December 13, 2009 -- Vol.13, No.1
Do Fundamentalism and Other Religious Variables Predict Domestic Violence?
One of many serious problems we face in the United States is a problem of the most intimate nature. In our states and cities and towns and counties—on our streets, maybe in our own families—this problem is played out in pain, and shame, and sometimes, with deadly consequences. It is a problem between lovers, husbands and wives, and intimate partners. It often involves children. They are likely victims, even if no one “lays a hand on them.” Our problem is domestic violence. How extensive is the problem? In their JAMA article, Grumbach, Rodriguez, McLoughlin, and Bauer (1999), state, “Physical violence is estimated to occur in 4 to 6 million intimate relationships each year in the United States.” Kemp (1998) reports an incidence rate of at least 120 per 1000 couples each year.
This study is mainly concerned with religious variables as predictors/correlates of domestic violence (DV). There is a “common sense” rationale for this connection, especially when one considers fundamentalism. The so called “patriarch thesis” or theory appears logical. When fundamentalist denominations teach that the husband is the proverbial “king of the castle” and his wife should obey him; it seems like a powder keg ready to explode. Added to what we already know concerning the many “extremist” religious groups with male dominance theories and their occasional appearance in the news media, we see how patriarchy can result in violence toward women. But, how does the notion of mainstream evangelicalism/fundamentalism as a correlate—either positive or inverse of DV, hold up under statistical scrutiny? And what about other measures of religious adherence or non-adherence?
Cunradi, Caetano, and Schafer (2002), in a study of religious affiliation and domestic violence found higher rates of domestic violence among liberal religionists than among fundamentalists. They conclude that a patriarchical family structure among many fundamentalist groups might actually contribute to less incidence of DV. Cunradi et al interpreted the results of the study to indicate that religious factors help moderate alcohol problems, which may be one of the central features of religion/domestic violence interactions.
In a partially contrasting study, Brinkerhoff, Grandin, and Lupri (1992) examined the impact of church attendance and denominational membership on DV. They conclude that these variables are not related to spousal violence. Again, no evidence for the “patriarch thesis” among conservatives was noted.
It may be important to note that the Brinkerhoff et al (1992) study involved Canadian subjects. This might suggest that there are regional differences in the way religion relates to domestic violence. Since religion includes many factors—social, cultural, ethnic, economic, educational, emotional, psychological, familial, and many more—there might be many regional variables that impact the issue of the relationship of religion and DV.
Considering fundamentalist populations in particular, what does the literature have to say? Lane (2000), in a study of Southern Baptist couples found a physical assault rate roughly equivalent to that of the population as a whole. Neither religious commitment nor religiousity was significantly related to domestic violence. Lane concludes that the “patriarchy thesis” is not supported, although statistical significance was found for age, with younger couples more likely to report incidents of violence.
In Utah, more than 95 percent of the Protestant fundamentalists are Mormons. Rollins and Oheneba-Sakyi (1990) found spousal abuse rates somewhat higher than the national average. However, employment status, family size, income, marital power differential, gender role orientation, and religiousity did not differentiate the rate of spousal violence.
Using a different paradigm, Laner (1985) points out that patterns involving Mormons and non-Mormons and violence during courtship are confusing. More non-Mormon women than Mormon women reported unpleasant and abusive treatment during courtship in both past and present relationships. Yet, more Mormon men admitted to being abusive.
As the foregoing indicates, although the picture is somewhat mixed, most investigators have found little relationship between religion and DV. Moreover, what is deemed significant is, in many cases, counterintuitive in terms of the common sense relationship suggested above.
