Margo Wilson’s Influence Continues

Background

The first author of this article originally met Margo Wilson in 1992 at the Second Annual Workshop of the Homicide Research Working Group on the FBI’s Quantico, Virginia campus. Margo presented a paper, “A Lifespan Perspective on Homicide Violence: The Young Male Syndrome,” coauthored with her life partner and collaborator, Martin Daly. As an early PhD student in sociology, this was my first exposure to Wilson and Daly’s seminal work on the evolutionary psychological perspective of violence. As fully elaborated on in their 1988 book, Homicide, this explanation of lethal violence is compelling. From this first meeting, I considered it a privilege to be professionally acquainted with Margo.

I was particularly interested in the work that Margo Wilson and Martin Daly published on the elevated risk of uxoricide (wife-killing) in Hamilton-Wentworth, Ontario for women with coresiding minor children who were not the biological offspring of their current partners (Daly, Wiseman, & Wilson, 1997). Margo was extremely helpful to me with both substantive and editorial advice as I and my coauthor worked to replicate the Canadian analysis of uxoricide for the city of Houston, Texas. We were ultimately successful in doing so (Brewer & Paulsen, 1999) and I felt quite indebted to Margo for her willingness to serve as an informal mentor for the project. She was always exceptional in her attentiveness to the “new-kids-on-the-block” as well as her seasoned colleagues.

The Present Study

The research herein is yet another extension of Wilson and Daly’s work, describing the overall and intimate partner SROKs over a 31-year period for Texas’ six largest cities: Austin, Dallas, El Paso, Fort Worth, Houston, and San Antonio (Figure 1). During the 1976-2007 time period, 10.1% of all U.S. homicides occurred in Texas. In turn, these six cities are among the 20 most populous in the United States, with a combined 2010 population of more than 6.8 million. Since 1993, these cities have constituted the Executive Issues Major Cities Program of the Police Research Center (PRC) within the College of Criminal Justice of Sam Houston State University. The PRC generates research, consultation, and training with these six cities and their chiefs of police.

OPEN IN VIEWER

Figure 1.

Texas Six Largest Cities and U.S. Rank by Population

Method

The unit of analysis in this study is incidents of murder and nonnegligent homicide, utilizing available data from a composite Uniform Crime Reports: Supplementary Homicide Reports (SHR) File for the years 1976 to 2007 (Fox & Swatt, 2009). The file provides incident-level information on criminal homicides in the United States, including location, circumstances, and method of offense as well as demographic characteristics of victims and offenders and the relationship between the two. The data are based on monthly reports to the Federal Bureau of Investigation (FBI) by local law enforcement agencies participating in the FBI’s Uniform Crime Reporting Program. To account for nonreported data (missing cases), a weighting scheme linked to FBI annual estimates of homicide counts by states and National Center for Health Statistics (NCHS) mortality data on decedent characteristics in coroners’ reports for deaths classified as homicide was adopted. For incomplete case data (e.g., missing characteristics on victims), a multiple imputation process based on a log-linear model for incomplete multivariate categorical data was implemented (Fox & Swatt, 2009).

It should be noted that Wilson and Daly’s analysis of spousal SROKs used city-specific homicide files, including historical relationship data. Their classification of spousal relationships included couples who, at the time of the homicide or previously, had a shared residence and sexual intimacy. This included registered and de facto current or former marriages and those boyfriend–girlfriend relationships where there was explicit evidence of current or past coresidency (Wilson & Daly, 1992). In keeping with subsequent studies (Gauthier & Bankston, 1997, 2004; Paulsen & Brewer, 2000) the present analysis is of the broader concept of intimate partner SROKs. That said, our classification of intimate partner homicides includes offenders and victims who, at the time of the killing, were married, common-law or divorced couples, or who were reported by the police as boyfriend–girlfriend or in homosexual relationships. Despite the lack of historical information on coresidency or sexual intimacy of those victim–offender relationships reported as boyfriend–girlfriend or homosexual, we reason that those not thought to be intimate in nature would have been categorized by the police as friends.

