Posted: November 28th, 2013
Term Project
Proposal
Variables Affecting Divorce Rates in 30 States
TABLE OF CONTENTS
PURPOSE STATEMENT AND MODEL.. …………………………………………………….3
EMPIRICAL ANALYSIS ………………………………………………………………………4
DATA DESCRIPTION………………………………………………………………………….7
DATA SOURCES……………………………………………………………………………….9
WORKS CITED …………………………………………………………………………….….10
Purpose Statement and Model
All around the United States and the whole world, the rate at which married couples have been divorcing has seen an alarming increase. This increase has been attributed to several factors. This creates the idea that there are those marriages that are more prone to end up in divorce than others are. Sociologists and marriage counselors use these variables to evaluate whether a given marriage will last for a long time or will end up in divorce. This paper will identify and analyze several variables in the effort of creating a parallel between the variables and the divorce rates. The study determines the effect of median income, median age, average household size and ethnicity on American divorce rates per 100,000 registrations. There exist other factors affecting divorce rates, for instance being endowed with efficient communication skills. These factors however, are hard to quantify and thus are outside the realm of this study (Blackburn, 2003).
Four independent variables have been specifically selected for this study depending on the availability of material, research and theoretical relevance of equation. The primary independent variable is the median income. This has been chosen as the main focus because of the results’ indication that it is the most important factor in the divorce rate. The paper will utilize a regression analysis in an attempt to explain the reason behind the rising rates of divorce. The four independent variables analyzed in this paper are the median income, median age, average household size and ethnicity.
The equation used in the attempt to analyze the different variables affecting divorce rates will be:
Divorce rate = F (median income_ median age_ average household size _ethnicity) + error term. The above equation is derived from the year 2000 cross sectional data set. The data set consisted of fifty observations from thirty states. The divorce rate is reached upon by taking the number of registered divorces per 100,000 registrations in a given state.
Empirical Analysis
The dependent variable, divorceper100k is the number of divorce cases filed per 100, 000 registered marriages within each state. This was arrived at by taking into account the total number of couples seeking formal divorce and correlating it with the total number of registered marriages from the given state. This data was obtained from the American Census Bureau. The bureau tracks marital events from different states in the US.
The primary independent variable is the median income of the married couple. Scientists from the University of Oklahoma undertook research on the causes of divorces and supported the idea that the median family income was a factor in the rates of divorce. Married couples who earned less than fifty thousand dollars per annum were more likely to have a divorce than those who earned more. Those who earn less have been shown to experience a considerably higher amount of stress than their counterparts who earned more (Johnson, & Jonathan, 1986). Their 1995 research indicated an affirmative relationship between the income level and the divorce rates. The reasoning behind this theory was that if the couple was well financed, the two do not have to face any of the issues associated with monetary shortages during and after divorce. This implies the couple concentrates on other issues, which is advantageous to the marriage. Greed in this case is taken to be the main issue for divorce (Lorenz, 2005). When the income sources are limited, a conflict is created between the limited funds and the limitless needs and requirements. In another instance, the establishment of a family unit creates two income earners whereby each of the partners becomes financially independent and thereby increasing the likelihood for divorce. Research shows that a good percentage of marital fights are brought about by money issues because of either having too much or lacking it altogether (Johnson, 2004).
The median age of the couple has been cited as one of the variables affecting divorce rates. Married couples between thirtythree to fiftytwo years of age depicted rates of divorce of 34%. Those in the age bracket of fiftythree to seventytwo years had the highest divorce rates at 37% while those in their senior ages of above seventytwo years had a divorce rate of 18%. The age of a couple is a major issue regarding whether the marriage will last or will end up in divorce. The main reason behind this is that if the median age is on the higher side, the individuals are bound to have attained financial independence and the capacity to handle financial stress more reasonably (United States Bureau of Census, 2005). Another reason is that with advancement in age, one is bound to have accumulated a considerable amount of experience in life’s issues and the ability to face such. It also endows one with the relevant knowledge on the best person one is suited to marry. Research indicates that the best median age offering the most advantages is the age of twentyfive years. Research indicates that there has been a steady rise in the median age at the first marriage in the last two decades. The median age at the first marriage has increased by more than three years for both men and women in the past two decades from twentytwo years for men to twentyfour years, and from twenty to twentytwo years for women.
Another of the variables affecting the divorce rates is the average household size. This is the number of persons in a given household. Research indicates that divorce rates were higher in homes where there was high average household size. The average household size in the United States of America is 2.5 persons per household. Families that depicted a number greater than this were more likely to end up divorced than those with fewer numbers (Trent, Scott, 1989). On the other hand, couples who had children together were likely to have a lasting relationship than those without children. The children were seen as a possible creation of a tighter bond between the two couples.
The divorce rate also varied considerably among the different ethnic communities in the United States of America. The United States Census Bureau’s 5Year American Community Survey indicated the percentage of each population divorcing over a fixed period of time (Somogyi, 1941). Their latest survey indicated 12.6% divorced couples among the American Indian and Alaskan natives, which was the highest number registered during the fiveyear period. The African American ethnic community was rated second with a percentage of 11.5%. The percentage of whites divorcing during the fiveyear period was 10.8%. The Native Hawaiian and Other Pacific Islanders who divorced during the time was 8% (Shelton, 1987).
Data Description
VARIABLE 
MAX 
MEAN 
MIN 
Divorce_Per_100k (Divorces per 1,000,000) 
7.0 Alabama 
4.19 
2.4 Arkansas 
Median age