The current investigation makes use of the 2000 Religious Congregations and Membership Study (RCMS). This raises a point concerning potential problems inherent in this study. Finke and Scheitle (2005) comment on the difficulty of making an accurate count of religious adherents in the US. Commenting on the data provided in the RCMS, they characterize it as the most complete enumeration available by counties, states, and nation. Yet, they find fault with the census, in which adherents accounted for about 50 percent of the total US population. They offer some corrected figures, which place the actual national level of adherents at about 63 percent. While they may offer some helpful data for regional and state comparisons, as the nature of the current study makes apparent, it is of limited value in the current investigation. Additionally, around 80 percent of Americans claim affiliation with a church or religious group. Obviously, there is still much to account for. Lastly, the RCMS corrections offered by Finke and Scheitle (2005) are partially based on assumptions, which may lead to over or undercounting, depending on the accuracy of the assumptions.
A second problem deals with domestic violence data. It is notoriously difficult to obtain in any uniform, consistent manner that allows “across state” comparisons. The current researcher believes that the approach taken here is one of many valid courses of investigation possible.
STUDY PART 1
The data displayed in the Table 1 (excluding the last column) is dependent on data from the text, Religious Congregations and Membership in the United States 2000 (Jones, Doty, Grammich, Horsch, Houseal, Lynn, Marcum, Sanchagrin, and Taylor 2002). The book represents the results of a religious census conducted in 2000 and contains data from 149 religious groups. The census included data from 139 Christian denominations (including Unitarian Universalists and Later-day Saints). It also includes data on independent Christian groups and numerical data on Baha’i, Islamic, and Jewish adherents.
As Religious Congregations and Membership points out, the total adherents represented in the study was 141,371,693 or 50.2 percent of the total US population. Yet about 80 percent of Americans claim membership in a religious group. The text offers two possible explanations for the discrepancy.
First, many folks claim membership, but that is not borne out by denominational data. The census reflects actual participation in religious congregations as reported by a given religious body. Second, it is plainly admitted that the listing is incomplete. There are 14 religious bodies each with a membership in excess of 100,000 not reporting data to the census. These account for a shortage of 31,040,360. There are also many independent churches and religious movements that report to no external data collection source. Attempts were made by the authors to be complete and accurate. Still, the bottom line is simply, “It is not known what percent of total religious adherents in America this represents,” (p. xiii).
A final issue deals with what data to utilize. The data provided by Religious Congregations and Membership provides the number of full members or communicants, number of attendees, and the total adherents. Denominations have a great deal of freedom in deciding what constitutes adherents, though the basic notion seems clear enough. An adherent is an individual over whom a denomination has influence—usually measured by attendance, even if somewhat sporadic, or affiliation. The current study utilizes the number or percent of adherents in computation. For ease of interpretation, all data (except domestic violence data) is presented as non-decimal percents (they have been rounded) or given as whole numbers. The following table displays data used in this study. It is accompanied by an explanation of each variable
Explanation of variables
Total Adherents (TA): TA reports the percent of the state population who are adherents as counted in the census.
Total Protestant Adherents (TPA): This is reported as a percent. The percent of total adherents who are Protestant is multiplied by the TA to arrive at the TPA per state population.
% Protestant Fundamentalist (%PF): Determining whether a denomination is mainline/liberal or fundamentalist/evangelical is at the heart of determining the percent of Protestants in a given state are fundamentalist/evangelical. To make this determination, five sources were used: first the membership list of the National Council of Churches (National Council of Churches); second, the membership list of the National Association of Evangelicals (NAE); third, the membership list of the Pentecostal/Charismatic Churches of North America (PCCNA); fourth, the charts, tables, and commentary from Smith’s article (1990). The NCC membership is viewed as mainline Protestant. The NAE and the PCCNA membership tend towards evangelicalism/fundamentalism. Smith’s article contains classification information for both mainline/liberal and evangelical/fundamentalist. Fifth, if a denomination appeared on none of the three organizational membership lists, and Smith’s article was not definitively helpful, Mead’s Handbook of Denominations in the United States (2005) became a source of classification. When it was impossible to determine denomination position with sufficient clarity, it became necessary to remove the denomination in question from the data used in the study. Usually these were very small denominations.
