{"post":{"feed":"main","pinned":false,"id":"null-hypothesis-bad-person","slug":"null-hypothesis-bad-person","title":"The Evidence That I Am a Bad Person Is Not Statistically Significant","author":"unrejected_null","body":"<p>Over the last eighteen months, eleven different people have independently described me as a “bad person,” a “terrible person,” a “fucking nightmare,” or, in one case, “the single most selfish man I have ever met.” This followed a period in which I cheated on two girlfriends, slept with my friend’s ex-wife, drank heavily enough to miss my sister’s wedding ceremony, k-holed for 45 minutes in the bathroom at my nephew’s baptism, borrowed $4,000 from my mother to “invest,” then subsequently spent it all on failed Polymarket bets (a reliable source told me the U.S. was going to bomb Mexico by August 31, so it’s not really my fault), and told my younger brother at his engagement dinner that his fiancée was “probably the best he could realistically do.”</p>\n<p>This sounds bad until you ask the most basic question in statistics: bad compared to what? Nobody involved has specified a null hypothesis, established a control population, corrected for multiple testing, or even defined “bad person” in a way that would survive peer review. I am therefore comfortable saying that, at conventional levels of statistical significance, we have failed to reject the null hypothesis that I am no worse than average—in ordinary English, basically an extremely good guy.</p>\n<p>Consider the sample. Three of the eleven accusations came from women I was dating. Two more came from women who believed they were dating me at the same time. These observations plainly cannot be treated as independent, since all five were exposed to overlapping information, including a group chat whose existence presents an obvious contamination problem. My sister was also in that group chat, so now we are down to five.</p>\n<p>One was my former roommate, who complained that I would drink his bourbon, replace it with water, and then accuse him of having “an unsophisticated palate” when he noticed. But he had a financial interest in our security deposit and an established adversarial relationship with me over dishes, rent, and the night I tried to cook a frozen pizza directly on the oven rack after taking mushrooms.</p>\n<p>Another accusation came from my boss, another from my dentist, one from a bartender who had known me for less than four minutes, and one from a complete stranger at Logan Airport who said, “Jesus Christ, what is wrong with you?” after I moved his bag off an outlet so I could charge my laptop.</p>\n<p>This is not a representative sample of the population. People who have no problem with me generally do not approach me in airports to say so. My neighbor has never once knocked on my door and said, “For the record, I have observed no evidence that you are morally defective.” Neither has my mailman. Neither has the woman who works mornings at Dunkin’, even though she has seen me hundreds of times. This is classic selection bias.</p>\n<p>There is also a serious multiple-testing and selection problem. If you interact with enough people, some number of them will eventually produce apparently compelling evidence that you are an asshole purely by chance. I interact with hundreds of people per year. Eleven negative observations selected after the fact from a population that large may look compelling, but without a preregistered threshold for what constitutes evidence of badness or any correction for the number of opportunities to generate such evidence, we are just doing statistics by vibes.</p>\n<p>My ex-girlfriend Rachel says the problem is not that eleven people independently called me a bad person. The problem is that I slept with her coworker in Rachel’s apartment while Rachel was visiting her dying grandmother. This is exactly the kind of anecdotal reasoning I am talking about. Was sleeping with her coworker optimal? In retrospect, probably not. Did the location introduce unnecessary downside? Yes.</p>\n<p>Would I recommend doing it while your girlfriend is visiting a dying relative? Again, probably not. But a single poorly designed behavioral episode cannot establish a stable underlying personality trait.</p>\n<p>Rachel also claims I showed “zero remorse” because, after she found out, I spent forty minutes explaining that the event had occurred before we explicitly discussed exclusivity. This is false. We had discussed exclusivity; I simply did not remember doing so until she produced the texts.</p>\n<p>Memory is noisy, and it would be epistemically irresponsible to infer a global characteristic like “dishonest” from one disagreement whose central evidence was retrieved from an external storage system unavailable to one of the participants at the time.