#how to identify confounding variables

[religion-is-a-mental-illness]

By: Christina Buttons

Published: May 31, 2024

The prominent science journal Nature has launched a new opinion article series on sex and gender. One paper in this series explores research attempting to search for a biological basis for trans-identity, arguing that such research could “pathologize” and “harm” the trans community. The authors discourage “investigations into the underlying bases of transgender identity” and propose various steps for researchers to incorporate transgender activism into their work to influence research outcomes — signaling the end of Nature’s commitment to pursuing scientific truth over ideology.

The article starts by reviewing neuroscientific studies aimed at finding the cause of trans-identity in the brain, identifying 83 papers from 1991-2024. It highlights the transgender brain-sex hypothesis, which suggests that trans-identified people have brain regions resembling those of the opposite sex. However, it neglects to mention that this hypothesis falls apart because the studies did not control for confounding variables such as sexual orientation.

The article does acknowledge that “the results of these analyses have been inconsistent.” Yet, when the media covers these studies, the public is often informed by headlines such as “transgender people are born that way,” “science proves trans people aren’t making it up,” and “attacks on trans people are also attacks on science itself.” You can read a simplified explainer I wrote debunking the brain-sex studies here.

The authors move on to the more plausible “own-body perception” theory, which proposes that reduced structural and functional connectivity between certain brain networks is responsible for gender dysphoria. However, these studies do not show a causal link, only an association. Abnormalities in body perception networks in the brain are also associated with many other conditions, including body dysmorphic disorder, anorexia, body integrity identity disorder, schizophrenia, and autism.

After reviewing the neuroscientific studies, the article’s language shifts into typical activist rhetoric, claiming that research into transgender identity can be “harmful.” The authors argue that if brain scans or some other objective test could assess whether someone is experiencing gender dysphoria, it could be used to prevent people from accessing cross-sex hormones and surgeries if they are not deemed “eligible.”

"A second possibility is that neuroscientific findings related to transgender identity will fuel transphobic narratives," the authors write, citing a “feminist perspective” social science journal article on "Transprejudice."

For example, they state, "Some people argue that allowing transgender women to access infrastructure, such as public toilets or women’s prisons, threatens the safety of 'real women'." It is odd and audacious for a serious science publication to use "real women" in quotations. Moreover, their source for this claim is an article about Kathleen Stock, who does not argue that transgender women threaten the safety of biological women. In fact, she explicitly states the opposite: "I am definitely not saying that trans women are particularly dangerous – they are definitely not."

The authors also take a dig at sexologist Ray Blanchard, claiming that autogynephilia “hasn’t held up to scientific scrutiny,” citing a "feminist analysis" paper by a trans activist. Apparently, they haven't spent any time on trans Reddit, where they would encounter a vast discourse on "gender euphoria boners."

The authors end by setting “four actions” for researchers studying transgender people to prevent further “harm” from being done. They suggest researchers set up an advisory board and multidisciplinary teams consisting of transgender people to consult on their study designs and “prevent the outcomes of neuroscientific and other studies from being described and published in an overly deterministic and simplistic way.” They also dictate what should and should not be studied, suggesting researchers "prioritize research that is likely to improve people’s lives" rather than searching for the cause of trans-identity.

The final suggestion is to “rethink how ethical approval is obtained,” which relates to an example they provided of a 2021 UCLA study that was suspended after significant backlash from transgender activists. The study aimed to examine the brains of trans-identified individuals by showing them images of themselves wearing tight clothes, intending to trigger gender dysphoria. Although the study obtained ethical approval from their research institute and the transgender participants provided informed consent, it seems they weren't the right transgender people to ask permission from. Their suggestion implies that researchers must obtain approval for their studies from transgender activists.

The authors seem aware of the implications of their recommendations, as they conclude their article by admitting their approach would limit scientific inquiry:

“Our aim is not to halt scientific enquiry. But when it comes to transgender identity, knowledge cannot be pursued in isolation from the many societal factors that shape how that knowledge is received and acted on.”

This statement translates to prioritizing activism over truth-seeking when the findings might be inconvenient or misaligned with political narratives and activist goals. Such a stance compromises the integrity and credibility of science, reducing it to a tool for activism rather than a means of uncovering and understanding reality.

It is disheartening to watch one of the world’s most prestigious scientific journals compromise their credibility by continuing to prioritize ideology over truth.

Besides, the authors' concerns about discovering a biological cause for trans-identity are misplaced. While there are biological traits associated with being transgender, such as same-sex attraction and gender nonconformity, “transgender” itself does not appear to be an inherent condition one can be born with. The concept of "transgender," as understood in Western cultures, is a cultural construct that doesn't have a direct equivalent in many non-Western societies.

Research into a cause for gender dysphoria would be difficult because the transgender population has become so heterogeneous. Even if one were predisposed to a psychiatric condition like gender dysphoria, predispositions are not predeterminations of a transgender outcome. The notion of transgender identities being fixed at birth is further contested by the increasing number of detransitioners and extensive research on desistance among children, suggesting that such identities can often be temporary coping mechanisms for young people in distress.

==

We're just supposed to accept that hacking off body parts and giving life-altering drugs and hormones is a completely normal part of life. And that wondering where this is all coming from, what's underlying it is the problematic part.

At its core, the point of this ideology is to pathologize the completely normal and normalize the pathological.

Carl Sagan warned us about this:

"The truth may be puzzling. It may take some work to grapple with. It may be counterintuitive. It may contradict deeply held prejudices. It may not be consonant with what we desperately want to be true. But our preferences do not determine what's true." ― Carl Sagan

Reality is not obliged to conform to people's wishes or preferences, and we are not obligated to lie or consign ourselves to ignorance in order to placate those wishes and preferences. We don't allow "if you find out what's true, it'll hurt our feelings" - i.e. blasphemy - for the religious. Why are we allowing genderist fanatics to get away with it, when it's still just an accusation of blasphemy?

When people say, "you're not allowed to go looking over here, it's a moral failing to do so," the correct response is to go, "now I want to go look over there even more."

-

"Sex is real... But the belief that we have a moral duty to accept reality just because it is real is, I think, a fine definition of nihilism." ― Andrea Long Chu, gender cultist and lunatic

“The facts may tell you one thing. But, God is not limited by the facts. Choose faith in spite of the facts.” ― Joel Osteen, religious nutcase and lunatic

[sylph-0f-space]

Insider facts to Sports Betting Exposed

For the vast majority, sports betting is extremely confounded. Be that as it may, truly, you needn't bother with a degree to comprehend how it functions, you should simply watch a ton of ESPN and assume responsibility for your numbers.

Some express that there are fundamental variables to sports betting, however to tell you sincerely, there are no enormous thoughts behind betting. Maybe, there might be mysteries to winning thousands in sports betting, yet these insider facts can be identified by anybody, even not-really sports-disapproved of people.

