3 ms·

I just yesterday skimmed the documentation for statsmodels weighted generalized linear models and they use data from an affair study there: https://www.statsmod

by fho 4y ago

I just yesterday skimmed the documentation for statsmodels weighted generalized linear models and they use data from an affair study there: https://www.statsmodels.org/devel/examples/notebooks/generated/glm_weights.html https://www.statsmodels.org/devel/examples/notebooks/generat... I have not looked at the original study, but from what I gather from the results basically nobody is having an affair (the median for time spend in an affair is still 0, the 3/4 quantile is still only 0.48 (time units) and the maximum is 57.60 (time units)). I am not sure what "time units" is ... it is probably in the original paper, but I did not bother to look that up. Even if it is years, that would mean that 75% of the people spend less than half a year in an affair over their lifetime. Once married this drops significantly and reduces further over the years of marriage.

conductr 4y ago
I think my wife has a rather binary view on affairs. Doesn’t need to be long, sustained, or repetitive. I’d wager many wives would be livid just by the presence of an account on AM even if no physical cheating occurs. Just the fact it was sought out or investigate is going to cause a riff in the relationship. Also if the models show nobody is cheating, the data should be considered suspicious. Maybe the word “affair” is an issue. It insinuates a long term situation when really a lot of this falls under hookups/prostitution.
- fho 4y ago
  > Also if the models show nobody is cheating, the data should be considered suspicious. Possible ... when the cost of getting caught is high, even if it is just a perceived risk, I would expect that people "cheat" on those questionnaires.
  - conductr 4y ago
    Exactly. I wouldn’t trust this data to be truthful to begin. We’re talking about the “worlds oldest profession”, I’d need proof of that. Even a place like Saudi has a known practice and general understanding that sinning goes down in Bahrain. The traffic on the bridge is all the evidence I needed to see in order to believe that.
- paulryanrogers 4y ago
  In some religious circles the mere thought of someone else is adultery. (Like you can get a divorce because they looked at others at the beach too often.) Which seemed extreme to me even when I used to believe. Oh and if you're single and 'lust' after someone else then you're also eternally damned. That always made ~dating~ [courtship] awkward.

by fho 4y ago

conductr 4y ago

I think my wife has a rather binary view on affairs. Doesn’t need to be long, sustained, or repetitive. I’d wager many wives would be livid just by the presence of an account on AM even if no physical cheating occurs. Just the fact it was sought out or investigate is going to cause a riff in the relationship. Also if the models show nobody is cheating, the data should be considered suspicious. Maybe the word “affair” is an issue. It insinuates a long term situation when really a lot of this falls under hookups/prostitution.

fho 4y ago
> Also if the models show nobody is cheating, the data should be considered suspicious. Possible ... when the cost of getting caught is high, even if it is just a perceived risk, I would expect that people "cheat" on those questionnaires.
- conductr 4y ago
  Exactly. I wouldn’t trust this data to be truthful to begin. We’re talking about the “worlds oldest profession”, I’d need proof of that. Even a place like Saudi has a known practice and general understanding that sinning goes down in Bahrain. The traffic on the bridge is all the evidence I needed to see in order to believe that.
paulryanrogers 4y ago
In some religious circles the mere thought of someone else is adultery. (Like you can get a divorce because they looked at others at the beach too often.) Which seemed extreme to me even when I used to believe. Oh and if you're single and 'lust' after someone else then you're also eternally damned. That always made ~dating~ [courtship] awkward.