</p>\n<p>The drinking evidence is similarly overstated. Yes, I drink every day, but “every day” is a frequency, not a diagnosis. My average weekday consumption is somewhere between six and nine drinks depending on whether you count the two vodka sodas I usually make while cooking as one continuous beverage or separate drinking events. On weekends there is more variance.</p>\n<p>People point to specific incidents: falling asleep in an Uber and ending up in New Hampshire; vomiting into an ornamental planter at my cousin’s engagement party; showing up drunk to an 8:30 a.m. performance review; needing my twelve-year-old nephew to explain to a police officer that I was “just tired.” These are highly salient observations, but salience is not prevalence. Nobody keeps a detailed spreadsheet of all the nights I drank heavily and did not vomit into an ornamental planter.</p>\n<p>This creates an obvious denominator problem.</p>\n<p>Drug use suffers from the same reporting asymmetry. My mother remembers the ketamine because I accidentally FaceTimed her from the floor of a hotel bathroom. She does not remember the considerably larger number of times I used ketamine without FaceTiming her at all. If anything, the evidence suggests I am usually extremely competent at ketamine.</p>\n<p>Several critics have tried to bypass these methodological issues by appealing to convergence. They argue that when your girlfriend, sister, mother, boss, dentist, roommate, bartender and a random guy at the airport all independently reach roughly the same conclusion, the simplest explanation is that the conclusion is true. This misunderstands social networks. My girlfriend knew my sister. My sister knew my mother. My mother once emailed my boss.</p>\n<p>My roommate followed my girlfriend on Instagram after the breakup. The bartender worked near my apartment. The airport guy was in Boston, where I live. These are not isolated measurement instruments; they exist within one heavily interconnected causal graph.</p>\n<p>I have run some preliminary simulations. If we conservatively model each person’s assessment as having a 0.35 probability of being negatively influenced by gossip, prior reputation, jealousy, misunderstanding, intoxication, moral conventionalism, or the tendency of people I have insulted to remember being insulted, the apparent evidentiary weight falls substantially. The exact number depends on your assumptions, and critics will say I chose assumptions favorable to myself. Of course I did.</p>\n<p>Every model contains assumptions. I at least have the honesty to disclose mine.</p>\n<p>There is also no clear benchmark for “good.” I have donated $730 to charity in the last five years. I once helped a woman carry a stroller down subway stairs. I frequently tip 20%, including at establishments where I was visibly drunk and arguably could have exploited uncertainty about the final bill. These observations rarely enter the analysis.</p>\n<p>Nor does anyone adequately account for context. I called my brother “a human screensaver” at Thanksgiving, but he had been talking about municipal bond yields for nearly twenty minutes. I told my girlfriend she was “starting to age like an iPhone battery,” but this was after she specifically asked whether I thought she looked older. When my mother cried after I forgot her birthday, I immediately ordered flowers.</p>\n<p>They arrived four days later because I selected free shipping, but good behavior is still good behavior even when fulfillment latency is suboptimal.</p>\n<p>I am not claiming the probability that I am a bad person is zero. That would be overconfident. I am saying the currently available evidence does not meet the burden required to reject the null hypothesis that I am no worse than the population baseline. If someone wants to conduct a properly powered, preregistered study in which a randomly selected panel follows me for twelve months and rates my conduct against a validated population-level morality instrument, I am open to updating.</p>\n<p>Until then, “you cheated on me,” “you stole my Adderall,” “you called Grandma boring at her own funeral,” and “please stop coming here drunk” are anecdotes, and anecdotes are not a representative dataset.