Here is a rundown of a portion of the privileged insights of sports betting, exposed:

1. The primary mystery to sports betting is to remain all alone. Try not to put down a bet because a group has many fans or highlights a headliner. Rather, make a bet because your nature tells you to. Depending on your companions' opinions or previous encounters won't take you some place. Have little to no faith in the oddmakers or the bookies either for they are just thinking about their own pockets.

2. Try not to bet in a solitary group until the end of a season. This mystery to sports betting won't just decrease your possibilities on acquiring money, however it will likewise restrict your decisions to the better coordinates with better chances from here on out. All the more in this way, a group perhaps a leader toward the beginning of a season yet you perhaps ignoring likely groups than could get through the playoffs. Something else, you can not dispose of the way that players are as yet human and even headliners can get harmed mid-season.

3. Appropriate your money among different matches. This mystery to sports betting is finished by numerous master bettors. Not exclusively will this increment you chances of winning, yet it will likewise safeguard your money. You would have no desire to lose in a simply a one-time-big-time bargain, isn't that right?

4. One more confidential to sports betting is to continuously remain alert. Oddmakers tend to change money lines and point spread in only a flicker of an eye and as frequently as could really be expected, particularly if it is one of those most-anticipated matches like a title battle in boxing or a Game 7 finals. Mediocre is truly invaluable during these times because you can win the two different ways. Know about new chances and make the most of those turns of events.

5. Pick a decent internet betting website. This mystery to sports betting is being disregarded by a lot of people. A few sites are not popular for not a really obvious explanation. It is because bettors trust them and they offer incredible client support. A decent internet betting webpage ought to make betting issue free and a delight for devotees

6. The last confidential to sports betting is to explore, endlessly research. No, it doesn't mean going to the library, it implies gathering as mush data on sports, be it on the papers, web or ESPN. What makes a difference is that you keep yourself refreshed in order to be not caught unaware when putting down bets. Oddmakers generally to this to think of precise and nearly balance chances, and it's additionally your obligation to do as such. You'll see that it will take care of over the long haul - when your financial balance is getting increasingly thick continuously.

[rad4learning]

Some context on that previous post, originally I had this in there:

"It is common to see claims like "[medical intervention] is good/bad for trans-identified people."

I don't know where the radblr post is :/ but I've seen discussion of this article https://doi.org/10.1093/jsxmed/qdaf026

The researchers do not make a causal claim about the meaning of their results, in the abstract conclusion we have "Gender-affirming surgery, while beneficial in affirming gender identity, is associated with increased risk of mental health issues, underscoring the need for ongoing, gender-sensitive mental health support for transgender individuals’ post-surgery."

Some on radblr, however, have. In this case, there would seem to be a pretty obvious confounder - people might be more likely to sign up for the surgery if they are experiencing worsening mental health. Maybe the worse outcomes overtime are because these are people who are already crashing. I don't have access to the full paper at the moment so apologies if that's comprehensively addressed in there.

I'm not claiming that if the results were the opposite there wouldn't be shouting from the rooftops that "trans surgery improves mental health"* or that this result is meaningless. I'm just saying, caution is warranted with causal claims and frankly we don't need to make those claims to think that there should be caution about medically indicating cosmetic surgery for people who tend to experience worsening mental health afterwards.

The abstract lists that matching was done on propensity scores using age, race and ethnicity. Propensity score matching, despite being everywhere, is not a great form of matching but matching is a good practice. It tries to make the two groups more comparable to each other by picking some variables that might be different between the groups & matter to the outcome of interest. Then you use those to make your groups that you analyse more similar to each other (simplifying a little here) even though the original groups are different. Researchers are often limited by what information is available. Clearly, age, race & ethnicity might not be the only important differences between US patients with gender dysphoria diagnosis who undergo surgery or not. I'm going to skip past the claim about beneficial impact from affirming gender identity because I'm just looking at the abstract and don't know what is underpinning that claim. "

but I have now found a copy of the paper and while it isn't a causal study, in the absence of reading a study that does look for a causal effect... it's pretty concerning even though the authors don't intend it to be. I wouldn't say that them selecting only patients without diagnosed mental health disorder rules out the possibility of selection bias being responsible for the different results but I do think it reduces how much it is plausible that that accounts for a lot of the gap given that the measure of worse mental health was also based on diagnosed mental health disorder. So I won't claim that the study proves that there is a causal effect but my impression so far is that I think it is pretty suggestive that that might exist.

[literaturereviewhelp]

Describe the influence "levels of evidence" have on practice changes. Identify the most reliable level of evidence and provide an example of the type of practice change that could result from this level of evidence. Initial discussion question posts should be a minimum of 200 words and include at least two references cited using APA format.   OR   The paper will consist of 3-5 pages of content, a cover page and a reference page.  The total page count with the cover page and the reference page should be 5-7 pages. Your paper should include an introduction and conclusion that summarize the contents of the entire paper. Your paper should be written in proper APA format.  This link will take you to the section of the APUS library that can assist you with formatting: apus.campusguides.com/content.php Paper topic:       How Have We Evolved in the Management Field? Write: Based on your readings and research, develop a timeline of five key management theory/principle milestones over time.  Please address the following questions for each of the milestones you elect to include on your timeline: - Why was this milestone significant for the period of time it was created?  What was going on in the world of work that allowed the environment to know the time was right for this particular milestone? - Which theorist “fathered” the principle?  What were the signs of the time which led the theorist to develop the management concept? - What are the highlights and limitations of the theory/principle? References: At least two references are required for this assignment.  You may use your textbook as a reference in addition to the two external references.   OR Statistical significance refers to the likelihood that the results of a study are not due to chance, while clinical significance refers to the practical importance of the results in terms of their impact on patient care. In other words, statistical significance is a measure of the strength of the evidence, while clinical significance is a measure of the relevance of the evidence to real-world situations. Using a quantitative research article from one of the previous topics, analyze the p-value. What is it? Is it statistically significant? If your p-value is not statistically significant, what is the clinical significance? Generalizability of research depends on a variety of factors. List three factors of generalizability, and discuss whether this research article is generalizable to the nursing problem you are researching. Initial discussion question posts should be a minimum of 200 words and include at least two references cited using APA format. For previous articles, check the attachments.     The Role of Confounding Variables in Causal Inference In research, independent variables are the ones that are controlled or varied in order to establish their correlation to the dependent variable. The dependent variable is the variable that is being observed in order to determine the effects of the independent variables. On the other hand, there are extraneous variables that refer to any variable that is not of interest to the study but which has the capability of affecting the dependent variable. There are two common strategies used to address extraneous variables: randomization and matching. Randomization helps in the allocation of the participants to groups in a random manner to eliminate bias. Matching: the subjects are matched in such a way that other extraneous variables are grouped together to avoid confounding. Another type is statistical control, in which the researcher employs statistical means for controlling for extraneous variables, such as through regression analyses (Byrnes and Dee, 2025). For instance, in the paper by Egami and Tchetgen (2024), the authors employed negative control variables (auxiliary variables) to control unmeasured network confounding in causal peer effect estimation. By including outcome and exposure variables with no relation to the treatment of interest, they couldn’t be influenced by confounding and determine accurate causal peer effects with dependence in the network (Egami and Tchetgen, 2024).             Reference Egami, N., & Tchetgen Tchetgen, E. J. (2024). Identification and estimation of causal peer effects using double negative controls for unmeasured network confounding. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 86(2), 487-511. https://doi.org/10.48550/arXiv.2109.01933 Byrnes, J. E., & Dee, L. E. (2025). Causal inference with observational data and unobserved confounding variables.  Ecology Letters,  28(1), e70023. https://onlinelibrary.wiley.com/doi/abs/10.1111/ele.70023   Read the full article