In the SHR 1976 to 2007 file there were 32 victim–relationship categories, recoded into five for purposes of this analysis: Intimate, Family, Friend, Acquaintance, and Stranger. For the present analysis, those victim–offender relationships recorded by the police as current or former, registered or common-law spouses, boyfriends/girlfriends or homosexual partners were designated as Intimate; all other familial categories (including step- and in-law relationships) were coded as Family; the Friend category remained as such; coworkers and other known relationships were coded as Acquaintances and the remaining stranger or unknown categories were coded as Stranger. We recognize that combining these last two categories is somewhat arbitrary but do not believe it detracts from the primary question at hand related to homicides between intimate partners.

The composite SHR file used for this analysis does not contain a category for ethnicity; so that, we are compelled to limit our examination to offender and victim race, recorded by the police as White (including Hispanic White), Black, and Other. This 31-year file contained 17 original weapon categories, recoded into three for this study. All long-guns and handguns were classified as Guns, the SHR Knife category remained the same and the remaining categories were designated as Other. This Other category included poison, explosives, fire, drugs, drowning, and asphyxiation.

A final factor we take into consideration in this study is the ages of offenders and victims, with an interest in how this affects intimate partner SROKs. Although the composite SHR data file includes a category for victims age 65 and above, the level of aggregation for older offenders is age 50 and older. Therefore, we employed the 50+ age grouping for both Intimate partner homicide offenders and victims and compared this to all offenders and victims younger than age 50.

After determining frequencies for each of the variables included in this analysis, chi-square tests of significance were employed for within-city differences in offender and victim sex, race, and age, as well as weapon used.

While we recognize the limitations of the SHR, we believe it continues to be the best single source of homicide data that includes the victim–offender relationship for the six cities of interest in this study. What is gained by this exploratory analysis is a more complete estimate of victim–offender relationships, victim and offender sex, age and race, and weapon used. What is lost is the detail regarding intimate partner forms (e.g., registered versus de facto relationships), victim and offender ethnicity, and circumstantial details of the killing. Still, the SHR are useful in covering a longer time span and allowing the generation of city-specific SROK profiles for a specific state; information that has heretofore been unexplored.

Results

The Role of Relationship Type in SROKs

Table 1 displays the overall and relationship-specific sex ratios of killing (SROKs) for the United States, Texas, and Texas’ six largest cities for the period from 1976 through 2007. Based on the SHR imputation accounting for all reported homicide and nonnegligent manslaughter victims in the respective categories, it is first remarkable that the overall SROKs range from only 10 to 14 (Table 1) for the United States, the state of Texas, and the six largest cities within this state. This means that for every 100 men who killed women, 10 to 14 women killed men. However, the picture changes considerably when the SROKs are disaggregated by victim–offender relationship. For the U.S. overall during this 30-plus year period, almost half as many women as men killed intimate partners, as indicated by the SROK of 48 (Table 1). This ratio is lower than that reported in previous research for a shorter time period (Wilson & Daly, 1992), but is almost four times the overall SROK. Compared to this national figure, it is also the case that women’s representation as intimate partner homicide offenders is significantly higher for Texas, with a statewide intimate partner SROK of 62. However, the range of intimate partner SROKS varies widely across the cities in this analysis, from a low of 24 in El Paso to a high of 82 in Houston.

OPEN IN VIEWER

Table 1.

Overall and Relationship-Specific Sex Ratios of Killing (SROKs) for the United States, Texas, and Texas’ Six Largest Cities, 1976-2007