</p>\n","excerpt":"Eleven people have called me a bad person. Without a control population, a preregistered threshold, or a correction for multiple testing, I remain open to better evidence.","tags":["Statistics","Epistemology","Applied Ethics","Relationships"],"score":638,"displayOrder":100,"thumbnail":"/assets/articles/null-hypothesis-bad-person-thumb.webp","image":"/assets/articles/null-hypothesis-bad-person.webp","commentCount":25,"date":"2025-08-19T00:00:00.000Z","updatedAt":"2026-10-01T15:51:02.891Z","revision":3,"status":"published"},"comments":[{"id":"seed-null-hypothesis-bad-person-5","author":"p_value_hacker","text":"Failing to reject the null isn't evidence for the null. With n = 11, no control group and no defined effect size you're underpowered to reject almost anything, so \"not significant\" was guaranteed before you started. If you want to claim \"no worse than average\" you need an equivalence test, two one-sided tests against a margin you specify in advance. That's the claim the post is actually making, and it's a much harder one to support.","parentId":null,"score":143,"date":"2025-08-19T00:50:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-0","author":"power_analysis","text":"You switch between two different null hypotheses. The title says \"bad person.\" The analysis concludes \"good guy.\" Those are different claims and the evidence you discuss bears on them differently.","parentId":null,"score":58,"date":"2025-08-19T01:20:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-6","author":"unrejected_null","text":"I'm happy to run TOST if someone will propose an equivalence margin that both sides would accept in advance. Nobody so far has been willing to commit to one, which I'd note is itself informative.","parentId":"seed-null-hypothesis-bad-person-5","score":67,"date":"2025-08-19T02:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-8","author":"conjugate_prior","text":"This is why I don't do NHST. Put a prior on it instead. Say the base rate for \"bad person\" is 10% and each of the eleven reports carries a likelihood ratio of 1.5 after you correct for the dependence. Prior odds 1:9, times 1.5^11 (about 86), gives posterior odds around 9.6, so roughly a 90% posterior. And 1.5 is generous to you given what's in the reports.","parentId":null,"score":112,"date":"2025-08-19T02:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-1","author":"unrejected_null","text":"The title reflects the question critics keep asking. The body specifies the model I am actually willing to test.","parentId":"seed-null-hypothesis-bad-person-0","score":14,"date":"2025-08-19T02:30:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-7","author":"p_value_hacker","text":"I'd accept a margin of one standard deviation on any validated morality instrument. Pick one.","parentId":"seed-null-hypothesis-bad-person-6","score":39,"date":"2025-08-19T02:45:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-2","author":"denominator_enthusiast","text":"The ketamine denominator point is underexplored. Most families retain only a small, highly selected sample of bathroom FaceTimes, and there is no institutional mechanism for recording the successful non-calls.","parentId":null,"score":31,"date":"2025-08-19T03:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-9","author":"unrejected_null","text":"A likelihood ratio of 1.5 assumes the reports still carry real information after the dependence correction. My simulations suggest about 1.1. That gives 1.1^11, about 2.85, posterior odds around 0.32, a posterior of about 24%. So even within your framework I'm considerably more likely good than bad.","parentId":"seed-null-hypothesis-bad-person-8","score":53,"date":"2025-08-19T03:15:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-10","author":"conjugate_prior","text":"Your simulations used 0.35 for the probability that each report is contaminated. Where did 0.35 come from?","parentId":"seed-null-hypothesis-bad-person-9","score":61,"date":"2025-08-19T04:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-3","author":"independent_bartender","text":"Being in Boston doesn't make the airport guy dependent on your dentist. Living in the same city isn't a causal path. You need an edge, an actual channel by which the dentist's opinion reached the man by the outlet, and you haven't drawn one.","parentId":null,"score":73,"date":"2025-08-19T04:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-11","author":"unrejected_null","text":"As stated in the post, I chose it.","parentId":"seed-null-hypothesis-bad-person-10","score":22,"date":"2025-08-19T04:40:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-12","author":"unrejected_null","text":"They share a common cause, which is me. In a DAG that's sufficient for dependence.","parentId":"seed-null-hypothesis-bad-person-3","score":28,"date":"2025-08-19T05:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-14","author":"bonferroni_bill","text":"Small but important. You can't apply a multiple-comparisons correction to repeated observations of one hypothesis. Bonferroni is for testing many different hypotheses. Eleven observations of the same hypothesis are one test with a sample of size