The question of how to classify LDS adherents must be addressed. Smith (1990) discusses the problem posed by classification of Mormons at some length. Certainly, most fundamentalist Christians hardly see themselves as of “one mind” with Mormons (Kennedy 1999). Still, when it comes to basic literalism, the literal approach to the Bible, extra-biblical documents, and authority remains. Smith classes Mormons with the fundamentalists. That practice is followed here.
To arrive at a state’s %PF then, the percent of TPA counted as fundamentalist/evangelical is multiplied my TPA. One may ask why evangelical groups are numbered with the fundamentalists. Alexander (2008) has shown, the major criterion for classification as fundamentalist should be biblical literalism. The Statement of Faith of the National Association of Evangelicals is, in essence, very similar to the position staked-out in The Fundamentals, the 1909 document that forms the theological justification of fundamentalism. The NAE statement represents an appeal to literalism compatible with the fundamentalist ideal (National Association of Evangelicals).
Domestic Violence Rate (DVR): This category utilizes data provided by two different reports published by the National Network to End Domestic Violence. The first, Domestic Violence Counts 2007: A 24-Hour Census of Domestic Violence Shelters and Services Across the United States reports concerning services offered in a non-duplicated count of individuals seeking help from identified domestic violence programs on September 25, 2007 (National Network to End Domestic Violence 2007). A second report, Domestic Violence Counts 2008: A 24 Hour Census of Domestic Violence and Services, is also used (NNEDV 2008). This census reports non-duplicated data collected on September 17, 2008. Services provided to victims in the unduplicated count typically included individual support or advocacy, children’s support or advocacy, legal-related services, transportation, group support or advocacy, housing, and childcare/daycare (see 2008 report).
The average number of programs reporting to the census in 2007 was 69 percent. By 2008, the number rose to 78 percent. The combined average for the two years is 73 percent. Any state with an agency reporting rate of less than 58 percent is considered low response (less that 80 percent of the average). In compiling the DVR the number of domestic violence victims receiving services as well as those seeking services but who could not be accommodated were averaged for the two years unless one of the years fell below a 58 percent response rate for that state. In that case, the year with a 58 percent average or above provided the only data utilized for that state. Finally, it is noted that five states failed to receive the 58 percent minimum either year. For these states, the year with the highest response rate is utilized.
Dividing the number of victims of DV served by reporting agencies by the 2007 population estimate for the given state (United States Census) and multiplying by 100,000 provides the average reported DVR per 100,000 population. This final number results in the DVR for the purpose of this study.
One might ask why the Domestic Violence Counts Data was used for this project. The reasoning behind its use is best summed up by the Executive Director of AARDVARC.org— An Abuse, Rape and Domestic Violence Aid and Resource Collection— in response to a request for assistance in locating resources for DV statistics by state. She writes, the main problem is
…[E]very state DEFINES domestic violence differently- vastly differently- so there is no uniform methodology…. The bottom line is you really have to go state by state, and know from the start that comparisons between any two states really require that you compare states with VERY similar statute definitions- comparing otherwise becomes an apples vs. oranges exercise…. [N]o one to date has developed a methodology that reasonably… gets close to being able to compile anything useful at a national level. (Catherine NeSmith, personal communication)
The Domestic Violence Counts data was collected with an identical methodology, involved all states, had a good overall response rate, and worked from the same definitional assumptions.
Describing the sample
Osborne and Overbay (2004) discuss the problem of outliers at length and recommend removing them from the data set. They recommend an acceptable cutoff point of 3 standard deviations. However, due to the erratic nature of the data (using Anderson-Darling Normality statistics), this investigation employed a “cut point” of 2.75 standard deviations. Looking at all variables, it was decided to remove Utah and Alaska from further data analysis. Removing Utah and Alaska from the data set did, in general, improve the normality situation to varying degrees. Table 2 shows the descriptive statistics from the new data set (N = 48)
It was hypothesized that religious variables, particularly fundamentalism, might be good predictors of domestic violence. There is much anecdotal evidence for this assumption. As recently detailed in the press, the Fundamentalist Church of Jesus Christ of Latter Day Saints certainly traumatized both women and children. This took place on both a psychological, “brainwashing” level, and an overt sexual/physical level. One might recall that the Branch Davidians, also a fundamentalist cult, was investigated for child abuse. The Children of God, a fundamentalist group originating in the days of the Jesus Movement (late 1960’s-early 1970’s), has a long history of child abuse and person abuse in general. Accounts such as related in Under the Banner of Heaven describe abuse of the deadliest nature in all too clear detail (Krakauker 2004).