[theevolvingimportanceofmedic]

Are Real-World Studies Reliable? Addressing Bias & Data Quality Issues

In an era where healthcare decisions are increasingly driven by data, real-world evidence (RWE) has become a crucial tool for assessing treatment effectiveness beyond controlled medical trials. Real-world data (RWD) provides insights into how medical interventions perform across diverse patient populations in routine practice. However, concerns regarding bias, data integrity, and regulatory compliance raise an important question: How reliable are real-world studies?

The Growing Importance of RWE

Unlike traditional clinical trials, which follow strict protocols and eligibility criteria, real-world studies rely on data from electronic health records (EHRs), insurance claims, patient registries, and even wearable devices. This shift allows researchers, policymakers, and healthcare professionals to evaluate the long-term safety, cost-effectiveness, and impact of treatments in real-world settings.

Medical affairs teams use RWE to support health economics research, inform market access strategies, and guide regulatory decision-making. However, ensuring the credibility of findings requires a proactive approach to addressing biases and enhancing data quality.

Common Biases in Real-world Studies

Real-world studies are vulnerable to multiple forms of bias, which can compromise their reliability:

Selection bias: Since real-world studies do not employ randomized patient selection, certain demographic or clinical groups may be overrepresented or underrepresented, leading to skewed results.

Confounding variables: Unlike randomized controlled trials (RCTs), real-world studies often lack mechanisms to isolate variables, making it difficult to establish causality.

Reporting bias: Incomplete or inconsistent data entry in electronic health records and insurance claims databases can introduce errors that affect study conclusions.

Publication bias: Studies with favorable outcomes are more likely to be published, creating an incomplete picture of a treatment’s true effectiveness.

Mitigating Bias in RWE

Several methodologies can help mitigate bias in RWE studies:

Propensity score matching (PSM): This statistical technique matches patients with similar baseline characteristics to reduce confounding.

Inverse probability weighting (IPW): A weighting method that adjusts for imbalances in patient characteristics, improving comparability.

Sensitivity analyses: Conducting multiple analyses with different assumptions helps assess the robustness of findings.

Use of linked datasets: Combining multiple data sources (e.g., EHRs, registries, and claims data) can improve data completeness and reduce missingness-related biases1.

Ensuring Data Quality in Real-world Studies

Improving the reliability of RWE requires stringent methodologies and advanced analytical tools. Strategies to enhance data quality include:

Systematic literature reviews: Conducting thorough literature reviews ensures that RWE studies incorporate all relevant data, reducing the risk of biased conclusions2.

Artificial intelligence in healthcare: AI-driven analytics can identify patterns, clean datasets, and account for missing variables, leading to more reliable insights3.

Standardized data collection: Implementing structured reporting systems across healthcare institutions ensures greater consistency and completeness in real-world data4.

Regulatory compliance: Adhering to guidelines set by regulatory bodies such as the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA) ensures that real-world studies meet rigorous scientific and ethical standards5.

The Role of Regulatory Compliance in RWE Reliability

To incorporate RWE into clinical decision-making, regulatory bodies have introduced stringent data governance frameworks. Ensuring compliance with Good Clinical Practice (GCP) and other regulations mitigates the risks associated with incomplete or biased data.

The FDA’s Real-World Evidence Frameworkestablishes standards for assessing RWD quality, study design, and applicability in regulatory submissions6.

The EMA emphasizes transparency and reproducibility in RWE submissions, ensuring that studies meet the highest scientific standards7.

For example, the FDA approved Palbociclib (Ibrance) for male breast cancer based on RWE from claims and EHR data rather than traditional clinical trials8. This case highlights how high-quality RWE can inform regulatory decisions when RCTs are impractical.

Future Outlook: Combining RWE with Clinical Trials

While RCTs remain the gold standard for evaluating treatment efficacy, RWE plays a complementary role by providing insights into long-term safety, patient adherence, and economic impact. Integrating real-world data with traditional research methodologies can create a more comprehensive understanding of healthcare interventions.

Advancements in AI-driven analytics, real-time data integration, and digital health monitoring are improving the accuracy of RWE studies. Organizations are increasingly leveraging these technologies to refine data accuracy and eliminate bias9. By embracing the best practices in systematic literature review, regulatory compliance, and data validation, real-world studies can offer valuable insights that drive evidence-based healthcare decisions.

The Path Forward

RWE is a powerful tool in modern healthcare, but its reliability depends on addressing biases and ensuring data integrity. Implementing standardized methodologies, leveraging artificial intelligence, and adhering to regulatory standards can help unlock the full potential of real-world studies and effectively disseminate findings across the healthcare ecosystem.

References

Schneeweiss S. Learning from big health care data. N Engl J Med. 2014;370(23):2161-3.

Wang SV, Pinheiro S, Hua W, et al. STaRT-RWE: structured template for planning and reporting on the implementation of real-world evidence studies. BMJ 2021;372:m4856.

Rajkomar A, Dean J, Kohane I. Machine learning in medicine. N Engl J Med. 2019;380(14):1347-58.

FDA. Real-world evidence: what is it and what can it tell us? [Internet]. 2023 [cited Feb 27, 2025]. Available from: Real-World Evidence

European Medicines Agency. Real-world evidence in regulatory decision-making [Internet]. 2022 [cited Feb 27, 2025]. Available from: https://www.ema.europa.eu/en/human-regulatory/post-authorisation/real-world-evidence

US FDA. Framework for FDA’s real-world evidence program [Internet]. 2018 [cited Feb 27, 2025]. Available from: https://www.fda.gov/media/120060/download

European Medicines Agency. Guideline on good pharmacovigilance practices (GVP) [Internet]. 2021 [cited Feb 27, 2025]. Available from: https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-good-pharmacovigilance-practices_en.pdf

US FDA. FDA approves Ibrance for male breast cancer based on real-world evidence [Internet]. 2019 [cited Feb 27, 2025]. Available from: https://www.fda.gov/news-events/press-announcements/fda-approves-ibrance-male-breast-cancer-based-real-world-evidence

Corrigan-Curay J, Sacks L, Woodcock J. Real-world evidence and regulatory decision making: where are we now? Clin Pharmacol Ther. 2018;104(5):822-9.