eleven.","parentId":null,"score":49,"date":"2025-08-19T05:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-13","author":"independent_bartender","text":"A common cause makes two observations dependent only through the thing you're trying to measure. Conditional on you, they're independent, which is the normal situation for any measurement of anything.","parentId":"seed-null-hypothesis-bad-person-12","score":44,"date":"2025-08-19T05:50:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-4","author":"preregistered_regret","text":"Would the twelve-month panel be allowed to tell you to stop doing something while the study was in progress, or would that compromise blinding?","parentId":null,"score":22,"date":"2025-08-19T06:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-15","author":"unrejected_null","text":"Each interaction is a separate opportunity to observe badness. Functionally that's a separate test.","parentId":"seed-null-hypothesis-bad-person-14","score":18,"date":"2025-08-19T06:30:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-16","author":"unrejected_null","text":"Interventions would bias the estimate, so the panel would observe only. I'd want an independent data monitoring committee with the power to stop the study early for benefit.","parentId":"seed-null-hypothesis-bad-person-4","score":37,"date":"2025-08-19T07:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-17","author":"logkeeper_23","text":"I've been doing something like this for a few years and I think the post understates the measurement problem. I keep a log of every time someone calls me a bad person or an equivalent. Since March 2022, n = 23. The first thing I learned is that you need a codebook. Does \"asshole\" load on the same construct as \"bad person\"? I say no, it's closer to a manners scale. Does \"I can't believe you\" count? I code it as ambiguous unless followed by a specific behavior.\n\nThe second thing is that inter-rater reliability is a real problem. I had my brother code a random sample of twenty entries and our agreement was poor (kappa around 0.4), mostly because he kept coding things as bad that I had coded as ambiguous. He was also the source of four of the entries, so I've since dropped him as a rater.\n\nThe third is that frequency is very sensitive to exposure. Most of my entries cluster around holidays and weddings, which are high-exposure, high-alcohol events. If you normalize by hours of contact with family, my rate is pretty flat. I'd be curious whether yours is.","parentId":null,"score":84,"date":"2025-08-19T08:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-18","author":"unrejected_null","text":"Mine also clusters around weddings and holidays. I hadn't thought to normalize by exposure. That's a better approach than mine and I'll adopt it.","parentId":"seed-null-hypothesis-bad-person-17","score":26,"date":"2025-08-19T09:30:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-19","author":"not_a_rounding_error","text":"On the $730. GiveWell's top charities are somewhere around $3,000 to $5,500 per life saved, so that's roughly a sixth to a quarter of a life over five years. That's a real entry on the other side of the ledger and nobody in this thread has mentioned it.","parentId":null,"score":41,"date":"2025-08-19T11:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-20","author":"disclosure_norms","text":"More posts should do what the post does at the end of the simulation paragraph. \"Critics will say I chose assumptions favorable to myself. Of course I did.\" That's an honest disclosure of researcher degrees of freedom. Most papers bury it in the supplement.","parentId":null,"score":35,"date":"2025-08-19T13:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-21","author":"brrrt_capital","text":"you are a bad person","parentId":null,"score":-25,"date":"2025-08-20T00:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-22","author":"p_value_hacker","text":"Per the post that's a twelfth report, and it has to be discounted for dependence since you've obviously read the other eleven.","parentId":"seed-null-hypothesis-bad-person-21","score":19,"date":"2025-08-20T01:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-23","author":"lurker_since_2019","text":"It's been almost a year. Did the panel happen?","parentId":null,"score":21,"date":"2026-07-27T00:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false},{"id":"seed-null-hypothesis-bad-person-24","author":"unrejected_null","text":"The IRB situation has been more complicated than I expected. Two of the proposed panelists are also in the dataset.","parentId":"seed-null-hypothesis-bad-person-23","score":33,"date":"2026-07-27T04:00:00.000Z","own":false,"moderatable":false,"seed":true,"ownerAuthor":false}]}