50 years Alaska 
30 years 
18 years Chicago 
INC (Median family income) 
$83,000 Arizona 
$60,000 
$45,000 Indiana 
Household size (% of households with 2 persons or higher) 
33.2% California 
24% 
15% Idaho 
Ethnicity (Combined % of whites and blacks by state) 
75% Colorado 
50% 
29% Delaware 
Courtesy of: United States Bureau of the Census, 2009.
Name:
Course:
Lecturer:
Date:
Variables Affecting Divorce Rates in 30 states
Data Description
The data was mainly obtained from surveys conducted by the federal government. The data on the changes in divorce rates from state to state was obtained from the journal of marriage and the family written by Paul Nakonezny. The data on the different divorcing partners and their income rates was obtained from a report published by the United States department of commerce, bureau of the census titled State and Metropolitan Area Data Book. The information was collected between the years 1997 and 1998.
The data on the divorce rates from state top state was from the United States statistical abstract. This data was obtained from the section two of the vital statistics on page fiftyseven. This data was collected throughout a tenyear period from the year 1998 to 2008. Another source of data on personal income is one from the United States census bureau. This is from a report they published in the year 2007 titled Statistical Abstract of the United States: 2007 (U.S. Census Bureau, 2007). The data onthe per capita income is mainly the most recent information obtainable. This information is based on the census figures obtained in the year 2005.
The data on the divorce rates among different ethnicites was obtained from the united states statistical abstract. The data was collected fromt the year 1998 until the report was published in the year 2003. the data was obtained from section oneof the report titled population on page two. Qadditional information was obtained in the journal written by shelton Beth’s journal of marriage and the family titled, “Variations in Divorce Rates by Community Size: A Test of the Social Integration Explanation (Shelton, 1987)
Limitations
One of the underlying limitations is the lack of recent data. Most of the surveys were conducted from the year 2005 going backwards. This limitation fails us in having a conclusive and more recent report that clearly illustrates the situation as it is presently. Another limitation is the lack of complete ranges. Because of the conservative approach taken in mapping populations, the ranges of most of the tabulated data is highly likely to be maximum or mean estimates of the limits of population distribution. With this in mind, some regions are recorded as having lower rates than they may actually depict in real life situations.
Another limitation is that some parts of the United States are still highly remote and thus very little relevant data is collected. This also translated to minimal information is known about the communities living in such areas. The report also includes a sampling error where the estimates are based on a sample. Since the estimates are based on a sample, they will give a false impression if a definite census is conducted on the income tax returns on every divorcee. The sample used in coming up with this report was one out of a multitude of other samples that had the possibility of being selected under the given sample design. Estimates derived from other samples provided a possibility of differing from the chosen sample. Although the chosen sampling method has proven to be reliable, its truthvalue is still debatable in highlighting real life situations.
The nonsampling error is another possible error that could be evident in this report. These errors have been proven to affect the overall estimates. If the errors are random, then their effect may cancel out and thus have an insignificant effect on the estimates. On the other hand, if the errors are systematic, they may end up significant effect giving a biased report. The above errors arise from failing to obtain all the information in the given variables. Another reason is the difference in the interpretation of the different source documents containing the variables, errors committed when estimating the missing data and the failure to detail or represent all the aspects of a given population.
Regression Analysis
A regression analysis is conducted to model the relationship between the independent variable, which is the divorce rates, and the variables affecting the rates at which divorces are occurring. The model equation for conducting a regression analysis on multiple linear regression analysis takes the form of
Where y= independent variable
= dependent variables
= parameter vector
= error term
On applying the above equation to the results of the thirty states, the regressed model is now tabled as shown below.
DEPENDENT VARIABLE: DEATH ADJUSTED R^{2} = 0.9979 n = 30  
Independent Variables 
Coefficient 
T Statistic 
Significance of t 
median income 
0.8700 
7.4879 
1.1986 
median age