Contact Us – info@turacoz.com

Visit Us at – Turacoz Healthcare Solutions: Medical Communications Company

[stevensaus]

When You Don’t Ask The Right Questions, Research Edition

Science reporting (and research) had its flaws prior to the current ideological imposition being put on it; these four articles not only illustrate the problems, but give you some good ideas of what to watch out for.

Cause And Effect

When there’s breathless coverage about the dangers of something — particularly something that isn’t part of “polite” society — then you have to be particularly concerned about any implications about causation. The most recent is research out of Canada, and the reporting sure makes it seem like ingesting THC could lead to schizophrenia… but they quite likely have the causality of the thing backward.

The most likely hypothesis? More people who were suffering from schizophrenia that was not yet addressed by the healthcare system self medicated in an attempt to control their own mental health. Self-medication — whether you’re using pot, booze, carbs, or something else — is not a good thing. But there is a huge difference between that and implying that THC use leads to an increased risk of developing schizophrenia. The two may be correlated, but that is very different than causation.

This is not just a bit of pedantry; the implications have real-world effects. To me, this study seems to indicate that the best intervention would outreach and increased availability of psychiatric services to those self-medicating.

Morning Happiness

Sample bias is still a real thing, as seen in this study that claims that “everyone is happiest in the morning.” A lot of psychological and sociological research already suffers from sample bias — including my own — because it uses a sample of convenience, often college students.

That problem is intensified when you realize that most studies of the general population do not take neurodivergence into account.

Sometimes those differences are minor, but significant.

In this case, getting up in the morning.

It’s fairly common for those with ADHD to be night owls. But with a prevalence estimated anywhere from nearly 15% to as low as 4% (and varying by gender), if you fail to separate out this confounding variable, your study results are trash and not generalizable.

Because the researchers did not distinguish between neurotypes, this study claiming that “everyone” is happiest in the mornings is equivalent to a 1950’s researcher saying that “children are happiest with right-handed writing materials.” Sure, that’s true for right-handed people, but for the 10%-12% of lefties, that’s absolutely wrong.

Burying The Most Important Part

Likewise, this study which claims that oral contraceptives protect against ovarian cancer does not separate out between the types of oral hormonal birth control. While I’ve never taken any, I know lots of women who have shared with me that different brands and formulations have very different effects on their bodies. Does that affect the results of this study? We cannot know because the researchers didn’t bother to check.

More concerningly to me is how the reporting buried the most stunning and useful bit of incidental findings I’ve seen in a long time: “We also identified blood biomarkers associated with ovarian cancer years before diagnoses, warranting further investigation.”

THEY IDENTIFIED BIOMARKERS THAT COULD LEAD TO A BLOOD TEST FOR OVARIAN CANCER. That’s huge! I’d say they should get a NIH grant, except… well.

That SOUNDS Like A Lot…

Speaking of the current political climate, that brings us finally to context. With the anti-trans bigots pumping their hatred into the public discourse, you’ve probably heard mention of transgender people who have “decision regret” after receiving gender-affirming care. This does happen, though it’s pretty rare. Measuring that rate is difficult, as different researchers used different metrics for what “counted” as “regret” (and after what kinds of procedures), leading to estimates ranging somewhere between 1% and 8%.

This figure has been weaponized by politicians and political opportunists (looking at you, Chloe Cole) as a reason to halt gender-affirming care.

Before you wring your hands too much about it or think they’re making a good point, it’s worth noting that studies of women who were treated for early breast cancer found rates of decision regret about their treatment as high as 69%.

The researchers — as reported by Oncology Nurse Advisor — found that “decision regret was influenced by multiple factors, including patient demographics, decision-making processes, and mental health, necessitating targeted interventions to mitigate its impact.”

If you’re wondering why that statistic isn’t getting the same kind of press, or why the suggested solution isn’t more and better support for transgender people like it is for women being treated for breast cancer, or why a political party isn’t making it a giant talking point,the answer is simple.

It’s bigotry, my friends.

What’s The Point?

The point?

For those doing the research: Do better with your research questions and conclusions. Communicate your findings clearly — including the nuance — to reporters, don’t trust them to understand.

For the rest of us: Think about what you read. Dig a little further, even if it seems to support your point of view. While reporters will misunderstand and (unintentionally) misrepresent findings, while politicians will deliberately lie, there is truth out there.

But it’s rarely as simple as a soundbite. Originally at ideatrash

[sunaleisocial]

Revealing causal links in complex systems

New Post has been published on https://sunalei.org/news/revealing-causal-links-in-complex-systems/

Revealing causal links in complex systems

Getting to the heart of causality is central to understanding the world around us. What causes one variable — be it a biological species, a voting region, a company stock, or a local climate — to shift from one state to another can inform how we might shape that variable in the future.

But tracing an effect to its root cause can quickly become intractable in real-world systems, where many variables can converge, confound, and cloud over any causal links.

Now, a team of MIT engineers hopes to provide some clarity in the pursuit of causality. They developed a method that can be applied to a wide range of situations to identify those variables that likely influence other variables in a complex system.

The method, in the form of an algorithm, takes in data that have been collected over time, such as the changing populations of different species in a marine environment. From those data, the method measures the interactions between every variable in a system and estimates the degree to which a change in one variable (say, the number of sardines in a region over time) can predict the state of another (such as the population of anchovy in the same region).

The engineers then generate a “causality map” that links variables that likely have some sort of cause-and-effect relationship. The algorithm determines the specific nature of that relationship, such as whether two variables are synergistic — meaning one variable only influences another if it is paired with a second variable — or redundant, such that a change in one variable can have exactly the same, and therefore redundant, effect as another variable.

The new algorithm can also make an estimate of “causal leakage,” or the degree to which a system’s behavior cannot be explained through the variables that are available; some unknown influence must be at play, and therefore, more variables must be considered.

“The significance of our method lies in its versatility across disciplines,” says Álvaro Martínez-Sánchez, a graduate student in MIT’s Department of Aeronautics and Astronautics (AeroAstro). “It can be applied to better understand the evolution of species in an ecosystem, the communication of neurons in the brain, and the interplay of climatological variables between regions, to name a few examples.”

For their part, the engineers plan to use the algorithm to help solve problems in aerospace, such as identifying features in aircraft design that can reduce a plane’s fuel consumption.

“We hope by embedding causality into models, it will help us better understand the relationship between design variables of an aircraft and how it relates to efficiency,” says Adrián Lozano-Durán, an associate professor in AeroAstro.

