Credibility and Burden of Proof in Administrative Hearings: Why Identical Cases Get Opposite Outcomes | Administrative Remedies

Two people walk into Social Security disability hearings on the same day with identical medical conditions, identical evidence, and attorneys representing both. One walks out with benefits. The other doesn't. The difference isn't the law or the facts — it's whether the judge believed them.

Credibility and burden of proof are the two analytical factors that decide what happens once a claimant is in front of a judge. Credibility determines how much of someone's testimony the judge actually credits. Burden of proof determines how much credited evidence they need to win. Together, they explain why cases that look identical on paper can come out completely differently — and why the same credibility profile can be survivable in one administrative system and fatal in another.

This episode picks up where the lifecycle episode left off, moving from how a record gets built to how it gets evaluated. Social Security overhauled its credibility framework in 2016 with SSR 16-3P, telling adjudicators to focus on whether reported symptoms are consistent with the medical evidence rather than whether the claimant is "a truthful person." But the inquisitorial structure of agency hearings — where the ALJ is investigator, questioner, and judge in the same proceeding — makes pure consistency analysis nearly impossible. The judge spends 45 minutes in a room with the claimant, watching them shift in their chair, hearing how they describe their pain. Demeanor findings get baked into the analysis even when the regulation says they shouldn't, and they're nearly unreviewable on appeal because the appellate court wasn't in the room.

The burden side raises a deeper structural problem. The Administrative Procedure Act says the proponent of an order bears the burden of proof — straightforward in an OSHA enforcement case where the agency is the proponent, but in a Social Security case the claimant is the proponent and the agency gets to define what evidence counts. That creates the worst of both worlds: an inquisitorial system where the judge has a duty to develop the record, but an adversarial-style burden rule where the claimant loses if the evidence comes out 50-50. Civil law systems that use the inquisitorial model don't operate this way. The U.S. borrowed the structure without borrowing the burden allocation that made it coherent.

The episode closes by comparing how these factors interact across different administrative systems. Social Security uses preponderance of the evidence — about 51% — but claimants typically have a paper trail of medical records that can survive a wobbly hearing. Asylum cases technically use a more generous standard — the Supreme Court said in Cardoza-Fonseca that as little as a 10% chance of persecution can satisfy the well-founded fear requirement — but asylum applicants often have no paper trail at all. The result is counterintuitive: the system with the lower formal threshold can be harder to win, because credibility carries the entire case, and trauma symptoms like flat affect and fragmented memory get read as evasiveness.

Listen Now

Spotify | Apple Podcasts | Listen on our site

What We Cover

Why two claimants with identical medical conditions and identical evidence can get opposite outcomes — and what credibility actually means in administrative law
How SSR 16-3P tried to shift Social Security from "do I believe this person" to "are these symptoms consistent with the record" — and why the gap between regulation and practice persists
The seven things ALJs are actually looking for: prior work history, statements over time, medical source observations, treatment gaps, demeanor, specificity, and acknowledgment of capability
Why an unexplained treatment gap looks like evidence that symptoms aren't severe — and why the same gap with an explanation becomes a barrier-to-care problem
How social media, surveillance, and pattern-of-overstatement evidence can collapse a case, and why one dramatic inconsistency matters less than a cumulative pattern
Why courts require ALJs to identify specific testimony they're rejecting and connect it to specific contradicting evidence — not just summarize the medical record
How the APA's burden rule puts the burden on the proponent — meaning the claimant in benefits cases and the government in enforcement cases
Why the Social Security system inherited the risks of both adversarial and inquisitorial models without the protections of either
How preponderance, clear and convincing, and beyond a reasonable doubt map onto different administrative contexts based on the severity of consequences
Why the same credibility problems are survivable in Social Security but often fatal in asylum cases — and what that says about how standards of proof actually function

Full Rough Transcript

Gwen: Hello, and welcome to Administrative Remedies, because you can’t fix what you don’t understand. Brought to you in part by the University of Tulsa College of Law. I’m Gwendolyn Savitz, an associate professor here at TU and the associate dean of research and intellectual life.

