Statistical Analysis of Divorce and Separation Cases

Brat Swam
6 min readApr 26, 2021

Statistics analysis of attorneys, minors, antagonism, duration. etc of divorce and separation cases in a populous county in USA.

The analysis is based on data that was publicly available. The data consists of 338510 distinct cases, 7820 attorneys, 4.65 million documents, 1.08 million calendar events, 289664 minors. The article tries to shed quantitative light on characteristics of divorce and separation cases.

In all the plots below, the horizontal axis is the number of cases. The vertical axis varies by the topic. This may appear visually a little unusual, but mathematically does not make any difference.

This work is not reviewed by anyone (and a work in progress). If any problem, mistake or any other issue is evident, please do not hesitate to communicate.

The text of the publication is organized into multiple sections, focusing on topics such as attorneys, minors, antagonism, etc.

Attorneys involved in cases:

The following chart shows the distribution of number of attorneys involved in cases. 0 attorney implies both parties self representing themselves. The highest number of attorneys involved in a case appears to be 31.

Number of attorneys per case vs number of cases

The following 2 charts show log/linear-log/linear plots of number of attorneys and the number of cases they are representing. The highest case load, 276284, belongs to both parties not using attorneys or self representing themselves . Next highest, 1658, belongs to the Independent Defender’s Office, which probably implies there is a criminal case connection to the divorce case.

Number of attorneys vs number of cases they represent.

The statistical measures for attorneys and cases are as follows (after ignoring the cases where both parties are self represented ).

weighted average: 9.99
weighted median: 60
median: 2
mode: 1

From the above plots and table it is evident that number of attorneys with more than 1 case rapidly decreases — only 80 attorneys have 10 cases and only 17.2 % have 10 or more cases. An approximate fit of straight line in the log-log plot yields a mono-nomial (e.g. y = bx^a ) with the value for exponent -1/19.22 = -0.052, which implies a very steep drop.

Antagonism between case parties:

The intensity or contentiousness of a case can be indirectly measured by the number of court documents and court calendar dates. In general, more disagreement should result in more court filings and more presence in court dates.

The following plots shows the number of documents vs number of cases.

Number of documents vs number of cases.

The following table lists the salient statistical measures for number of documents vs number of cases.

weighted average: 13.73
weighted median: 26
median: 7
mode: 1

In the log-log plot above, appears to have 2 regions. The first region with higher number of documents with a slower decreases, and the second region with lower number document with a steeper decrease. The number of cases in the first region is much smaller. The first region probably represents high-conflict cases. The inflection point between the 2 regions appears to be about 8000 documents and 10 cases.

The following plots show number of calendar events vs number of cases. In general there are many more cases with lower number of calendar event.

Number of calendar event vs number of cases.

The following lists the salient statistical metrics. The 0 for mode probably reflects the scenario where no additional documents have been filed after the suite was opened.

weighted average: 3.19
weighted median: 13
median: 1
mode: 0

Presence of minors:

The follow plot shows the number of minors in a case vs number of cases. It is interesting to see a downward trend here too. This likely follows the distribution of children in the greater general population of the county — less number of households have more children. 0 minors had the highest number of cases, followed by 1 and 2 minors.

Number of minors in a case vs number of cases.

The following lists the salient statistical metrics on minors. The 1 for mode probably reflects the scenario where no additional documents have been filed after the suite was opened.

weighted average: 0.86
weighted median: 2
median: 1
mode: 0

Case duration:

The next metric we look at is the distribution of duration. The duration is range when the first document was filed or calendar date scheduled to date of the last document filed or calendar date scheduled.

The following plot shows the duration of a case vs number of cases when calendar date is used.It appears vast majority of cases are under 21 years old. Further more, number of cases start drastically falling at about 1 years, and further at about 12 years. These durations have significances. 1 year signify the parties are able to resolve differences quickly and amicably , 12 years signify onset of teenage and 21 years signify full adulthood. Both 12 and 21 years probably imply significant weight placed on the minors’ voices.

Duration of cases vs number of cases.

The following lists the salient statistical metrics on duration of cases . The 0 for mode probably reflects the scenario where no additional events were scheduled after the suite was opened, and the case was completed or a case just filed (although these cases expected to be very low).

weighted average: 1.76
weighted median: 7.33
median: 0.25
mode: 0.0

Very similar trends are observed when duration is based on the date of the documents filed.

Duration of cases vs number of cases.

The following lists the salient statistical metrics on duration of cases . The 0 for mode probably reflects the scenario where no additional events have been scheduled after the suite was opened, and the case was completed or recently filed cases ( expected to be very low )

weighted average: 2.35
weighted median: 6.75
median: 0.83
mode: 0.0

The following plots show the case duration vs number of cases parameterized for 0 through 3 number of minors . Qualitatively, there appears to be no difference among cases with 0 minors to those with 1–3 minors. This is somewhat counter intuitive to the notion that divorce with minors is more acrimonious, hence longer. This similarity may be due to property and alimony issues, which also tend to be fairly acrimonious.

Duration of cases vs number of cases for minors involved from 0 through 3.

The following lists the salient statistical metrics on duration of cases for 0 through 3 minors involved. From these numbers, cases with 0 minor appear to have less lengthy duration in mean and median.

number of minors: 0
number of cases: 91416
weighted average: 1.06
weighted median: 5.75
median: 0.083
mode: 0.0

number of minors: 1
number of cases: 56436
weighted average: 2.16
weighted median: 7.83
median: 0.33
mode: 0.0

number of minors: 2
number of cases: 29593
weighted average: 2.50
weighted median: 7.75
median: 0.58
mode: 0.0

number of minors: 3
number of cases: 9869
weighted average: 2.81
weighted median: 8.3
median: 0.67
mode: 0.0

Conclusions:

The following are some conclusions that could be drawn based on the above.

  1. The number of cases per attorney appears to follow a mono-nomial distribution.
  2. The case duration appears to have inflection points at about 1, 12 and 21 years. Probably corresponding to quick amicable resolution, onset of teenage years for minors, and attainment of complete adulthood respectively.
  3. There appears to be no qualitative differences in case duration for cases with minors and cases with minors. However, a reduction is average and median duration is observed for cases with no minors.

--

--