The engineers, along with MIT postdoc Gonzalo Arranz, have published their results in a study appearing today in Nature Communications.

Seeing connections

In recent years, a number of computational methods have been developed to take in data about complex systems and identify causal links between variables in the system, based on certain mathematical descriptions that should represent causality.

“Different methods use different mathematical definitions to determine causality,” Lozano-Durán notes. “There are many possible definitions that all sound ok, but they may fail under some conditions.”

In particular, he says that existing methods are not designed to tell the difference between certain types of causality. Namely, they don’t distinguish between a “unique” causality, in which one variable has a unique effect on another, apart from every other variable, from a “synergistic” or a “redundant” link. An example of a synergistic causality would be if one variable (say, the action of drug A) had no effect on another variable (a person’s blood pressure), unless the first variable was paired with a second (drug B).

An example of redundant causality would be if one variable (a student’s work habits) affect another variable (their chance of getting good grades), but that effect has the same impact as another variable (the amount of sleep the student gets).

“Other methods rely on the intensity of the variables to measure causality,” adds Arranz. “Therefore, they may miss links between variables whose intensity is not strong yet they are important.”

Messaging rates

In their new approach, the engineers took a page from information theory — the science of how messages are communicated through a network, based on a theory formulated by the late MIT professor emeritus Claude Shannon. The team developed an algorithm to evaluate any complex system of variables as a messaging network.

“We treat the system as a network, and variables transfer information to each other in a way that can be measured,” Lozano-Durán explains. “If one variable is sending messages to another, that implies it must have some influence. That’s the idea of using information propagation to measure causality.”

The new algorithm evaluates multiple variables simultaneously, rather than taking on one pair of variables at a time, as other methods do. The algorithm defines information as the likelihood that a change in one variable will also see a change in another. This likelihood — and therefore, the information that is exchanged between variables — can get stronger or weaker as the algorithm evaluates more data of the system over time.

In the end, the method generates a map of causality that shows which variables in the network are strongly linked. From the rate and pattern of these links, the researchers can then distinguish which variables have a unique, synergistic, or redundant relationship. By this same approach, the algorithm can also estimate the amount of “causality leak” in the system, meaning the degree to which a system’s behavior cannot be predicted based on the information available.

“Part of our method detects if there’s something missing,” Lozano-Durán says. “We don’t know what is missing, but we know we need to include more variables to explain what is happening.”

The team applied the algorithm to a number of benchmark cases that are typically used to test causal inference. These cases range from observations of predator-prey interactions over time, to measurements of air temperature and pressure in different geographic regions, and the co-evolution of multiple species in a marine environment. The algorithm successfully identified causal links in every case, compared with most methods that can only handle some cases.   

The method, which the team coined SURD, for Synergistic-Unique-Redundant Decomposition of causality, is available online for others to test on their own systems.

“SURD has the potential to drive progress across multiple scientific and engineering fields, such as climate research, neuroscience, economics, epidemiology, social sciences, and fluid dynamics, among others areas,” Martínez-Sánchez says.

This research was supported, in part, by the National Science Foundation.

[churchofnix]

To effectively ask for evidence and double-blind, peer-reviewed studies to support scientific claims, follow these guidelines:

Be Specific: Clearly identify the claim you are questioning. Instead of general queries, pinpoint the exact aspect you want evidence for. This helps ensure you get precise and relevant information.

Request Evidence: Ask for empirical evidence that supports the claim. Use language such as, "Could you provide the empirical evidence supporting this claim?" or "What studies back up this assertion?"

Double-Blind Studies: Specify the type of study you are looking for. Double-blind studies are crucial in eliminating bias. For instance, "Are there any double-blind studies that validate this claim?" or "Please share any double-blind research conducted on this topic."

Peer Review: Emphasize the importance of peer-reviewed studies. Peer review ensures that the research has been evaluated by other experts in the field. Ask, "Can you provide peer-reviewed studies that support this claim?" or "Which peer-reviewed journals have published research on this subject?"

Quality and Source: Inquire about the quality and source of the studies. Ask questions like, "Which journals were these studies published in?" or "What is the sample size and methodology of the studies cited?"

Critical Analysis: Encourage a critical evaluation of the evidence. You might say, "How do these studies account for potential confounding variables?" or "What limitations were identified in these studies?"

Replication: Ask if the findings have been replicated. Replication is key in validating results. For example, "Have these results been replicated in other studies?" or "Is there consensus in the scientific community about these findings?"

Scientific Consensus: Check if the claim is widely accepted by the scientific community. You can ask, "What is the consensus among experts regarding this claim?" or "How does this claim align with the current scientific understanding?"

Example Request:

"Could you provide empirical evidence supporting this claim? Specifically, I am looking for double-blind, peer-reviewed studies published in reputable journals. It would be helpful to know the sample sizes, methodologies, and any potential limitations identified in these studies. Additionally, have these findings been replicated, and what is the general consensus among experts in this field?"

Importance of This Approach

Myths and pseudoscience can spread misinformation and lead to harmful consequences. By demanding rigorous evidence and well-conducted studies, we uphold scientific integrity and ensure that claims are based on reliable and validated information. This approach fosters critical thinking and helps distinguish between scientifically supported facts and unfounded assertions.

[dataanalyst75]

Peer-graded Assignment: Testing a Multiple Regression Model of Gross Domestic Product per capita

from Regression Modeling in Practice - Wesleyan University - Week 3

Starting from the hypothesis for a linear association between "Internet users (per 100 people)" – quantitative explanatory variable - and "Gross Domestic Product (income) per capita (person) in constant 2000 US$ " – quantitative response variable – hereinafter what is being found from a multiple regression analysis.

Firstly, being assessed whether the results support the hypothesis for the association between “income per person” and “Internet users”.

Secondly, an evaluation is conducted whether there is any evidence of confounding by adding to the model one explanatory variable at a time, making it easier to identify which of the variables could be confounding variables.

Furthermore, using multiple regression, other variables are being added to the model in order to evaluate multiple predictors of incomeperperson quantitative response variable. The results of the associations between all explanatory variables and income per capita are reported, where for “all explanatory variables” are intended other likely predictors, included internet users.

The following regression diagnostic plots are generated as well, making a description of the regression model in terms of the distribution of the residuals, model fit, influential observations, and outliers:

1.       q-q plot

2.       standardized residuals for all observations

3.       leverage plot

Hereinafter " Gross Domestic Product per capita in constant 2000 US$" ("Internet users (per 100 people)" respectively) is referred to as simply “incomeperperson” (“Internetuserate” respectively). 

Testing a multiple linear regression model: how to evaluate the fit of the model based on the significance of reggression coefficients, confidence intervals, and by the R square value

From the linear regression test, it arises that Internetuserate is significantly, positively associated with incomeperperson: incomeperperson increases as internetuserate increases.