Marc: And I’m Marc Roark. I’m the dean of the College of Law.

Gwen: We’ll be breaking down complex doctrines with real-life analogies and examples to demystify the world of administrative law for everyone trying to understand how government actually works.

Marc: Agencies are the main way the federal government gets things done. It’s not through Congress for reasons we’ll be addressing over the course of this series.

Gwen: Marc, I’d like you to imagine two Social Security disability hearings happening on the same day in the same building in different hearing rooms.

Marc: OK.

Gwen: So room A, you have Sarah. She testifies that she has debilitating back pain that prevents her from working. She has medical records that show she has a herniated disc and the ALJ denies her claim. In room B, you have Michael. He testifies that he has debilitating back pain that prevents him from working and also has medical records showing a herniated disc. But the ALJ approves his claim.

Marc: So same medical condition, but different outcomes.

Gwen: Right. Same condition, same evidence. In this case, we’ll say they’re both represented by attorneys. They both had vocational experts testify. But one gets benefits and one doesn’t.

Marc: Okay. What’s the difference?

Gwen: The ALJ believed Michael. The ALJ didn’t believe Sarah. Credibility determined everything.

Marc: That seems deeply problematic. I mean, how can something that important come down to whether a judge personally believes you?

Gwen: It is problematic, but it’s also unavoidable. The last time we talked about Kathleen’s case, we followed it through the whole system, from when she applied to when she was ready to go to federal court. Today, we’re talking about why one person walks out with a favorable decision and another one doesn’t. We’re specifically looking at two factors, credibility, whether the judge believes you, and burden of proof, who has to prove what and how convincing the evidence has to be.

Marc: So the analytical framework behind all of these different outcomes.

Gwen: Right. And once you understand how judges think about these two factors, you can understand why cases that look identical on paper can start to come out completely differently.

Marc: We should start with what we actually mean when we’re saying that a judge is evaluating credibility.

Gwen: What we normally mean is that credibility means whether you’re telling the truth.

Marc: All right.

Gwen: Yes. Here, it’s a little more layered than that. So it’s not just whether the person is intentionally lying, although if they are, they don’t have a lot of credibility. But it’s also, are they trying to be slightly deceptive? Are they just honestly mistaken about the facts? Are they overstating symptoms? Or are they minimizing things, saying they can’t do stuff that they seem to be capable of? Or even their memory. Is their memory reliable? Do their perceptions seem trustworthy?

Marc: So someone could be completely honest and still have credibility issues.

Gwen: Right. You could genuinely believe that you can’t lift 10 pounds. But if medical evidence shows you can lift 20, you’re not lying. You’re mistaken. But the judge still won’t credit your testimony.

Marc: So why does credibility matter so much in administrative cases specifically?

Gwen: Because here, so many of these cases are turning on subjective facts. The medical record might show a herniated disc. That’s objective. The MRI could prove that. But how much pain does it cause? There’s the real problem. How limiting is it? How much does it prevent you from working? That’s subjective. Only you can testify to that. And if the judge doesn’t believe you, you lose, even if you genuinely experience those symptoms.

Marc: So is there any actual regulatory guidance on how judges can evaluate this, or is this just intuition?

Gwen: There is detailed guidance. Social Security overhauled the framework here in 2016 with SSR 16-3P, which replaced an older ruling that used the word credibility explicitly. And they changed it pretty significantly. Here’s the key language.

Marc: In evaluating an individual’s symptoms, our adjudicators will not assess an individual’s overall character or truthfulness in the manner typically used during an adversarial court litigation. The focus of the evaluation of an individual’s symptoms should not be to determine whether he or she is a truthful person.

Marc: So it’s not about whether they’re a truthful person, then what is it about?

