# Determinants of Divorce (GSS)

Using the General Social Survey data, I try to investigate which factor is the most determinant of divorce. For this purpose, I use the logistic regression. (note : I have updated the analysis.)

Below, I display the variables used in the regressions.

DIVORCE. If currently married or widowed: Have you ever been divorced or legally separated? 1=Yes. 2=No. (These are the initial values. I reversed them, so that higher values means “have been divorced”.)

DIVORCE5. Divorces during the last five years. 0 = None, 1 = Prior 4 years, 2 = Last year, 3 = Both. (Recoded as follows in SDA program : DIVORCE5 (d:1-3). The dummy variable created means that the cases coded 1 through 3 receive a code of 1. In other words, case coded 0 receives a code of 0.)

EVSTRAY. Have sex other than spouse while married. 1=yes. 2=no. 3=not married. (I have suppressed the value 3 and reversed 1 with 2, so that the higher value in the variable means “having sex while married”.)

SEX. 1 = male, 2 = female.

WORDSUM. Vocabulary test (a proxy for IQ, correlation = 0.71; 0.83 for g). Should not be taken as a measure of general intelligence however. It is a ten item/question variable, having a rather low reliability, of about 73. See “Reliability and Stability Estimates for the GSS Core Items from the Three-wave Panels, 2006–2010” (Michael Hout & Orestes P. Hastings, 2012).

REALINC. Family income on 1972-2006 surveys in constant dollars (base = 1986).

POLVIEWS. 1 = Extremely liberal, 4 = Moderate, 7 = Extremely conservative. We hear a lot of talk these days about liberals and conservatives. I’m going to show you a seven-point scale on which the political views that people might hold are arranged from extremely liberal – point 1 – to extremely conservative – point 7. Where would you place yourself on this scale?

ATTEND. 0 = Never, 8 = More than once week. How often do you attend religious services?

AGE. Respondent’s age.

RACE. 1 = White, 2 = Black, 3 = Other.

Year. Year of survey.

The first table shows the results with DIVORCE (N=13204) as dependent, and the second table shows the results with DIVORCE5 (N=6194) as dependent variable. With DIVORCE, being white (versus black) is negatively associated with divorce and there is no gender effect. However, with DIVORCE5, there is no black-white difference, and women are more likely to divorce than men. In both DIVORCE and DIVORCE5, wordsum and political views play no role. Being religious is negatively associated with divorce. Being rich decreases the likelihood of being divorced, only if we use DIVORCE5.

Now, when I restrict the sample to whites only, the results are similar, so I won’t display them. For black sample, Wordsum has a large positive coefficient for DIVORCE5 but the effect is weak for DIVORCE. Furthermore, the negative effect of being religious (on divorce) is less for the black sample both DIVORCE and DIVORCE5.

Next is another logistic regression with extramarital_affair as dependent variable. First table is for blacks (N=780) and second table for whites (N=6609).

Generally, conservative and religious people tend to have less extramarital sex. Also, as expected, females are less likely than males to have extramarital sex. The only difference between the black and white samples is that intelligence is associated with more extramarital sex among whites.

How to interpret all this ? My guess is that people who are atheist or liberal are seeking novelty, excitement, or new experiences since atheists and liberals are more likely to break conventional traditions. Men are usually known to be far more adventurous than are women. An article entitled “Cybersex: The New Affair Treatment Considerations” states that :

While percentages of extramarital sex (EMS) vary from study to study, it can be estimated that 50-60% of married men and 45-55% of married women engage in extramarital sex at some time or another during their marriage and almost half come to therapy because of it. EMS appears in several different forms, only some of which are sexual in nature. Recently, couples are coming to therapy with a new type of affair: the Cyber-Affair. According to the President of the American Academy of Matrimonial Lawyers, this type of Internet infidelity has been greatly underestimated due to the Internet’s current popularity as a technological advancement. In addition, healthcare professionals are often unfamiliar with the dynamics associated with the relatively new concept of cyber-affairs and the electronic process of “virtual cheating” and thus often do not consider the behavior as infidelity.

Aren’t modern societies ashamed enough ? Should they dig holes even deeper ?

Syntax :

RECODE divorce (1=1) (2=0) INTO DIVORCE_DICHOTOMY.
EXECUTE.
RECODE divorce5 (0=0) (1 thru highest=1) INTO DIVORCE5_DICHOTOMIZED.
EXECUTE.

RECODE evstray (1=2) (2=1) (ELSE=SYSMIS) INTO Extramarital_Affair.
EXECUTE.

RECODE race (1=2) (2=1) (ELSE=SYSMIS) INTO BW.
EXECUTE.

RECODE year (1991 thru 1995=1) (1996 thru 1999=2) (2000 thru 2003=3) (2004 thru 2007=4) (2008 thru 2012=5) INTO Year5.
EXECUTE.

COMPUTE SQRTrealinc=SQRT(realinc).
VARIABLE LABELS SQRTrealinc ‘square root of R income in constant dollars’.
EXECUTE.

DESCRIPTIVES VARIABLES=age year COHORT WORDSUM SEI realinc SQRTrealinc POLVIEWS ATTEND
/SAVE
/STATISTICS=MEAN STDDEV MIN MAX.

FREQUENCIES VARIABLES=Zsei Zrealinc SQRTrealinc ZSQRTrealinc Zage Zyear Zwordsum Zpolviews Zattend
/FORMAT=NOTABLE
/HISTOGRAM NORMAL
/ORDER=ANALYSIS.