In addition, after partialing out the part of the association that can be accounted for by one potential confounder at a time, it can be concluded that Internetuserate is still significantly associated with incomeperperson.

However, it is about a relationship not linear, but curvilinear. Indeed, firstly, if a straight linear regression is being drawn through the scatter plot points, most points would be really far away from the line, especially those around Internetuserate of 30 or higher: it sounds like incomeperperson appears to decrease, meaning that there is a lot of prediction error. Secondly, the R-square is about 66%, indicating that the linear association of Internetuserate is capturing only about 66% of the variability in incomeperperson.

Definitely, the best fitting line is not straight, rather it is a quadratic curve which captures the non-linear nature of the association, i.e. which picks up on the curvilinear part of the relationship. The quadratic line is capable of capturing the association at medium and higher Internetuserate. The evidence of this is given by adding a second order polynomial term to the regression model so that a significantly better fitting model is being returned. Indeed, the quadratic term is positive, and significant, indicating that the curvilinear pattern in the scatter plot is statistically significant (p <0.0001). A positive linear coefficient and a positive quadratic coefficient means that the curve is concave, such that starts high, then goes down, and then starts to go up again. In addition, the R-Square increased at roughly 77%, which means that adding the quadratic term for Internetuserate increase the amount of variability in incomeperperson. Testing for cubic, third-order polynomial, it turns to worsen the p-value as well as not improving the fit of the model.

After adjusting for potential confounding factors such as

·         alcohol consumption per adult (age 15+), litres urban population (% of total) - “alcconsumption”;

·         urban population (% of total) – “urbanrate”;

·         total employees age 15+ (% of population) – “employrate”;

·         residential electricity consumption, per person (kWh) – “relectricperperson”,

from multiple linear regression model testing, it comes up that,

·         Internetuserate is significantly, positively associated with incomeperperson. Therefore, Internetuserate is a predictor over incomperperson as the association between this explanatory variable and the response variable remains significant after controlling for all other variables in the multiple regression model.

Hereinafter the p-value, the estimated regression coefficient and its 95% confidence interval, useful to evaluate the overall fit of the predicted values (model) of the incomeperperson response variable:

o   Beta = 297.19,

o   p <0.0001,

o   95% confidence interval: from 227.29   to 367.09

·         alcconsumption is significantly, but negatively associated with the response variable (Beta = -358.131171, p = 0.0011) after partialing out the part of the association that can be accounted for all other explanatory variables in the multiple regression model.

o   Beta = -296.009399,

o   p < 0.0256,

o   confidence interval: from -555.23 to -36.79

·         urbanrate is no longer associated with incomeperperson after controlling at least for the Internetuserate, as the p value is much higher than 0.05. This is a case of a confounding, where urbanerate is no longer significant when at least Internetuserate is included in the model. The null hypothesis of no association between urbanrate and incomeperperson cannot be rejected after adjusting for Internetuserate and the other explanatory variables in the model. The confidence interval includes a value of zero in that range: this is another evidence of no association as, in a multiple regression analysis, a 95% confidence interval that includes a value of zero means that the p-value is greater than 0.05.

o   Beta = 54.8,

o   p = 0.1189,

o   confidence interval: from -14.27 to 123.87

·         employrate is also significantly and positively associated with incomeperperson, such that higher percentage of total population, that has been employed, reports a greater income per capita.

o   Beta = 139.15,

o   p = 0.0247,

o   confidence interval: from 18.03 to 260.27

·         relectricperperson is also significantly and positively associated with incomeperperson, such that higher residential electricity consumption, per person, reports a greater income per capita.

o   Beta = 0.97,

o   p = 0.0496,

o   confidence interval: from 0.0019 to       1.95

The linear regression model itself is characterized by a r square value of about 72%, where the r square indicates the amount of variability in the incomeperperson that is explained by explanatory variables.

Evaluation of regression model for evidence of misspecification

The further evaluation of the regression model for evidence of any misspecification is being conducted limiting the focus on 2 explanatory variables: Internetuserate and relectricperperson.

Residual plots are chosen as regression diagnostics to try to understand the cause of any misspecification by assessing visually any specification error.

Firstly, a centred explanatory variable, relectricperperson is being added to the regression equation featured by the linear term of the centred Internetuserate explanatory variable.

The intercept value of 8698.69 is the value of the incomeperperson at the mean of Internetuserate and relectricperperson. So the incomeperperson, when Internetuserate and relectricperperson are at their mean, is 8698.69 US$. The results also show that the coefficients for the linear and quadratic Internetuserate variables remain significant after adjusting for relectricperperson. relectricperperson is also statistically significant. The positive regression coefficient indicates that countries with a high residential electricity consumption, per person, tend to have a greater incomeperperson. Looking at r square, Internetuserate and residential electricity consumption, per person, together explain about 69.1% of the variability in incomeperperson.

Q-Q Plot of residuals analysis for incomeperperson

The goal is now to take a look at residual variability, i.e. how large the residuals are as well as whether regression assumptions are being met and whether there are any outlying observations that might be unduly influencing the estimation of the regression coefficients.

In order to check the residual variability, i.e. whether the residuals from the regression model are normally distributed, a Q-Q Plot is being used. It arises that the residuals generally follow the straight line, but deviate somewhat at the higher quantiles. This means that residuals do not follow perfect normal distribution. This indicates that the linear Internetuserate term might not fully estimate the linear association observed in the scatter plot. There may be other explanatory variables to be included eventually to improve estimation of the observed linear association.

Standardised residual analysis

To search for outliers, a plot of the standardized residuals by country for each of the observations is being examined. Looking at this plot, it comes up that most of the residuals fall within one standard deviation of the mean, i.e. between -1 or 1. Few countries (6 observations, i.e. 4,7%) have residuals that are more than 2 standard deviations above or below the mean of zero: this is a warning sign that it could be about some outliers. However, 95,3% of the values of the residuals falls between 2 standard deviations of the mean.

There are 3 observations that are three or more standard deviations from the mean: these might be extreme outliers.

In terms of evaluating the overall fit of the model, i.e. how well the model fits the data based on the distribution of the residuals, the plot results show that 2.3% of observations (3) has standardized residuals for the absolute value greater than 2.5, and 4.7% (6 observations) have an absolute value greater than or equal to 2.

Due to this 2.3% higher than the limit of 1%, there's evidence that the level of error within the model could be not unacceptable, i.e, the model's a fairly poor fit to the observed data. In order to improve the fit of the model, more explanatory variables should be included to better explain the variability in incomeperperson response variable.

Residual plot analysis for relectricperperson

When the analysis comes to look at relectricperperson residual plot to determine how specific this explanatory variable contribute to the fit of the model, it looks like there is a shaped pattern to the residuals where the absolute values of the residuals are significantly larger at higher values of the explanatory variable, but get smaller, that is closer to zero, as the explanatory variable decreases.