Gwen: Here, we’re talking about whether there’s symptom consistency. The regulatory intent was to shift from, do I believe this person, to are these reported symptoms consistent with the medical evidence and the record as a whole? That’s a meaningful distinction. They were trying to make it more analytical and less character-based.

Marc: But how does that work in practice?

Gwen: In practice, the ALJs are still making character judgments. It’s hard not to when you’re sitting across from someone for 45 minutes watching their body language and listening to how they describe their limitations. The regulation says to focus on consistency, but human psychology says you’re going to form an impression of the whole person.

Marc: That gap between what the regulation requires and what actually happens in the hearing room is one of the central tensions in administrative adjudication.

Gwen: So with that, what are the judges actually looking for? Let’s look at the regulatory framework, because these aren’t supposed to be judicial hunches. They’re legally required considerations.

Marc: 20 CFR 404.1529: Because symptoms such as pain are subjective and difficult to quantify, any symptom-related functional limitations and restrictions that your medical sources or non-medical sources report, which can reasonably be accepted as consistent with the objective medical evidence and other evidence, will be taken into account in reaching a conclusion as to whether you are disabled. We will consider all of the evidence presented, including information about your prior work record, your statements about your symptoms, evidence submitted by your medical sources, and observations by our employees and other persons.

Gwen: So prior work history, your statements, the medical sources, the judge’s observations, these are the categories that the regulation tells the ALJ to evaluate. Let’s talk about how they actually do it in practice.

Marc: So the regulation mentioned consistency.

Gwen: Yes, and that’s huge. That part of the regulation talks about consistency with the medical records, but here’s another part that’s worth looking at.

Marc: We will consider whether there are any inconsistencies in the evidence and the extent to which there are any conflicts between your statement and the rest of the evidence, including your history, the signs and laboratory findings, the statements by your medical sources or other persons about how your symptoms affect you.

Gwen: Here it’s explicitly talking about inconsistencies. What you said in your initial application versus what you told your doctor versus what you wrote in statements and what you testified at the hearing. Do they all match? If you told the doctor in 2022 you couldn’t walk two miles, but at the hearing in 2024 you say you’ve never been able to walk more than a block, judges notice. And they question everything else you say.

Marc: Okay, what if the condition, though, actually got worse?

Gwen: Then you have to explain it. In 2022, I could walk two miles, but since my surgery in 2023, I can only walk one block. With explanation, that inconsistency becomes consistent. What judges are looking for is whether you acknowledge and explain the change or whether you pretend the earlier statement never happened.

Marc: Okay, so what about the medical evidence then?

Gwen: Yeah, so we want things to be consistent with the medical evidence. If you testify that you have severe debilitating migraines three times a week, but your medical records show you’ve seen a doctor for headaches twice in the past year total, you’re not on any migraine medication, you haven’t had any imaging done, then the judge won’t believe the headaches are really that severe.

Marc: Right. Probably not.

Gwen: What if someone can’t afford treatment? Here, the context matters a lot. Normally, the judge might think if you truly had these really terrible migraines, you’d be seeking the most aggressive treatment you possibly could. But if you come in and say, I’d love to see a neurologist but I don’t have insurance and I can’t afford the $300 appointment, then that explains the gap. Judges are supposed to also look at barriers to care, but they only know about it if you tell them about it.

Marc: An unexplained treatment gap looks like evidence your symptoms aren’t that bad. An explained treatment gap is a barrier to access. Same facts, but completely different credibility implications.

Marc: OK, what about how you present yourself, your body language, your demeanor?

Gwen: Yes, this is the most subjective factor, and it can be kind of problematic. Judges are paying attention to eye contact. They’re paying attention to whether you seem genuine or if you’ve rehearsed this a lot, whether you’re forthcoming or evasive, whether you appear to be in pain while you’re testifying.

Marc: So that seems like it introduces an enormous bias. Some people are nervous. Some come from cultures where eye contact with authority figures is disrespectful. Trauma survivors might be emotionally flat. And some people are just better actors.

