotation). RESULTS Sex Differences in Perceived Causes of the Breakdown Most respondents indicated that the breakdown had multiple causes. The most commonly nomi- nated causes are presented in Table 1. The most frequent complaint is sexual incompatibility, which was mentioned by 56%o of the men and 407o of the women. The next most common com- plaint is lack of common interests/lack of com- munication. With respect to other complaints, there is somewhat less agreement between husbands and wives, with wives more likely to mention the husband's lack of time at home and his adultery, cruelty, and drinking, while husbands focused on the wife's adultery and fric- tion with relatives. It is worth noting, however, that over one-fourth of the men mention their own lack of time at home, the top complaint among the wives. The sex difference is highly sig- nificant (F = 7.78, df = 16,311,p < .000: Table 2, column 1). The univariate analyses show that the significant differences occur on complaints of husband's adultery, drinking, cruelty, and lack of time at home; housing and financial problems (wives nominated these more frequently); and JOURNAL OF MARRIAGE AND THE FAMILY T A UFIT 1 D r[1> ' r A' T A TCE'C f' A NAD I A I".R DD A VNrAn\ i'XT August 1984 553 This content downloaded from 211.80.94.134 on Mon, 19 Dec 2016 05:27:55 UTC All use subject to http://about.jstor.org/terms
wife's adultery and sexual incompatibility with the 1:2 and 1:3 contrast significant.Hus- (husbands nominated these more frequently). band's cruelty is concentrated in period 1,and the contrast between this and all other onset periods is Perceived Onset of the Breakdown significant. Both men and women,but particularly women, perceived the marriage as beginning to break Socioeconomic Status down early.Fifteen percent of the wives dated the SES also was significantly associated with per- onset of breakdown within the first three months ceived causes of breakdown (F 1.59,df= of marriage.Most couples,however,did not sepa- 16,272,p =.023),the two significant contribut- rate until much later (Figure).The sex difference ing causes being husband's drinking (p =.011) is highly significant (F=2.47,df=16,319,p< and husband's cruelty (p<.000).The relation- .001). ship is an inverse one.Only 170 of women mar- Were particular problems associated with ried to men in higher professional or managerial earlier or later recognition that the marriage was occupations complained about alcohol,whereas failing?Table 2(column 3)shows that they were 70%%of women married to unskilled men and all (multivariate F 2.47,df=16,319,p<.000). women whose husbands were unemployed or re- The univariate analyses show that five causes are tired made this complaint.The trends are vaguer associated with particular breakdown periods: among the smaller number of men (13)who saw husband's drinking (p <.000),husband's cruelty their own drinking as a problem,but none of (p<.000),sexual incompatibility (p =.004),dis- these were in the two top-SES categories.Com- agreements over children (p =.014),and an plaints of husband's cruelty were most common “other woman”(p<.032).Mean values(Table among wives of men in the semiskilled(470%)un- 3)indicate that complaints of sexual incompatibil- skilled (57%)and unemployed/retired categories ity rise from a high base in onset period 1 (0-12 (50%).(It should be borne in mind that cruelty months)to a peak in period 3(6-10 years),fol- was self-defined by respondents and included be- lowed by a sharp decline in period 4(11+years). haviors other than physical cruelty.There was The contrast between periods 1:4 and 3:4 was sig- also a relationship between SES and perceived nificant beyond the.05 level.The "other onset of the breakdown,with semiskilled and un- woman"'cause rises steadily from period 1 skilled men being particularly likely to attribute through to period 4,with the gains from 1:3 and the breakdown to their wife's adultery in the 1:4 being significant.Disagreements over children period between 2 and 10 years after marriage. are concentrated in period 2(2-5 years),but only the contrast between periods 2:4 is significant. Religion Husband's drinking is most commonly mentioned The multivariate F value for respondent's reli- by those who nominate onset periods 1 and 4, gion is not significant,but there is a trend towards significance with respect to two causes,husband's lack of time at home (p=.024)and husband's FIGURE.PERCEIVED ONSET OF BREAKDOWN drinking (p =.041).In both cases the complaint AND TIME TO SEPARATION was most common among Roman Catholics and least common among those with no religious affil- Per Cent iation.When the sample is divided by sex,it 40 -----Onset Women emerges that religious beliefs have an opposite ef- -Onset Men fect on men and women with respect to sexual Time to Separation complaints (p =.003).Women with no religious 30 affiliation were particularly likely to see sexual in- compatibility as a cause of the marriage break- down,whereas it was the men with a religious af- filiation (of any kind)who were likely to make 20 this attribution.Spouses'religion was not a sig- nificant influence,although the univariate values show that lack of communication is somewhat 10 more commonly complained of by those with a Roman Catholic spouse,least by those with one of no religious affiliation (p =.049).The interac- tion between spouse's religion and sex of respon- dent,however,is highly significant (F=1.79, 1 Year 5 Years 10 Years 15 Years 20+Years df=16,306,p =.005),and four complaints 554 JOURNAL OF MARRIAGE AND THE FAMILY August 1984 This content downloaded from 211.80.94.134 on Mon,19 Dec 2016 05:27:55 UTC All use subject to http://about.jstor.org/terms