This behaviour is accordingly with the other regression diagnostic plot that shows that the model does not predict incomeperperson as well for countries that have high values of relectricperperson.

It seems that relectricperperson is featured by a curvilinear pattern to these observations, where the residuals get larger for countries for which relectricperperson exceeds 1152.85 kWh. This suggests that the association between relectricperperson and incomeperperson could be curvilinear. As a result, a second order polynomial term for relectricperperson might be added to the model as well.

Due to the fact that multiple explanatory variables are in place, it would be appropriate to take a look at the contribution of each individual explanatory variable to the model fit, controlling for the other explanatory variables. Therefore the partial regression residual plot is being returned.

Regression residual plot analysis for relectricperperson analysis

Looking at the partial regression residual plot for the relectricperperson explanatory variable, it comes up that it is about a scatter plot, showing the effect of adding relectricperperson as an additional explanatory variable to the model that includes only Internetuserate explanatory variables. Next step is to access whether relectricperperson residuals show a linear or nonlinear pattern. Taking a look at the plot for relectricperperson, accordingly with the plot of the residuals at different values of relectricperperson without adjusting for Internetuserate variables, the partial residual regression plot for relectricperperson might indicate a non linear association. This because many of the residuals are pretty far from the straight line after the value of 25 on the x axis,  indicating that a curvilinear shape might improve the residuals.

This would be additional support for adding a polynomial term for relectricperperson to the model.

Outlier and Leverage diagnostics analysis for incomeperperson

Looking at the leverage plot, 5 outlier observations are being shown. However, they have small, close to zero, leverage values, indicating that they do not seem to have a strong influence on the estimation of the regression parameters.

There are few cases (2 observations) with higher leverage values, that seems having not so much influence on the estimation of the predicted value of incomeperperson.

It is present 1 observation that is both high leverage and outlier and surely has influence on the estimation of the predicted value of incomeperperson.

Hereinafter the SAS code to return the output results of the aforementioned analysis (see jpgfiles attached as well)

/Program for GAPMINDER data set/ PROC IMPORT DATAFILE ='/home/u63783903/my_courses/gapminder_pds.csv' OUT = imported REPLACE; RUN; DATA new; set imported; /* the gapminder csv dataset is being uploaded and imported to the SAS - the dataset is being read and prepared for use */

LABEL incomeperperson = "Gross Domestic Product per capita in constant 2000 US$" Internetuserate="Internet users (per 100 people)" alcconsumption="alcohol consumption per adult (age 15+), litres" urbanrate="urban population (% of total)" employrate="total employees age 15+ (% of population)" relectricperperson="residential electricity consumption, per person (kWh)";

if incomeperperson ne . and alcconsumption ne . and Internetuserate ne . and urbanrate ne . and employrate ne . and relectricperperson ne . ; PROC MEANS; var Internetuserate alcconsumption urbanrate employrate relectricperperson; run;

centering quantitative explanatory variables by subtracting the mean value from the actual value for each observation, and this to be done for each explanatory variable; data new2; set new;

additional data management needs the creation of a new temporary data set; if incomeperperson ne . and alcconsumption ne . and Internetuserate ne . and urbanrate ne . and employrate ne . and relectricperperson ne . ;

centered variables being used in the regression analysis; Internetuserate_c=Internetuserate-39.4496407; alcconsumption_c=alcconsumption-7.3729134; urbanrate_c=urbanrate - 61.2255118; employrate_c=employrate-57.6897637; relectricperperson_c=relectricperperson-1152.85; run;

check coding to assess whether the variables are properly centered by calculating the mean of each centered variable using the means procedure; PROC MEANS; var Internetuserate_c alcconsumption_c urbanrate_c employrate_c relectricperperson_c; run;

linear regression model with the calculation of the 95% confidence levels also included; PROC glm; model incomeperperson=Internetuserate_c/solution clparm; run;

PROC glm; model incomeperperson=Internetuserate_c relectricperperson_c/solution clparm; run;

PROC glm; model incomeperperson=Internetuserate_c urbanrate_c/solution clparm; run;

PROC glm; model incomeperperson=Internetuserate_c employrate_c/solution clparm; run;

PROC glm; model incomeperperson=Internetuserate_c alcconsumption_c/solution clparm; run;

add quadratic term; PROC glm; model incomeperperson=Internetuserate_c Internetuserate_c*Internetuserate_c/solution clparm; run;

scatterplot for x variable, Internetuserate and y incomeperperson with linear and quadratic regression line the 95% confidence interval for the regression line being printed as well; proc sgplot; reg x=Internetuserate y=incomeperperson / lineattrs=(color=blue thickness=2) degree=1 clm; reg x=Internetuserate y=incomeperperson / lineattrs=(color=green thickness=2) degree=2 clm; yaxis label="Gross Domestic Product per capita in constant 2000 US$"; xaxis label="Internet users (per 100 people)"; run;

add cubic term; PROC glm; model incomeperperson=Internetuserate_c Internetuserate_cInternetuserate_c Internetuserate_cInternetuserate_c*Internetuserate_c/solution clparm; run;

linear regression model; PROC glm; model incomeperperson=alcconsumption_c/solution clparm; run;

linear regression model; PROC glm; model incomeperperson=alcconsumption_c Internetuserate_c /solution clparm; run;

add quadratic term to test whether adding a second order polynomial term to the regression model returns a significantly better fitting model; PROC glm; model incomeperperson=alcconsumption_c alcconsumption_c*alcconsumption_c/solution clparm; run;

add cubic term; PROC glm; model incomeperperson=alcconsumption_c alcconsumption_calcconsumption_c alcconsumption_calcconsumption_c*alcconsumption_c/solution clparm; run;

linear regression model; PROC glm; model incomeperperson=urbanrate_c/solution clparm; run;

linear regression model; PROC glm; model incomeperperson=urbanrate_c Internetuserate_c/solution clparm; run;

add quadratic term; PROC glm; model incomeperperson=urbanrate_c urbanrate_c*urbanrate_c/solution clparm; run;

add cubic term; PROC glm; model incomeperperson=urbanrate_c urbanrate_curbanrate_c urbanrate_curbanrate_c*urbanrate_c/solution clparm; run;

linear regression model; PROC glm; model incomeperperson=employrate_c/solution clparm; run;

linear regression model; PROC glm; model incomeperperson=relectricperperson_c/solution clparm; run;

linear regression model; PROC glm; model incomeperperson=relectricperperson_c Internetuserate_c/solution clparm; run;

add quadratic term; PROC glm; model incomeperperson=relectricperperson_c relectricperperson_c*relectricperperson_c/solution clparm; run;

add cubic term; PROC glm; model incomeperperson=relectricperperson_c relectricperperson_crelectricperperson_c relectricperperson_crelectricperperson_c*relectricperperson_c/solution clparm; run;