Gwen: That is totally true. What makes it worse is that demeanor findings can be essentially unreviewable on appeal. If the ALJ writes, I observed the claimant was evasive and avoided eye contact, an appellate court can’t contradict that, even if, as we’ll see, they can still try and play with the facts a little. They weren’t in the room.

Marc: So this most subjective, the most bias-prone factor can also be the one with the least appellate check.

Gwen: And the factor most likely to be wrong is the one most insulated from correction. Right. So courts have upheld demeanor as a permissible factor as long as it’s not the only basis for a credibility finding. But in practice, demeanor observations can color the entire analysis. If the ALJ’s first impression is this person seems evasive, that impression affects how they read every other piece of evidence.

Marc: Okay. Are there factors that cut the other way then, things that help credibility?

Gwen: Yes. There are two important ones. The first is specificity. Can you describe your experience in concrete detail? “It really hurts” is a pretty vague way of describing things. “It’s a sharp, stabbing pain that starts in my lower back and radiates down my right leg to my knee. It’s worse first thing in the morning, usually a 7 out of 10, but it spikes to a 9 if I’ve been standing for more than a few minutes.” That’s going to be pretty specific.

Marc: That second description, it sounds like someone who actually lives with that pain every day, not someone that’s reciting a script.

Gwen: Right. But in order to be able to talk like that, you have to have the language to describe it. Your response is how judges are generally going to react to it. Specificity suggests authenticity. Vagueness says you might be describing something you’ve been told to say rather than what you’ve actually experienced.

Marc: But you’ve probably had the experience of going to a doctor because of some sort of pain, and they try and supply potential descriptors. Is it sharp? Is it throbbing? You might not have really been able to describe it without them offering these possible options.

Gwen: And this isn’t just true about pain. This works across all other kinds of testimony. Specific details about your daily routine. Specific descriptions of what you can and can’t do. Specific examples of how your condition limits you. Concrete beats general basically every time.

Marc: Okay, what about the other?

Gwen: The other thing is admitting what doesn’t help you. This can seem counterintuitive to people, but we want to see whether you’re actually acknowledging your own capabilities. “I can’t do anything in the kitchen” sounds like an exaggeration. “I can make sandwiches and microwave meals, but I can’t stand long enough to cook a full dinner. My spouse does most of the cooking.”

Marc: Sounds like you’re being honest.

Gwen: Real disability has nuance. If you claim a total inability to do anything, judges get skeptical. If you honestly describe what you can and can’t do, you seem more trustworthy about the limitations you are describing.

Marc: So now we go to the other side, the thing that can really sink a case. And I want to focus on one in particular because it illustrates something important about how the modern system works.

Gwen: Social media?

Marc: It is social media. ALJs aren’t supposed to go into social media on their own, but it can be introduced by the claimant or it can be introduced through a fraud referral. This could be things like claiming you can’t get out of bed, but then posting yourself standing at a concert for hours.

Gwen: But going to one concert doesn’t mean you’re not disabled. People with chronic pain push through for important events all the time.

Marc: Right, and a good representative could make that argument. The problem wouldn’t automatically be the concert itself, but the contradiction. If you say bedridden most days but have photos contradicting that claim, you’re affecting your credibility. If this person testifies, I have severe pain most days, but occasionally I push through for something important and pay for it afterwards with days of recovery, that’s going to be a lot more consistent with the photos.

Gwen: The exaggeration destroys the credibility, not the activity itself.

Marc: So the lesson isn’t don’t go to concerts. It’s don’t overstate your limitations.

Gwen: Yes. And here it’s worth noting that this social media stuff, it’s used by insurance companies a lot more aggressively. They’ll actively look for things that seem to contradict what somebody’s claiming. Social Security really is supposed to be barred from it aside from these few instances, but it still matters.