United States

United States				Texas
Offender				Offender
	Man	Woman	SROK		Man	Woman	SROK
Intimate	53,335	25,846	48	Intimate	5,264	3,242	62
Family	50,931	13,628	27	Family	5,316	1,313	25
Acquaintance	265,725	19,777	7	Acquaintance	26,207	2,066	8
Stranger	133,145	5,739	4	Stranger	13,354	591	4
Overall	503,136	64,990	13	Overall	50,141	7,212	14
Austin				Dallas
Offender				Offender
	Man	Woman	SROK		Man	Woman	SROK
Intimate	111	49	44	Intimate	550	433	78
Family	85	27	32	Family	691	176	25
Acquaintance	599	48	8	Acquaintance	4,011	378	9
Stranger	240	11	5	Stranger	2,677	132	5
Overall	1,035	135	13	Overall	7,929	1,119	14
El Paso				Fort Worth
Offender				Offender
	Man	Woman	SROK		Man	Woman	SROK
Intimate	111	27	24	Intimate	243	166	68
Family	92	34	37	Family	244	65	27
Acquaintance	391	19	5	Acquaintance	1526	125	8
Stranger	250	8	3	Stranger	765	28	4
Overall	844	88	10	Overall	2778	384	14
Houston				San Antonio
Offender				Offender
	Man	Woman	SROK		Man	Woman	SROK
Intimate	801	659	82	Intimate	321	180	56
Family	805	209	27	Family	386	100	26
Acquaintance	5,249	410	8	Acquaintance	2,273	142	6
Stranger	3,724	146	4	Stranger	1,196	52	4
Overall	10,579	1,424	13	Overall	4,176	474	11

Note; (p<.05)

Offender Race and Weapon in Intimate Partner SROKs

As shown in Table 2, a range of 86% to 97% of these intimate partner SROKs were classified as intraracial, though Hispanic and non-Hispanic Whites are not disaggregated in these data. As of 2000, approximately 60% of San Antonio’s population was of Hispanic or Latino origin, and this was the ethnic classification for more than 80% of El Paso’s residents (U.S. Census Bureau, 2009). This is a likely explanation for the low White intimate partner SROK in El Paso, given the cultural tradition of male-dominated Latino households (Macionis, 2003, p. 347). This finding is also in keeping with previous studies of SROKs in the United States and Texas (Gauthier & Bankston, 2004; Paulsen & Brewer, 2000).

OPEN IN VIEWER

Table 2.

The Intimate Partner SROKs by Offender Race and Weapon for Texas’ Six Largest Cities, 1976-2007

Austin

Austin				Dallas
Offender				Offender
	Man	Woman	SROK		Man	Woman	SROK
Offender race				Offender race
White	88	32	36	White	257	135	53
Black	22	17	77	Black	286	297	104
Other	1	0	0	Other	7	1	14
Weapon				Weapon
Gun	72	31	43	Gun	345	288	83
Knife	14	12	86	Knife	82	117	143
Other	25	6	24	Other	123	28	23
El Paso				Fort Worth
Offender				Offender
	Man	Woman	SROK		Man	Woman	SROK
Offender race				Offender race
White	98	24	24	White	124	64	52
Black	11	3	27	Black	119	102	86
Other	2	0	0	Other	0	0	0
Weapon				Weapon
Gun	70	15	21	Gun	154	118	77
Knife	11	10	91	Knife	29	32	110
Other	30	2	6	Other	60	16	27
Houston				San Antonio
Offender				Offender
	Man	Woman	SROK		Man	Woman	SROK
Offender race				Offender race
White	382	214	56	White	248	121	49
Black	404	440	109	Black	72	59	82
Other	15	5	33	Other	1	0	0
Weapon				Weapon
Gun	485	444	92	Gun	196	99	51
Knife	148	187	126	Knife	42	59	140
Other	168	28	17	Other	83	22	27

Note: Intraracial incidents ranged from 86% of cases in Fort Worth to 97% in Austin. A range from 94% in Fort Worth and San Antonio to 98% in El Paso was heterosexual incidents. Available data in this SHR file did not include Ethnicity. Therefore Race, in this analysis, refers to White (both Hispanic and non-Hispanic), Black, and Other. p<.05

Table 2 provides a closer look at the distribution and characteristics of Intimate partner SROKs specifically for Texas’ six largest cities over this 31-year time period. As seen in Table 2, five of the six cities have significantly higher Intimate partner SROKS for Black offenders than Whites (including Hispanic Whites) or others. In Dallas and Houston, more Black women than Black men committed lethal violence against an intimate partner, with Black intimate partner SROKS of 104 and 109, respectively.