This might lead to the suggestion of certain propositions:
1. Fundamentalism, especially any variety that discourages questioning and holds to an ultimate unquestionable authority (in this case the Bible or another religious book or figure) might cause adherents to display increased rates of domestic violence. This is especially true if a pattern of male authoritarianism is taught based on said unassailable authority. Situations such as this occurred in extreme cases such as those described above. This might be true on a lesser scale for fundamentalist religious denominations in the mainstream of American culture, which often promote male authoritarianism.
2. Some religious factors might be protective factors. In other words, while popular American Christian fundamentalism might lead to an increase in domestic violence, perhaps, American style Catholicism, noted by Newport (2009) for its independence of thought, might be a protective factor when it comes to domestic violence.
3. Finally, although rejected by the vast majority of researchers, perhaps religion in general contributes to the problem.
To explore these notions, all four religious factors measured in this study: total adherents for all religions by state (TA); Catholic total adherents by state (CTA); Protestant total adherents by state (PTA); and percent of Protestant fundamentalists (%PF) were used as independent variables in a multiple regression study. The dependent variable for the study was the domestic violence rate (DVR).
Taken as a whole, there is no significant effect for the religious factors measured in this study and DVR. Some of this may be due to the odd distribution of the data set. However, a post hoc question is raised by this analysis. By examining Table 1, it is apparent that the distributions are far from normal. Therefore, the question is raised concerning a different type of analysis- an analysis of the contiguous states (leaving out Utah) by region. This appears to possibly be a better paradigm, since regions often share many characteristics, and, therefore, might control for unknown confounding variables affecting the analysis. With that in view, a bivariate correlation for each of the five traditionally defined regions of the US using %PF and DVR was implemented. The regions are shown in Table 3.
On a region-by-region parametric bivariate regression analysis using %PF as an independent variable, no significance was found, except for the West. In view of the small size of the individual data sets, it seemed appropriate to analyze the data using the Spearman Rank Order Correlation Test, since nonparametric methods are less susceptible to error due to sample size or relative shape of the distribution. The results of Spearman’s Rank Order Test reveals a high level of correlation and significance for %PF and DVR for the West region (rho = 0.9524, two-sided p-value = -.0011).
Reflection on regional analyses
The effect for fundamentalism as a correlate or predictor of domestic violence on the West appears both strong and definite. However, this raises an issue addressed earlier. The West has a relatively large Mormon population. Could that be the cause of the observed relationship? Mormons make up about 31 percent of fundamentalist adherents in the West.
To repeat the question: Is it possible that the effect is mainly due to the presence of LDS adherents as opposed to fundamentalism in general, of which LDS adherents make up only a part? To test this notion, a second Spearman Rank Order Test was conducted using the percent of adherents of the LDS faith in each state. Is the “Mormon Factor” sufficient in and of itself to explain the high correlation and positive results on the bivariate regression and the Rank Order Test for %PF and DVR in the West? Returning again to the Rank Order Test, it seems it is not (rho = 0.36, two-sided p-value = 0.39). It appears to be total fundamentalism in the West, as opposed Mormonism alone that proved correlative/predictive.
A return to the data
The main concern of this study deals with fundamentalism as a correlate or predictor of DVR. Still, a question of interest might be: Do any of the other variables serve as predictors of DVR on a regional basis? To answer this question, each variable, excluding %PF, were utilized in a bivariate regression analysis as an initial screening. If significance was found, in view of considerations previously mentioned, the Spearman Rank Order Test was conducted using that variable. The Table 4 shows significant effects found on subsequent Rank Order Tests.