Marc: And this can kind of connect to the broader pattern. What can destroy credibility isn’t usually one dramatic lie. It’s a pattern of overstatement that makes the judge question everything. You say you can’t walk, but surveillance shows you walking to your car. You say you can’t lift anything, but you just posted about helping a friend move. Each individual discrepancy might potentially be explainable.

Gwen: But as the judge starts to see these stack up, the cumulative effect can be devastating.

Marc: So the pattern matters more than any single inconsistency.

Gwen: Yeah. And once credibility collapses, it can take the whole case with it. The judge stops crediting your pain testimony, starts discounting your descriptions of limitations, and suddenly the residual functional capacity, the assessment of what you can still do, shifts dramatically. You can go from “can sit for four hours” to “can sit for six,” and that two-hour difference is enough to put a lot of jobs back on the table that you can potentially do.

Marc: Credibility doesn’t just affect whether the judge believes you then. It affects the factual findings that drive the entire legal analysis.

Gwen: Right. And that’s why it could be dispositive. But they have to do it correctly. If they don’t properly document why they’re finding certain things, it won’t work. And courts take this seriously. If an ALJ is going to discount your testimony, they can’t just say, I didn’t believe her and move on.

Gwen: As an example, there was a Ninth Circuit case where the Ninth Circuit didn’t believe the ALJ. The ALJ had denied disability benefits, but the entire credibility analysis was essentially, I’ve summarized the medical evidence and the claimant’s symptoms are not as bad as she says. Read what the court had to say about that.

Marc: We hold an ALJ does not provide specific, clear, and convincing reasons for rejecting a claimant’s testimony by simply reciting the medical evidence in support of his or her residual functional capacity determination. To ensure that our review of the ALJ’s credibility determination is meaningful and that the claimant’s testimony is not rejected arbitrarily, we require the ALJ to specify which testimony she finds not credible and then provide clear and convincing reasons, supported by evidence in the record, to support that credibility determination.

Marc: So, she said, I don’t believe her and then listed the medical evidence, but never connected the two, never said which specific testimony she rejected or which specific evidence contradicted it.

Gwen: Right. And the court was saying that’s not enough. You have to identify the specific testimony you’re rejecting and explain why. A summary of the medical evidence isn’t the same as an explanation of why you don’t believe the claimant.

Marc: So that kind of seems like it should be obvious.

Gwen: Yeah, you might think that, but this pattern, these cursory credibility findings followed by a medical evidence summary, this was so common that courts had to keep saying it over and over again. And the regulatory framework now reinforces this. The ruling we read earlier says adjudicators must provide specific reasons for the weight given to the individual’s symptoms. It must be consistent with and supported by the evidence and be clearly articulated so the individual and any subsequent reviewer can assess how the adjudicator evaluated the individual’s symptoms.

Marc: Okay. So the system is trying to discipline the credibility determination and force the adjudicator to show their work.

Gwen: Right. Because without that sort of discipline, this credibility determination just becomes a black box. The ALJ says not credible. Nobody knows why. And the reviewing court can’t determine whether these findings were based on legitimate inconsistencies or if this was just a gut feeling about whether this person looked like they were really in pain.

Gwen: So let’s switch to the other half of why these outcomes can differ. At least the other half of what we’re talking about this episode. There is certainly plenty more to come. Because even if you’re perfectly credible, you still have to meet your burden of proof.

Marc: Who bears the burden matters in every legal proceeding. But the default rule is usually straightforward. The party bringing the action has to prove their case.

Gwen: Right. And that’s pretty much what the APA says.

Marc: 5 U.S.C. 556(d): Except as otherwise provided by statute, the proponent of a rule or order has the burden of proof.

Marc: Proponent bears the burden. It sounds straightforward enough.

Gwen: It can be, but you have to ask, who’s the proponent? Because that can change depending on what kind of case you’re in. In an enforcement case, OSHA is citing a company for a safety violation. In that case, the agency is the proponent. The agency is seeking an order imposing the penalty, so the agency bears the burden of proving the violation.