0.3499 
1.5102 
0.20003 
average household size 
0.7525 
0.5023 
0.6003 
ethnicity

1.2109 
0.9998 
0.9241 
A new equation for obtaining the relationship between the death rate and the various variables would now be written as
Divorce rate = F (0.8700median income +0.3499median age_0.7525average household size 1.2109ethnicity) + error term.
In order to perform a logged regression analysis, a slight a slight adjustment is performed by taking away the dummy variable ethnicity as it would include values of zero. After the slight modification, the new expression for deriving the estimates would therefore look like, Divorce rate = F (median income_ median age_ average household size) + error term. After performing the logged regression analysis, the results obtained are tabled in the figure below.
DEPENDENT VARIABLE: LDEATH ADJUSTED R^{2} = 0.9953 n = 33  
Independent Variables  Coefficient  T Statistic  Significance of t 
median income 
0.9278  13.4125  5.70004 
median age 
0.1244  1.4998  0.1568 
average household size 
0.0107  .0498  0.89995 
The new estimate equation would therefore take the form of Divorce rate = F (0.9278median income_0.1244 median age_0.0107 average household size) + error term.
Multicollinearity was tested to check whether two or more of the independent variables had a linear relationship with each other. This would end up causing a biased or false coefficient. The net effect would be having an unreliable result from the regression model. The test was performed by regressing al the independent variables and comparing the resulting R^{2} values to those of the entire model. After performing the above, the results were tabled below
Multicollinearity Test  
Independent Variable 1  Independent Variable 2  Simple Regression R^{2 }Result 
median income  median age  0.9201 
median income  average household size  0.1499 
median income  ethnicity

0.01002 
median age  median income  0.22001 
median age  average household size  0.01004 
average household size  ethnicity

0.17101 
From the above results, it is evident that multicollinearity does not occur in this case. From the calculations, the R2 for the entire was model was. After performing a regression on all the pairs of the variables, none was equal to the value for the entire model.
The results of the regression analysis were in line with the predictions made earlier on. The positive coefficient signs indicated that the variable would have a positive effect on the divorce rate whereas a negative coefficient means that the variable would have a negative effect on the dependent variable. Studies indicate that a true adjusted R^{2} on the regression analysis should be less than one but greater than or equal to zero. This coincides with the above results as the adjusted R^{2} from both regressions are very close to one and less than one. This means that the regression model accounts for over 99% of the variations in the dependent variable. This also goes ahead to explain that the above chosen independent variables indicate a correct likelihood of a given couple filing for divorce.
Works Cited
Blackburn, McKinley. “The Effects of the Welfare system on Marital Dissolution.” Journal of Population Economics 16 (2003): 477500.
Johnson, John. “Do Long Work Hours Contribute to Divorce?” Topics in Economic Analysis & Policy 4.1 (2004): 123.
Johnson, William, and Jonathan Skinner. “Labor Supply and Marital Separation.” The American Economic Review 79.3 (1986): 455469.
Lorenz, Scott. “Sex and Money Top Two Reasons for Marriage Problems, says author of ‘The Marriage Medics’.” Press Release Newswire. 21 February 2005. Retrieved from <http://www.prweb.com/releases/2005/2/prweb210502.htm>
Shelton, Beth Anne. “Variations in Divorce Rates by Community Size: A Test of the Social Integration Explanation.” Journal of Marriage and the Family 49.4 (1987): 827832.
Somogyi, Stephen. “Differential Divorce Rates by Religious Groups.” The American Journal of Sociology 46.5 (1941): 665685.
Trent, Katherine, and Scott South. “Structural Detriments of the Divorce Rate: A CrossSectional Analysis.” Journal of Marriage and the Family 51.2 (1989): 391404.
United States Bureau of the Census. State and County Quick Facts. Feb. 2005.
United States Bureau of the Census. Statistical Abstract of the United States. 2002.
Data Sources
The data for percentage divorce rate per state was found using the Journal of Marriage and the Family.
Source: Nakonezny
The data on state to state family average income levels was found in the State and Metropolitan Area Data Book.
http://www.census.gov/statab/www/smadb.html
The data on the divorce rates from state to state was found in the United States Statistical Abstract.
http://www.census.gov/prod/2004pubs/04statab/vitstat.pdf
The data on divorce rates among different ethnicities was found in the U.S. Statistical Abstract.
http://www.census.gov/prod/2003pubs/02statab/pop.pdf
Place an order in 3 easy steps. Takes less than 5 mins.