What about non-adherents? Do they have a greater propensity for DV than adherents? To examine that notion, the Spearman Rank Order Test was applied to the percent of non-adherents per state for both the West and the Midwest. In the West, a correlation was found, but it was hardly of a magnitude or significance (rho = 0.29, two-sided p-value = 0.055) to merit consideration compared to the magnitude or significance of positive correlation of TA and DVR . In the Midwest, there was a significant correlation between non-adherence and domestic violence, but it was an inverse correlation. Non-adherents were less likely to be associated with DV (rho = -0.64, two-sided p-value = 0.050). To make certain the results from the two tests were finding a consistent pattern, tests were also conduced for non- adherence in the other regions (Northeast, Southeast, and Southwest). In no case did the correlation prove significant.
It might be pointed out that we are only dealing with two regions of the United States. The states used in calculations for the West represent about 19 percent of the US population—nearly one-fifth of the US. The states included in the Midwest represent slightly over 20 percent of the total population—fully one-fifth (based on estimates for 2007 from the US Census). The combined population represents 40 percent of the US. Some might attribute the effects observed in this study to the presence of significantly more Protestants in each state. However, in both regions, there are only about 6 percent more Protestants on the average than Catholics. It is quite possible for a relative increase in Catholics mirroring an increase in DVR in a given region to show a positive relationship to DVR at the same time as TPA or %PF.
Others might argue that it is actual religious attendance and not adherence that makes a difference. Sometimes, writers will even list religious involvement as a safeguard against domestic violence (Ellison and Anderson 2001, Ellison Trinitapoli, Anderson, and Johnson 2007). Yet, using the regions set forth in Religious Congregations and Membership in the United States, it is possible to establish the percent of adherents who are attendees for each region (Jones et al 2002). When the average percent of victims of domestic violence per region is compared, the correlation between religious service attendance and domestic violence is not significant—neither positively significant nor inversely significant as many researchers have hypothesized (two-sided p-value = 0.193).
STUDY PART 2
METHODS AND RESULTS
Part two of this study examines six factors that have been linked to male perpetrators of domestic violence. The factors are summarized in a list compiled from U.S. government resources (MedicineNet). Identical factors are collaborated by Kemp (1998). The six factors are as follows:
What follows is a description of each variable in terms of how it was measured. An attempt was made to obtain national correlation statistics using the six variables and domestic violence. However, distributions were far from normal, and such comparisons became difficult. It was decided to simply compare data from the variables from the two regions (West and Midwest) where significance was found for religious variables. The research question in this case was simply: How did the religious variables in each region compare with the six “non-religious” variables? This section concludes with a table (Table 5) showing the six variable’s coefficients where variables were significant.
Explanation of the variables
Low Income (LI): Data supplied by the U.S. Census provides the percent of families living below the federal poverty level (U.S. Census 2007 data).
Low Educational Attainment (LEA): Using data from the National Kids Count Program of the Annie E. Casey Foundation, the high school dropout rate from 2003 was averaged with the dropout rate from 2007 (Kids Count). The rationale for this procedure is that an individual dropping out at age 17 in 2003 would be 21 years of age in 2007 when most of the data used in this study was collected. Theoretical models seem to indicate that young age is often a factor in domestic violence. By averaging the two rates together, the data represents a wide sweep of young adults.
Unemployment Rate (UER): This statistic is a measure of unemployment rate by state for November 2008 released on December 19, 2008 (CNN Money.com).
Young Age (YA): Using data from the three-year estimates from the 2005-2007 American Community Survey, the percent of adults aged 18-24 in each state in the region is estimated to arrive at YA (U.S. Census).
School Violence Rate (SVR): While VCR provides one measure of the violence that might have surrounded a child during his/her formative years; it might be useful to attempt to measure a source closer to a child during his/her “growing-up” years. A statistic that might prove useful for that purpose is a measure of students who were in schools with relatively more or less students bringing a weapon (a firearm) to school. The Report on the Implementation of the Gun-Free Schools Act in the States and Outlying Areas School Year 2003–04 provides the number of violations per 1,000 enrollment by state (U.S. Department of Education).