Table 2 also displays SROKs based on the weapon used to commit lethal violence against an intimate partner. In all six cities, SROKs were significantly (p < .01) higher in cases in which a knife was the cause of death. There was actually an excess of female- over male-perpetrated Intimate partner homicides committed with a knife in Dallas, Fort Worth, Houston, and San Antonio. In comparison to knife homicides, intimate partner SROKS were lower when a gun was the cause of death, though the six cities varied. Firearm SROKs were near parity in Dallas (83), Fort Worth (77), and Houston (92). Table 2 also reveals that incidents in which a weapon other than a firearm or knife was used accounted for the smallest number of incidents perpetrated by women or men, and the lowest intimate partner SROKs across all six cities. Though firearms and knives were the most likely weapons used by either sex in all six cities, this latter finding indicates that men were much more likely than women to use a variety of other methods of committing criminal homicide. One possible explanation is that, with the exception of killing with poison or other drugs, the other methods reported (e.g., drowning or strangulation) require relatively greater physical strength, typically the case for a male partner.

The Role of Age in Intimate Partner SROKs

One dimension of intimate partner SROKs as yet unexplored in previous research is their distribution based on the ages of victims and offenders. Specifically, we are interested in whether the likelihood of these relatively high levels of women intimate partner homicide offending continues into older age. As necessitated by the available data and earlier explained, we disaggregate the offender age into the categories of under age 50 and age 50 or older, to compute age-specific overall and intimate partner SROKs.

At least two notable findings are shown in Table 3. One is that women’s representation among homicide offenders is much lower for homicide overall than for intimate partner homicide, regardless of offenders’ age (except for older offenders in El Paso, where both overall and intimate partner SROK’s were very low). The other is that there is much more variability among cities in the sex ratios of intimate partner homicide than in overall criminal homicide offending, regardless of the offender’s age.

OPEN IN VIEWER

Table 3.

A Comparison of SROKs for Younger and Older (Age 50+) Offenders in Texas’ Six Largest Cities, 1976-2007

	Younger offenders (Age 15-49)		Older (Age 50+) offenders
City	Overall SROKs	Intimate partner SROKs	Overall SROKs	Intimate partner SROKs
Austin	13	46	9	28
Dallas	14	82	18	62
El Paso	10	26	20	17
Fort Worth	13	65	20	96
Houston	13	84	17	71
San Antonio	11	58	17	46

When looking at the intimate partner SROKs in homicides with younger offenders, we see that they range from 26 in El Paso to 84 in Houston. If we treat El Paso as an outlier, intimate partner SROKs of offenders under age 50 range from 46 in Austin to a high of 84 in Houston. In comparison, overall SROKs in homicides with younger offenders ranged from 10 in El Paso to 14 in Dallas. Furthermore, intimate partner SROKs among older (age 50 and above) offenders range from a low of 17 in El Paso and 28 in Austin to highs of 71 in Houston and 96 in Fort Worth. Among older offenders, the overall SROKs ranged from 9 in Austin to 20 in El Paso and Fort Worth.

SROKs Over Time

A reduction in national levels of intimate partner homicide since the 1990s has been well documented (Dugan, Nagin, & Rosenfeld, 1999; Wells & DeLeon-Granadas, 2004). As shown in Figure 2, this trend is borne out in the intimate partner SROKs for Texas and its six largest cities over the 31-year time frame of this investigation. The intimate partner SROK dropped by 48% for the U.S. overall and by more than 67% for the state of Texas. Across Texas’ six largest cities, the reduction in intimate partner SROKs was 34% to 82%, with the largest decreases in Fort Worth, Houston, and San Antonio. The reduction in SROKs reflects a decline in men’s versus women’s risk of intimate partner homicide victimization from the period 1976-1985, to the period 1986-1995, to the period 1996-2007, across all geographic units including El Paso (data not shown).