COMPUTE wtssall_oversamp=wtssall*oversamp.
EXECUTE.

WEIGHT BY wtssall_oversamp.

LOGISTIC REGRESSION VARIABLES DIVORCE_DICHOTOMY
/METHOD=ENTER BW sex Zage Zyear Zwordsum Zrealinc Zpolviews Zattend
/CLASSPLOT
/PRINT=GOODFIT CORR ITER(1) CI(95)
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

LOGISTIC REGRESSION VARIABLES DIVORCE5_DICHOTOMIZED
/METHOD=ENTER BW sex Zage Zyear Zwordsum Zrealinc Zpolviews Zattend
/CLASSPLOT
/PRINT=GOODFIT CORR ITER(1) CI(95)
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

USE ALL.
COMPUTE filter_\$=(BW=1).
VARIABLE LABELS filter_\$ ‘BW=1 (FILTER)’.
VALUE LABELS filter_\$ 0 ‘Not Selected’ 1 ‘Selected’.
FORMATS filter_\$ (f1.0).
FILTER BY filter_\$.
EXECUTE.

LOGISTIC REGRESSION VARIABLES DIVORCE_DICHOTOMY
/METHOD=ENTER sex Zage Zyear Zwordsum Zrealinc Zpolviews Zattend
/CLASSPLOT
/PRINT=GOODFIT CORR ITER(1) CI(95)
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

LOGISTIC REGRESSION VARIABLES DIVORCE5_DICHOTOMIZED
/METHOD=ENTER sex Zage Zyear Zwordsum Zrealinc Zpolviews Zattend
/CLASSPLOT
/PRINT=GOODFIT CORR ITER(1) CI(95)
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

WEIGHT BY wtssall.

USE ALL.
COMPUTE filter_\$=(BW=2).
VARIABLE LABELS filter_\$ ‘BW=2 (FILTER)’.
VALUE LABELS filter_\$ 0 ‘Not Selected’ 1 ‘Selected’.
FORMATS filter_\$ (f1.0).
FILTER BY filter_\$.
EXECUTE.

LOGISTIC REGRESSION VARIABLES DIVORCE_DICHOTOMY
/METHOD=ENTER sex Zage Zyear Zwordsum Zrealinc Zpolviews Zattend
/CLASSPLOT
/PRINT=GOODFIT CORR ITER(1) CI(95)
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

LOGISTIC REGRESSION VARIABLES DIVORCE5_DICHOTOMIZED
/METHOD=ENTER sex Zage Zyear Zwordsum Zrealinc Zpolviews Zattend
/CLASSPLOT
/PRINT=GOODFIT CORR ITER(1) CI(95)
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

FILTER OFF.
USE ALL.
EXECUTE.

WEIGHT BY wtssall_oversamp.

LOGISTIC REGRESSION VARIABLES Extramarital_Affair
/METHOD=ENTER BW sex Zage Zyear Zwordsum Zrealinc Zpolviews Zattend
/CLASSPLOT
/PRINT=GOODFIT CORR ITER(1) CI(95)
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

USE ALL.
COMPUTE filter_\$=(BW=1).
VARIABLE LABELS filter_\$ ‘BW=1 (FILTER)’.
VALUE LABELS filter_\$ 0 ‘Not Selected’ 1 ‘Selected’.
FORMATS filter_\$ (f1.0).
FILTER BY filter_\$.
EXECUTE.

LOGISTIC REGRESSION VARIABLES Extramarital_Affair
/METHOD=ENTER sex Zage Zyear Zwordsum Zrealinc Zpolviews Zattend
/CLASSPLOT
/PRINT=GOODFIT CORR ITER(1) CI(95)
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

UNIANOVA Extramarital_Affair BY sex Year5
/METHOD=SSTYPE(3)
/INTERCEPT=INCLUDE
/PLOT=PROFILE(Year5*sex)
/EMMEANS=TABLES(sex)
/EMMEANS=TABLES(Year5)
/EMMEANS=TABLES(sex*Year5)
/PRINT=LOF OPOWER ETASQ HOMOGENEITY DESCRIPTIVE
/CRITERIA=ALPHA(.05)
/DESIGN=sex Year5 sex*Year5.

WEIGHT BY wtssall.

USE ALL.
COMPUTE filter_\$=(BW=2).
VARIABLE LABELS filter_\$ ‘BW=2 (FILTER)’.
VALUE LABELS filter_\$ 0 ‘Not Selected’ 1 ‘Selected’.
FORMATS filter_\$ (f1.0).
FILTER BY filter_\$.
EXECUTE.

LOGISTIC REGRESSION VARIABLES Extramarital_Affair
/METHOD=ENTER sex Zage Zyear Zwordsum Zrealinc Zpolviews Zattend
/CLASSPLOT
/PRINT=GOODFIT CORR ITER(1) CI(95)
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

UNIANOVA Extramarital_Affair BY sex Year5
/METHOD=SSTYPE(3)
/INTERCEPT=INCLUDE
/PLOT=PROFILE(Year5*sex)
/EMMEANS=TABLES(sex)
/EMMEANS=TABLES(Year5)
/EMMEANS=TABLES(sex*Year5)
/PRINT=LOF OPOWER ETASQ HOMOGENEITY DESCRIPTIVE