The religious variables appear to fare well compared to the non-religious variables usually considered as correlates of domestic violence. In the West, both %PF and TPA fare better than every measure considered. VCR was inversely related, but that appears to be due to relatively high crime rates in urban areas in Nevada (which has a rather low DVR) and the West Coast. Regarding the Midwest, LI, LEA, VCR, were not significant. The three significant religious variables, TA, TPA, and Non-adherence, were slightly better than ACR. SVR proved to be roughly equivalent to the religious correlates, though it was better than TPA. Among the non-religious variables, YA and UER proved to be the stronger correlates, though not appreciably more so than TA or Non- adherence (rho = 0.70, 0.67, and 0.64, respectively).
It is important to begin this discussion with a disclaimer, although it is rather obvious. This study deals with variables at the macro-level, using trends of tens of thousands, hundreds of thousands and millions of people. Therefore, it would be a misnomer to attempt to extend the results and conclusions of this study to the micro, or individual level. The further applications of the conclusions of this study “move away from” this macro-level, the more likely they are to lead those applying the conclusions in ways leading to mistaken assumptions. For example, to say that fundamentalism is not a factor in domestic violence in individual cases in the Southeast goes far beyond implications derived from the data set utilized in this study. Likewise, for example, to conclude that alcohol abuse does not play a role in domestic violence in individual cases in the West is taking the data out of context.
It is important for the reader to bear in mind that while this study reflects data collected and analyzed at the macro-level, by far, the studies reviewed in the introduction to this paper are at or much closer to the micro-level. Those studies deal, for the most part, with identified populations representing certain religious outlooks identified before data collection. Such careful identifications are absent from the current investigation.
Utilization of such data, however, is not a necessary component of the current investigation. The goal is to examine data reflecting much larger numbers of people (in the millions) and to look at large (macro) patterns of domestic violence and religious orientation. There are both benefits and limitations to this approach.
On the positive side, a major benefit is the sheer number of cases, or cases represented, taken into account in the current study. Further, the larger scope of data allows the creation of a method for determining the fundamentalist character of a given state. A method for comparing states on relative levels of domestic violence also proved useful. Although the data is far too irregular for analysis when applied to the nation as a whole, the study reveals rather conclusive results on a regional basis, and it is possible to draw some conclusions regarding large numbers of the population. These conclusions have solid statistical support.
Of major concern is the inability to generalize from the larger macro-levels to micro-levels. The farther one moves from the particular to the more general, the less confounding variables are controllable. Therefore, it would be straying too far afield to attempt to apply statements from such a general study to small groups or particular cases. In addition, although regionalizing did greatly normalize distributions, the results beg the question: Why are similar patterns found in the West and the Midwest not discovered elsewhere?
Research at both the macro-level and micro-level have a place in answering research questions—especially in a study such as this. One distinct benefit of this study is that it virtually bypasses the possibility of intentional, personal false disclosure, or at the very least provides greater immunity to that effect.
As Ring, Nash, MacDonald, Glennon, and Glancy (1998) point out, religion is a culture-related phenomenon. As they demonstrate, religion involves the economic, sociological, and ideological components of culture and all they entail. In this case, there are two immediate implications. First, it may help account for regional differences in the current investigation. Second, it deals directly with the issue of exactly “what” the study measures. It may well be argued that what is measured is religion, but recall that religion includes many components.
Several variables were examined in this study. They include:
Post hoc analysis brought more variables into play:
Correlations of these variables and DV noted as significant in this study were moderate to highly significant. Correlations with domestic violence are worthy of further study, perhaps using a regional paradigm first and moving toward a case study paradigm. Other important areas of research are the accuracy of reporting and the peculiarity of distribution(s). As far as the conclusions drawn from this investigation, additional areas of research become apparent. Why, for example, is a significant correlation noted involving DV and Protestantism but not Catholicism—a condition true in at least 40 percent of the US population? Why is the relationship of fundamentalism to DV so strong in the West? Mormonism is not a sufficient explanation. Why would non-adherents, at least in one region be more likely to abuse than religious adherents? Can some of the components of religion, such as those mentioned above, be isolated for further investigation?