Marc: And in a benefits case?

Gwen: In a benefits case, you are the proponent. You are seeking the order granting benefits. So you bear the burden. Same APA rule, same statutory language, but completely different practical consequences depending on which side of the table you’re sitting on.

Marc: So the burden follows whoever is asking the government to take action.

Gwen: Right. And in Social Security, Congress spelled out exactly what that burden looks like.

Marc: 42 U.S.C. 423(d)(5): An individual shall not be considered to be under a disability unless he furnishes such medical and other evidence of the existence thereof as the commissioner of Social Security may require.

Marc: Unless he furnishes such evidence. So not only do I bear the burden, the agency gets to define what evidence I need to bring.

Gwen: Yeah, that’s the practical bite. You’re the proponent, so you need to prove your case. And Social Security gets to decide what that proof looks like.

Marc: But we’ve already talked about how Social Security is an inquisitorial system. The ALJ has an affirmative duty to develop the record, to get the evidence needed for a fair decision. We spent an entire episode on that. But you’re also telling me that the claimant bears the burden of proof. The evidence is 50-50. The claimant loses, even though the ALJ was supposed to be helping develop the evidence.

Gwen: This is one of the deepest structural tensions in Social Security adjudication. Yeah, in civil law systems, the inquisitorial model doesn’t work that way. When the judge is responsible for developing the evidence, the concept of burden on the claimant operates differently. The tribunal’s obligation to find the truth means the system doesn’t just shrug when the evidence is in equipoise and say, you lose.

Marc: Right. But our system borrowed some inquisitorial features without adopting the inquisitorial approach to burden. We gave the ALJ the duty to develop the record, but we kept the rule that the proponent bears the burden of persuasion. So if the ALJ’s investigation produces a balanced record, some evidence of disability, some evidence of capability, the claimant loses.

Gwen: The system helps you build your case, but not necessarily enough. And if it doesn’t, that’s your problem.

Marc: It really feels like the worst of both worlds. You get an investigator who might develop the evidence against you, but you still bear the burden as if you were in an adversarial system with your lawyer building your own case.

Gwen: Yes, this is a hybrid problem we were talking about in the inquisitorial model episode. The system inherited the risks of both parents, and the practical consequence is exactly what you’d predict. The burden placement determines the outcome in every close case. If the evidence is 50-50 in a Social Security case, you lose. You’re the proponent and you didn’t carry your burden. If it’s 50-50 in an OSHA enforcement case, the employer wins. The agency was the proponent and it didn’t carry its burden.

Marc: So same evidence, different outcomes, entirely because of who’s the proponent.

Gwen: Yes. So the burden tells you who has to prove something. The standard of proof shows how convincing that evidence has to be. And administrative law has its own landscape here.

Marc: So preponderance of the evidence is the common one.

Gwen: Right. It’s most common. And so there we’re thinking more likely than not, like 51 percent. If the evidence tilts slightly in your favor, you’ve met the standard. And a lot of administrative cases use this. Social Security does. OSHA. This is what we use in standard employment disputes.

Marc: So then what’s the next level?

Gwen: All right, then we move to clear and convincing evidence. More than a preponderance, the evidence must be highly probable. Some immigration proceedings use this. Here’s the contrast.

Marc: The service has the burden of establishing by clear and convincing evidence that in the case of an alien who has been admitted to the United States, the alien is deportable.

Marc: All right, so compare this to OSHA, which just has to prove its case by a preponderance. Immigration uses clear and convincing. OSHA uses preponderance.

Gwen: So the standard theoretically tracks the severity of the consequences. Deportation here could mean returning to persecution and losing everything you’ve built. Congress decided that requires the government to be more certain. OSHA penalties are serious, but less catastrophic. Preponderance is sufficient.

Marc: And beyond a reasonable doubt?

