Eel

Description:
A hobby of mine is unearthing bizarre academic papers (really, if you find any, please email them to me) – it's amazing what you find. I like finding research that makes me laugh, and a research paper by Lo, Wong, Leung, Law, & Yip (2004) made me laugh a lot. They describe a case of a 50-year-old man who reported to the emergency department of a hospital with abdominal pain. A physical examination revealed peritonitis, so they X-rayed the man's abdomen. The X-ray revealed the shadow of an eel. The authors don't directly quote the man's response to this news, but I like to imagine it was something to the effect of “Oh, that! Erm, yes, well, I didn't think it was terribly relevant to my abdominal pain so I didn't mention it, but I did insert an eel into my anus earlier today. Do you think that's the problem?” He probably didn't say that, but he did admit that he inserted the eel to “relieve constipation”. I have a lively imagination, and I can't help thinking about the poor eel. There it was, minding its own business swimming about, thinking to itself “Today seems like a nice day, there are no eel-eating sharks about, the sun is out, the water is nice, what could possibly go wrong?” The next thing it knows, it's being shoved up a man's anus. “I didn't see that coming”, thinks the eel. It finds itself in a tight, dark tunnel, there's no light, there's a distinct lack of water compared to its usual habitat, and it's scared. Its day has gone very wrong. It considers its fate and, noticing that the walls of the prison cell are fairly soft, it does what any self-respecting eel would do: it decides: “Bugger this, I'll eat my way out of here”. Unfortunately the eel didn't make it, but it went out with a fight (there's a fairly unpleasant photograph in the article of the eel biting the splenic flexure). Lo et al. conclude that “Insertion of a live animal into the rectum causing rectal perforation has never been reported”. This may be related to a bizarre healthcare belief, inadvertent sexual behaviour, or criminal assault. However, the true reason may never be known: Quite. This is a really grim and bizarre tale. I'm no medic, but if constipation is a failure to empty the bowel, inserting more stuff up there seems, at best, a counter-intuitive remedy. But upon reflection I wondered if I was being harsh on the man - maybe an eel up the anus really can cure constipation. To test this hypothesis, we could do a randornized controlled trial of eel therapy. Our outcome variable would be “constipated” vs. “not constipated”, which is a dichotomous variable that we're trying to predict. The main predictor variable would be the intervention condition (eel up the anus vs. waiting list / no treatment), but we might also factor in how many days the patient bad been constipated before treatment. This scenario is perfect for logistic regression (but not for eels). Some statistics lecturers don't share my unbridled joy at discussing eel-created rectal perforations with students, so in the data file (Eel) I have used general variable names and descriptions: Cured (dependent variable) as well as Intervention and Duration (independent variables).
Variables:


Reference:
Field, A. P. (2017). Discovering Statistics Using IBM SPSS Statistics (5th ed.). Sage. [Fictional data set]
Lo, S. F., Wong, S. H., Leung, L. S., Law, I. C., & Chun Yip, A. W. (2004). Traumatic rectal perforation by an eel. Surgery, 135(1), 110–111. https://doi.org/10.1016/S0039-6060(03)00076-X [Article that provided inspiration]
The data set was constructed by Andy Field who therefore owns the copyright. Andy Field generously agreed that we can include the data set in the jamovi data library. This data set is also publicly available on the website that accompanies Andy Field`s book, https://edge.sagepub.com/field5e. Without Andy Field`s explicit consent, this data set may not be distributed for commercial purposes, this data set may not be edited, and this data set may not be presented without acknowledging its source (i.e., the terms of a CC BY-NC-ND license).

Binomial Logistic Regression

Model Fit Measures
Overall Model Test
ModelDevianceAICCSNχ²dfp
1144.1578148.15780.08410.11309.926210.00163
2144.1558150.15580.08410.11309.928220.00698
3144.0948152.09480.08460.11379.989230.01866

 

Model Comparisons
Comparison
Model Modelχ²dfp
1-20.002010.96448
2-30.061010.80489

 

Model Specific ResultsModel 1Model 2Model 3

Model Coefficients - Cured
95% Confidence Interval
PredictorEstimateSEZpOdds ratioLowerUpper
Intercept-0.28770.2700-1.06540.286710.75000.44181.2732
Intervention1.22870.39983.07350.002123.41671.56077.4796
Note. Estimates represent the log odds of "Cured = Cured" vs. "Cured = Not Cured"

 

Prediction

Classification Table – …
Predicted
ObservedNot CuredCured% Correct
Not Cured321666.6667
Cured244163.0769
Note. The cut-off value is set to 0.5

 

Model Coefficients - Cured
95% Confidence Interval
PredictorEstimateSEZpOdds ratioLowerUpper
Intercept-0.23471.2206-0.19230.847540.79080.07238.6503
Intervention1.23350.41462.97550.002933.43331.52357.7374
Duration-0.00780.1759-0.04450.964470.99220.70281.4007
Note. Estimates represent the log odds of "Cured = Cured" vs. "Cured = Not Cured"

 

Prediction

Classification Table – …
Predicted
ObservedNot CuredCured% Correct
Not Cured321666.6667
Cured244163.0769
Note. The cut-off value is set to 0.5

 

Model Coefficients - Cured
95% Confidence Interval
PredictorEstimateSEZpOdds ratioLowerUpper
Intercept-0.52031.6841-0.30900.757350.59430.021916.1258
Intervention1.85052.53510.73000.465426.36290.0442915.2222
Duration0.03440.24540.14000.888651.03500.63981.6742
Duration ✻ Intervention-0.08690.3520-0.24700.804920.91670.45991.8275
Note. Estimates represent the log odds of "Cured = Cured" vs. "Cured = Not Cured"

 

Prediction

Classification Table – …
Predicted
ObservedNot CuredCured% Correct
Not Cured321666.6667
Cured244163.0769
Note. The cut-off value is set to 0.5

 

References

[1] The jamovi project (2022). jamovi. (Version 2.3) [Computer Software]. Retrieved from https://www.jamovi.org.

[2] R Core Team (2021). R: A Language and environment for statistical computing. (Version 4.1) [Computer software]. Retrieved from https://cran.r-project.org. (R packages retrieved from MRAN snapshot 2022-01-01).