I leave the reader with these questions, because they serve as important reminders of what we can (and cannot) conclude from this study. We may safely conclude that a relationship between fundamentalism, Protestant adherence, religious adherence in general, and religious non-adherence are significantly related to domestic violence (the last variable inversely) in 20-40% of the population. We may safely conclude that the traditional non-religious risk factors for domestic violence do not, for the most part, correlate with any greater significance than the religious factors, and not as well overall in the regions examined (when assessed using data at a macro-level). This is significant and may provide direction for further study. Finally, it is useful to bear in mind this caveat: The conclusions drawn from a macro-level study carry certain limitations (as do micro-level investigations), and the data utilized in drawing conclusions are only as reliable as the reporting methods and sources.
Alexander, James. 2008. Stories of a Recovering Fundamentalist: Understanding and Responding to Christian Absolutism. Authorhouse: Bloomington, IN.
Binkerhoff, Merlin, Elaine Grandin, and Eugen Lupri. 1992. Religious Involvement and Spousal Violence: The Canadian Case. Journal for the Scientific Study of Religion 31:15-31.
CNN Money.com. Unemployment State by State. Available at http://money.cnn.com/pf/features/lists/state_unemployment <Viewed on June 27, 2009>.
Cunradi, Carol, Raul Caetano, and John Schafer. 2002. Religious Affiliation, Denominational Homogamy, and Intimate Partner Violence Amoung U.S. Couples. Journal for the Scientific Study of Religion 41:139-151.
Ellison, Christopher and Kristin Anderson. 2001. Religious Involvement and Domestic Violence Among U.S. Couples. Journal for the Scientific Study of Religion 40:269-286.
Ellison, Christopher, Jenny Trinitapoli, Kristen Anderson, and Byron Johnson. 2007. Race/Ethnicity, Religious Involvement, and Domestic Violence. Violence Against Women 13:1094-1112.
Finke, Roger and Christopher Scheitle. 2005. Accounting for the Unchurched: Computing Correctives for the 2000 RCMS Data. Review of Religious Research 47:5-22.
Grumbach, Kevin, Michael Rodriguez, Elizabeth McLoughlin, and Heidi Bauer. 1999. Screening and Intervention for Intimate Partner Abuse: Practices and Attitudes of Primary Care Physicians. JAMA, The Journal of the American Medical Association 282:468-474
Huck, Schuyler. 2000. Reading Research and Statistics. Longman: New York.
Kemp, Alan. 1998. Violence in the Family: An Introduction. Brooks/Cole: Pacific Grove CA:
Kennedy, John. 1999. Are Mormons Christians? Today’s Christian available at http://www.christianitytoday.com/tc/1999/marapr/9r2068.html <Viewed on June 4, 2009>.
Kids Count. Teens Who Are High School Drop Outs. Available at http://datacenter.kidscount.org/data/acrossstates/Rankings.aspx?ind=73 <Viewed on June 27, 2009>.
Jones, Dale E., Sherri Doty, Clifford Grammich, James E. Horsch, Richard Houseal, Mac Lynn, John P. Marcum, Kenneth M. Sanchagrin. and Richard H. Taylor. 2002. Religious Congregations and Membership in the United States 2000: An Enumeration by Region, State and County Based on Data Reported for 149 Religious Bodies. Glenmary Research Center: Nashville.
Krakauer, Jon. 2004. Under the Banner of Heaven: A Story of a Violent Faith. Anchor: New York.
Lane, Donald. 2001. Family Relationship Dimensions and Violence Among Married Couples In Southern Baptist Churches. Dissertation Abstracts International Section A: Humanities and Social Sciences 61:4566.
Laner, Mary. 1985. Unpleasant, Aggressive, and Abusive Activities in Courtship: A Comparison of Mormon and non-Mormon College Students. Deviant Behavior 6:145-148.