Gwen: This is criminal cases. Agencies can impose pretty serious penalties, fines, license revocations, even as we’ve seen deporting you, but they don’t send you to prison, theoretically.

Marc: Okay, here’s what I keep thinking about. In a regular trial, a jury applies the standard of proof to evidence they’re hearing for the first time. They have fresh eyes. But in the system we’ve been describing all season, the ALJ who applies the standard of proof is the same person who developed the evidence, questioned the witnesses, and formed impressions throughout the investigation. Applying a preponderance when you’ve spent months building the record is a very different cognitive exercise than applying it when you’re a juror hearing a cold case.

Gwen: Right. And that is a problem. This circles back to the confirmation bias and the ownership effects that we were talking about in that inquisitorial model episode. And the ALJ who ordered the consultative exam and reviewed the file for months isn’t applying the standard of proof from a blank slate. There’s no question there. They’re applying it to evidence that they have a psychological investment in. The standard is supposed to be a neutral measuring stick, but the person holding the stick isn’t neutral the way a juror would be.

Marc: And that’s still constitutionally permissible under Withrow.

Gwen: Yes, it’s permissible. It might not be perfect, but it is permissible.

Gwen: So now let’s put these together, the reason we’re talking about both of them in this episode, because they don’t operate in isolation. They can interact to determine outcomes. And we’ve already seen the credibility piece. So both Sarah and Michael from the opening, same herniated disc, same imaging, same burden of proof. The medical condition was identical and the credibility analysis wasn’t. And that’s what produced different outcomes.

Marc: Right. So that’s credibility doing its work. But now we can move to a system where testimony does different work.

Gwen: Asylum.

Marc: Okay. Different system and a different burden.

Gwen: Right. In Social Security, as we’ve said, the person needs to prove they’re disabled by a preponderance, that it’s just more likely than not. And even when credibility is shaky there, there’s usually some sort of paper trail to fall back on. We probably have medical records. We might have imaging. We can have treatment notes, consultative exams. The objective evidence can carry a case even when the testimony is a little wobbly.

Marc: OK, then what’s the standard in an asylum case?

Gwen: Here, the applicant has to establish a well-founded fear of persecution on account of a protected ground, race, religion, nationality, political opinion or membership in a particular social group. And the Supreme Court said in Cardoso-Fonseca that this can be satisfied by showing as little as a 10 percent chance of persecution.

Marc: So the substantive standard is actually more generous than a preponderance. That sounds like it should be easier to win.

Gwen: It does on paper. But here’s what can make asylum cases so different. In Social Security, Sarah had medical records, imaging, treatment history, potentially hundreds of pages of objective evidence. But in asylum, the applicant often has none of that. They might not have a police report from the country they fled. They might not have hospital records documenting an assault or government documents that prove the political activity that made them a target. And there can certainly be cases where the primary, if not the only evidence, is just their testimony.

Marc: So credibility isn’t just important. It’s often the entire case.

Gwen: Yes, it can be. So let’s take someone with Michael’s credibility profile. He’s the one who received Social Security Disability. He is consistent, specific, corroborated with documentation. So in an immigration case, that might mean that he’s describing the persecution clearly, that the dates line up, that the account is internally consistent, and the immigration judge could credit his testimony. So a well-founded fear based on a 10% threshold supported by credible testimony, he can clear the bar.

Marc: And someone with Sarah’s credibility profile?

Gwen: So she was the one who lost. Same claim, but she might have told the asylum officer that the attack happened in March and then told the judge it happened in April. Her written declaration could describe three attackers, and at the hearing she might say two. She might have difficulty making eye contact, and her affect could be flat when she’s describing events that the judge thinks should be emotionally devastating. The judge could read these inconsistencies, evasiveness, and lack of emotion as credibility problems.

Marc: But flat affect and fragmented memory could also be symptoms of the trauma she’s describing.