OPEN IN VIEWER

Figure 2.

Intimate partner SROKs over time for the United States, Texas, and Texas’ six largest cities, 1976-2007

Discussion and Conclusions

The primary objective of this study was to describe sex ratios of killing and intimate partner sex ratios of killing in a state representing 10% of all U.S. homicides over the 1976-2007 time period. Texas’ six largest cities are among the 20 most populous nationally, with 2010 populations ranging from 649,121 for El Paso to 2,099,451 in Houston.

This exploratory study yielded a number of important findings. We found, for example, that women’s versus men’s representation among intimate partner homicide offenders was significantly higher for Texas than the United States over the course of these 30 years. However, the range of Intimate partner SROKs varied considerably, from a low of 24 in El Paso to a high of 82 in Houston.

As related to weapon use in the commission of these approximately 60,000 homicides in Texas’ six largest cities, firearms and knives were the most likely weapons used by either sex in all six cities. We also found that the lowest sex ratios of killing for both intimate partner and nonintimate partner homicides were those involving weapons other than guns or knives, indicating that a relatively larger number of men than women used means such as fire or strangulation to commit their lethal acts.

We found that intimate partner SROKs among older (age 50 and above) offenders ranged from a low of 17 in El Paso to a high of 96 in Fort Worth. This relatively greater representation of older women as intimate partner homicide offenders calls for further examination in later studies, using city-specific data that would include information about possible histories of domestic violence that finally culminated in homicide.

On examining the basic race classification available in the SHR data, we learned that five of the six largest Texas cities had significantly higher intimate partner SROKs for Black offenders than Whites (including Hispanic Whites) or others. In two cities, Dallas and Houston, there were more Black women than Black men committing lethal intimate partner violence. In four of the six cities, we also found an excess of female- over male-perpetrated intimate partner SROKs when the weapon used was a knife, whereas the intimate partner SROKs were near parity in gun-related incidents for half of the cities we studied.

Another important finding was that there was a much wider range of intimate partner SROKs than overall SROKs, in both the younger (less than age 50) and older offender age categories. Though overall SROKs ranged from 10 to 14 for younger offenders and 9 to 20 for older offenders; intimate partner SROKs ranged from 26 to 84 for younger offenders and 17 to 96 for older offenders.

Finally, we found a significant (p <. 01) reduction in the intimate partner sex ratios of killing from the 1976-1985 to the 1996-2007 time period, statewide and for each of these cities. While the intimate partner SROK was reduced by 48% for the United States, it fell by more than 67% for the state of Texas and was reduced between 34% to 82% across Texas’ six largest cities.

Along with providing useful baseline data on overall and intimate partner sex ratios of killing over a 30-year time frame in urban areas of Texas, the present study also generates questions for future research. For example, while all six of these cities are governed by the same penal code related to homicide, there is likely to have been variation in city- or county-specific law enforcement responses to intimate partner nonlethal and lethal violence. Similarly, the present data offer no indication of the domestic violence resources available in each of these cities. In keeping with previous research showing that the availability of such resources may affect the saving of lives of women and of men in intimate relationships (Dugan et al., 1999), it would be interesting to track individual cities over time by their SROKs and resources available to abused women. Equally important in future research would be a temporal analysis of race/ethnicity-specific intimate partner sex ratios of killing, allowing for a comparison of how each of these cities tracks with national data indicating a long-term trend of greater reductions in female- than male-perpetrated Intimate partner homicide, particularly among Black couples.

In conclusion, Margo Wilson’s innate curiosity was borne out in part by the question of how homicide in intimate dyads is different from homicide generally. One answer to this question is found in her work with collaborator Martin Daly and others on the intimate partner sex ratios of killing. Study of this phenomenon has continued to the present day. The business of protecting women and men in intimate partner relationships, as well as the business of continued research in this critical aspect of human relationships, has been well served by Margo’s remarkable questions and insights. The present study has been particularly motivated by the extensive evidence that she and her collaborators provided to our understanding of differences between women and men in the volume, types, and motives for lethal violence.