Mead, Frank, Samuel Hill, and Craig Atwood. 2005. Handbook of the Denominations in the United States (12th Edition). 2005. Abington: Nashville.
MedicineNet. Intimate Partner Violence: Fact Sheet. Available at http://www.medicinenet.com/script/main/art.asp?articlekey=41728 <Viewed on June 4, 2009>.
National Association of Evangelicals. Statement of Faith. Available at http://www.nae.net/index.cfm?FUSEACTION=nae.statement_of_faith <Viewed on June 4, 2009>.
National Association of Evangelicals. Current NAE Members. Available at http://www.nae.net/index.cfm?FUSEACTION=nae.members <Viewed on June 4, 2009>.
National Council of Churches. Member Communions and Denominations. Available at http://www.ncccusa.org/members/index.html <Viewed on June 4, 2009>.
National Institute an Alcohol Abuse and Alcoholism. Volume and Ethanol Consumption for States, Census Regions, and the United States, 1970-2006. Available at http://www.niaaa.nih.gov/Resources/DatabaseResources/QuickFacts/AlcoholSales/default.htm <Viewed on June 4, 2009>.
National Network to End Domestic Violence. 2008. Domestic Violence Counts: 2008 A 24-Hour Census of Domestic Violence Shelters and Services. Available at www.nnedv.org <Viewed on June 4, 2009>.
National Network to End Domestic Violence. 2007. Domestic Violence Counts 07: A 24-Hour Census of Domestic Violence Shelters and Services Across the United States. Available at www.nnedv.org <Viewed on June 4, 2009>.
Ne Smith, Catherine. 2009. Personal communication.
Newport, Frank. 2009. Catholics Similar to Mainstream on Abortion, Stem Cells: Catholics Actually More Liberal on Some Issues. Available at http://www.gallup.com/poll/117154/catholics-similar-mainstream-abortion-stem-cells.aspx <Viewed on June 4, 2009>.
Osborne, Jason and Amy Overbay. 2004. The Power of Outliers (and why researchers should ALWAYS check for them). Practical Assessment and Evaluation 9 available at http://pareonline.net/getvn.asp?v=9&n=6 <Viewed on June 4, 2009>.
Pentecostal/Charismatic Churches of North America. PCCNA Member Churches. Available at http://pccna.org/Member_Churches.php <Viewed on June 4, 2009>.
Ring, Nancy, Kathleen Nash, Mary MacDonald, Fred Glennon, and Jennifer Glancy. 1998. Introduction to the Study of Religion. Orbis: MaryKnoll, New York.
Rollins, Boyd and Yaw Oheneba-Sakyi. 1990. Physical Violence in Utah. Journal of Family Violence 5:301-309.
Smith, Tom. 1990. Classifying Protestant Denominations. Review of Religious Research 31:225-246.
U.S. Census Bureau. 2005-2007 American Community Survey 3-Year Estimates. Available at http://factfinder.census.gov/servlet/CTGeoSearchByListServlet?ds_name=ACS_2007_3YR_G00_&_lang=en&_ts=264128411329 <Viewed on June 27, 2009>.
U.S. Census Bureau. Population Estimates. Available at http://www.census.gov/popest/states/NST-ann-est.html <Viewed on June 4, 2009>.
U.S. Census Bureau. R1707. Percent of People Below Poverty Level. Available at http://factfinder.census.gov/servlet/GRTTable?_bm=y&-geo_id=01000US&-_box_head_nbr=R1701&-ds_name=ACS_2007_1YR_G00_&-redoLog=false&-format=US-30&-mt_name=ACS_2007_1YR_G00_R1701_US30&-CONTEXT=grt <Viewed on June 26, 2009>.
U.S. Department of Education. Report on the Implementation of the Gun-Free Schools Act In the States and Outlying Areas School Year 2003–04. Available at http://www.ed.gov/about/reports/annual/gfsa/gfsa03-04rpt.pdf <Viewed on June 29, 2009>.