Gwen: Yes, exactly. Emotional numbing, inconsistent recall, difficulty engaging with authority figures. These are clinical features of PTSD. But the immigration judge sees credibility red flags. And once you’re in an asylum case, if the judge discounts the testimony, there might not be anything left. This is if we don’t have medical records from the home country or police reports or imaging.

Marc: So this generous standard doesn’t necessarily help if the only evidence that could meet it is testimony that the judge has just discounted.

Gwen: OK, so a lower standard of proof, but higher dependence on credibility.

Marc: Yes, that can be the interaction. In Social Security, the standard is higher, but the evidence base is broader. Credibility problems could be survivable because the paper record can also do independent work. In asylum, the standard might be more generous, but the evidence base can be narrower. Credibility problems can be fatal if there’s nothing else that the judge can rely on.

Gwen: Same credibility problems, completely different consequences, depending on which system you’re in.

Marc: Yes. Credibility determines how much of your evidence the judge credits. The burden determines how much credited evidence you need. And the available evidence base determines whether anything can compensate if credibility falters. All three interact, and the interaction looks completely different depending on which agency’s hearing room you’re sitting in.

Gwen: OK, so here’s what stays with me. The system asks you to carry the burden of proving your case in a proceeding where the judge is simultaneously helping you and evaluating you. And credibility is how the judge resolves that tension. It’s the mechanism that decides whether the investigator-judge’s dual role works for you or against you.

Marc: Yeah. So if the judge finds you credible, the inquisitorial features can help you. The judge is developing evidence to support your claim. If the judge doesn’t find you credible, those same features can hurt you. The judge can develop evidence that undermines your claim, and you still bear the burden of persuasion.

Gwen: And the credibility finding is based on the ALJ sitting in the room with you, watching you testify, hearing your voice, reading your body language. That’s why it matters that this is a live hearing and not just a paper review.

Marc: Right. It matters that it’s a live hearing. It matters which ALJ is the one conducting the hearing. And that raises a question that we spent two episodes circling without really asking. We’ve been looking at two specific factors, credibility and burden of proof, that shape what happens in the hearing room, but those are two inputs into a larger question.

Gwen: Which is what produces hearing outcomes in general.

Marc: Right. Because if credibility is making this much difference, and the burden is interacting with credibility in this way, and different ALJs will assess credibility differently and apply the burden differently, then how much does the outcome depend on which specific ALJ ends up in the room with you? Same claim, same record, different judge. How different is the likely outcome?

Gwen: Right. That’s the question we look at next. We’ve been looking at the inside of the hearing. Next episode, we step back and ask what hearing-level outcomes actually look like across the system, what determines them, how much variation there is, and what that tells us about what administrative adjudication actually produces. Credibility and burden of proof are two pieces of a larger picture.

Marc: And I’m guessing the larger picture isn’t reassuring.

Gwen: Yeah, the larger picture is that credibility and burden aren’t the only things that matter, and maybe not even the things that matter most. What matters most may be structural features of the system that operate before either credibility or burden comes into play. That’s the next episode.

Marc: So that does it for today’s episode on Administrative Remedies. Thank you for joining us today. Please, if you enjoy this podcast and enjoy this episode, give us a like on Spotify, iTunes, or whatever platform you’re listening on. And be sure to tune in next time where we’ll continue to dive into the contours of administrative law, because remember, you can’t fix what you don’t understand.

Related Guides

The Lifecycle of an Administrative Case — This episode picks up where Lifecycle left off: that one mapped how the record gets built, this one explains how the judge evaluates it once it's in front of them.
The Adjudication Spectrum — The structural tension at the core of this episode — inquisitorial duties paired with adversarial burdens — only makes sense once you understand where Social Security falls on the adjudication spectrum.
Who Decides: ALJs vs. Non-ALJs — Credibility findings are nearly unreviewable on appeal precisely because of who the ALJ is and how they sit in the proceeding; this guide explains the role.

Why Two Identical Disability Cases Can Have Opposite Outcomes