wife's adultery and sexual incompatibility (husbands nominated these more frequently). Perceived Onset of the Breakdown Both men and women, but particularly women, perceived the marriage as beginning to break down early. Fifteen percent of the wives dated the onset of breakdown within the first three months of marriage. Most couples, however, did not sepa- rate until much later (Figure). The sex difference is highly significant (F = 2.47, df = 16,319, p < .001). Were particular problems associated with earlier or later recognition that the marriage was failing? Table 2 (column 3) shows that they were (multivariate F = 2.47, df = 16,319, p < .000). The univariate analyses show that five causes are associated with particular breakdown periods: husband's drinking (p < .000), husband's cruelty (p < .000), sexual incompatibility (p = .004), dis- agreements over children (p = .014), and an "other woman" (p < .032). Mean values (Table 3) indicate that complaints of sexual incompatibil- ity rise from a high base in onset period 1 (0-12 months) to a peak in period 3 (6-10 years), fol- lowed by a sharp decline in period 4 (11 + years). The contrast between periods 1:4 and 3:4 was sig- nificant beyond the .05 level. The "other woman" cause rises steadily from period 1 through to period 4, with the gains from 1:3 and 1:4 being significant. Disagreements over children are concentrated in period 2 (2-5 years), but only the contrast between periods 2:4 is significant. Husband's drinking is most commonly mentioned by those who nominate onset periods 1 and 4, FIGURE. PERCEIVED ONSET OF BREAKDOWN AND TIME TO SEPARATION Per Cent 40 - -- Onse Onse - - Time 30 - 20 - 10 - \ K\ J 1 Year 5 Years 10 Years 15 YE t Women t Men to Separation i ears 20+ Years with the 1:2 and 1:3 contrast significant. Hus- band's cruelty is concentrated in period 1, and the contrast between this and all other onset periods is significant. Socioeconomic Status SES also was significantly associated with per- ceived causes of breakdown (F = 1.59, df = 16,272, p = .023), the two significant contribut- ing causes being husband's drinking (p = .011) and husband's cruelty (p < .000). The relation- ship is an inverse one. Only 17070 of women mar- ried to men in higher professional or managerial occupations complained about alcohol, whereas 700% of women married to unskilled men and all women whose husbands were unemployed or re- tired made this complaint. The trends are vaguer among the smaller number of men (13) who saw their own drinking as a problem, but none of these were in the two top-SES categories. Com- plaints of husband's cruelty were most common among wives of men in the semiskilled (47%0) un- skilled (57%0) and unemployed/retired categories (500o). (It should be borne in mind that cruelty was self-defined by respondents and included be- haviors other than physical cruelty.) There was also a relationship between SES and perceived onset of the breakdown, with semiskilled and un- skilled men being particularly likely to attribute the breakdown to their wife's adultery in the period between 2 and 10 years after marriage. Religion The multivariate F value for respondent's reli- gion is not significant, but there is a trend towards significance with respect to two causes, husband's lack of time at home (p = .024) and husband's drinking (p = .041). In both cases the complaint was most common among Roman Catholics and least common among those with no religious affil- iation. When the sample is divided by sex, it emerges that religious beliefs have an opposite ef- fect on men and women with respect to sexual complaints (p = .003). Women with no religious affiliation were particularly likely to see sexual in- compatibility as a cause of the marriage break- down, whereas it was the men with a religious af- filiation (of any kind) who were likely to make this attribution. Spouses' religion was not, a sig- nificant influence, although the univariate values show that lack of communication is somewhat more commonly complained of by those with a Roman Catholic spouse, least by those with one of no religious affiliation (p = .049). The interac- tion between spouse's religion and sex of respon- dent, however, is highly significant (F = 1.79, df = 16,306, p = .005), and four complaints 554 JOURNAL OF MARRIAGE AND THE FAMILY August 1984 This content downloaded from 211.80.94.134 on Mon, 19 Dec 2016 05:27:55 UTC All use subject to http://about.jstor.org/terms
181 1000.1 月 得 娟 3 s001 绳 的 鲤 2001 3 的 好 娟 S3S 理 娟 系 导3 3 3 3 得 剪 是 爱 60o 000:) 29.g (10) (ZO) EE'IZ 000.1 15S (600) uonestunwwo sasne uewoM ay1O ButsnoH aouasqe s,pueqsnH zuques s pueqsnH Al[anIo s,pueqsnH salolu!s,pueqsnH [epueut August 1984 JOURNAL OF MArriage AND THE FAMILY 555 This content downloaded from 211.80.94.134 on Mon,19 Dec 2016 05:27:55 UTC All use subject to http://about.jstor.org/terms
TABLE 2. PERCEIVED CAUSES OF BREAKDOWN AND DEMOGRAPHIC/STRUCTURAL VARIABLES Onset Onset Religion Spouse's Age at x x x Spouse's Religion Mar- Causes Sex Onset Sex SES SES Religion Sex Religion x Sex riage Multivariate F (p) 7.78 2.47 1.58 1.59 ns ns ns ns 1.79 1.87 (.000) (.000) (.008) (.023) (.55) (.56) (.14) (.27) (.005) (.000) Univariate F (p) Sexual Communication 4.31 4.56 (.039) (.004) Other man 24.9 2.49 (.000) (.06) Other woman 15.27 2.96 6.80 (.000) (.032) (.000) Housing 6.63 (.01) Children 3.62 (.014) Husband's absence 5.9 (.02) Husband's drinking 21.33 7.38 (.000) (.000) Husband's gambling 8.34 (.004) Husband's cruelty 39.37 12.38 (.000) (.000) Wife's interest Husband's interest Financial 4.58 (.03) Wife's ill health Other 5.83 (.003) 3.05 (.049) 2.85 (.011) 3.33 (.036) 3.13 2.72 (.045) (.03) 3.20 (.042) 7.61 (.000) 3.76 (.024) 3.32 (.04) 4.57 (.011) 9.46 (.000) 3.95 (.02) 3.08 (.016) 3.85 (.004) 4.12 (.03) (Continued on next page) Lfn (A Relatives C-4 0 z CT1 0 tTI Hoz C"1 t3 0-= 3, r;3 "Ti This content downloaded from 211.80.94.134 on Mon, 19 Dec 2016 05:27:55 UTC All use subject to http://about.jstor.org/terms