PROC glm; model incomeperperson = Internetuserate_c alcconsumption_c urbanrate_c employrate_c relectricperperson_c/solution clparm; run;

EVALUATING MODEL FIT *;

multiple regression adding variables to the model in order to evaluate multiple predictors over quantitative response variable, incomperperson - the calculation of the 95% confidence levels is also included ;

PROC glm; model incomeperperson=Internetuserate_c relectricperperson_c /solution clparm; run;

request regression diagnostic plots to evaluate the overall fit of the predicted values of the response variable to the observed values, and look for outliers ; PROC glm PLOTS(unpack)=all; model incomeperperson=Internetuserate_c relectricperperson_c /solution clparm; output residual=res student=stdres out=results; run;

plot of standardized residuals for each observation To evaluate the overall fit of the predicted values of the response variable to the observed values, and look for outliers. A plot of the standardized residuals for each of the observations is being examined ; proc gplot; label stdres="Standardized Residual" country="Country"; plot stdres*country/vref=0; run;

using proc reg to get a partial regression plot;

partial regression plot; PROC reg plots=partial; model incomeperperson=Internetuserate relectricperperson /partial; run;

[shristisahu]

Causal Analysis: A Detailed Guide for Comprehensive Understanding

Originally Published on Quantzig: Causal Analysis: Comprehensive Guide for Detailed Understanding

Key Takeaways for Causal Analysis

Causal analysis offers a powerful framework for identifying and understanding the relationships between various factors influencing outcomes. This approach allows businesses to pinpoint the drivers of success and challenges, enabling informed, data-driven decisions on pricing strategies, marketing campaigns, and other critical areas. By leveraging historical data to predict future outcomes, causal analysis facilitates proactive decision-making. Systematic application helps businesses refine and optimize operations, marketing efforts, and strategies, allowing them to adapt to evolving market dynamics and consumer behavior, thereby ensuring sustained growth and success.

Introduction

In the rapidly evolving field of data analytics, comprehending the cause-and-effect relationships between variables provides invaluable insights for data-driven decision-making. Causal analysis is a potent tool that helps untangle the complex web of interrelated factors. This article delves into the realm of causal analysis, exploring its practical applications and transformative impact on optimizing outcomes.

Causal Analysis

Evolution of Causal Models

Traditional machine learning models are excellent at detecting patterns and making predictions but often fall short in explaining the underlying reasons for these patterns. Identifying mere statistical correlations is not enough; understanding the causal mechanisms is crucial for informed decision-making. To achieve this, it is essential to uncover the causal relationships within the data, going beyond mere mathematical correlations. While traditional machine learning highlights patterns, causal learning provides deeper insights into the forces driving these trends.

Causal Inference and Use Cases

Causal inference aims to determine whether a specific treatment or action led to the observed result, focusing on root causes rather than just apparent effects. A root cause is the fundamental reason behind an occurrence, even if it seems distant from the observed effect.

Causal models help address several critical questions:

What is the typical impact of a treatment on the outcome, and how does it function?

How will the treatment affect an individual unit, positively or negatively?

Is there a cause-and-effect relationship between the treatment and the outcome?

What is the impact of the treatment on customer profitability, and which treatment should be recommended in scenarios with multiple options?

Causal Model Implementation Methods

Implementing causal models in real-world scenarios involves various approaches. Here are several methods to construct effective causal models:

Matching Matching involves grouping parameters with similar propensity scores, indicating the similarity between treated and untreated units. Differences within these matched groups help calculate metrics such as the average treatment effect (ATE) and the average treatment effect on the treated (ATT).

Stratification This method divides a larger group into smaller, more similar groups based on specific characteristics. Within these groups, treated and control units are matched as closely as possible, and the treatment effect is calculated for each subgroup to control for selection bias.

Doubly Robust Learning Used when data sets have numerous variables or when relationships between variables are not accurately represented by statistical models, this method is particularly effective for categorical treatment variables with measured confounding variables.

Forest-Based Estimator Suitable for data with many attributes, this flexible nonlinear method approximates variable treatment effects and confidence intervals.

Meta-Learners Effective for scenarios with multiple response variables, meta-learners include:

T-Model: Evaluates conditional expectations for control and treatment groups independently.

S-Model: Uses all indicators without giving special treatment to the treatment variable.

X-Model: Utilizes control group data to evaluate treatment effects, managing overfitting.

Domain Adaptation Model: Represents covariate shifts among treatment arms.

Double Machine Learning (DML) It addresses regularization and overfitting biases through orthogonalization and cross-fitting, providing accurate estimates of individualized treatment effects for classification and single-response variable problems.

Methods of Causal Analysis

To perform successful causal analysis, data analysts follow a structured process:

Define the issue and identify relevant factors.

Collect data through surveys, experiments, or existing datasets.

Use data exploration techniques like descriptive statistics and visualization to uncover trends and connections.

Apply statistical methods such as regression analysis to establish causation.

Consider alternative explanations and complicating factors to ensure robust causal links.

Applications in Business Decisions

Causal analysis enables businesses to move beyond simple correlations by examining variables affecting specific outcomes. For instance, a retail company aiming to increase revenue might analyze consumer behaviors, historical sales data, marketing efforts, pricing strategies, and seasonal patterns. By identifying influential elements, the company can allocate resources effectively, refine marketing plans, and optimize pricing strategies to boost sales.

In healthcare, understanding causal relationships between patient outcomes and treatment plans can lead to improved care. By analyzing patient data, medical professionals can identify the most effective treatments for specific conditions, enhancing patient outcomes and reducing healthcare costs.

Beyond Predictive Analytics

While predictive analytics forecasts future trends, causal analysis reveals the underlying drivers of these trends. Identifying causal factors allows organizations to take proactive measures to achieve desired outcomes. Whether optimizing supply chains, enhancing customer experiences, or reducing product defects, causal analysis provides the insights necessary for taking proactive steps and achieving positive results.

Conclusion

Causal analysis is essential for comprehensively understanding consumer behavior, business operations, and market trends. By integrating traditional machine learning with advanced data analytics techniques, businesses can uncover cause-and-effect relationships within their data. Through rigorous causal models and inference, organizations can make informed decisions about pricing strategies, promotional campaigns, and other critical operations.

Systematic data exploration helps identify variables influencing revenue growth, consumer behaviors, and pricing while highlighting seasonal trends. Embracing causal analysis empowers businesses to optimize strategies, enhance consumer experiences, and achieve sustainable success in a competitive marketplace.

Contact us.

[support1212]

baddest predict,

baddest predict,

In an age where data reigns supreme and predictive analytics promise to revolutionize decision-making processes across industries, there lurks a shadowy side: the baddest predict. This phenomenon, characterized by inaccurate forecasts, flawed models, and unforeseen consequences, underscores the inherent risks of relying too heavily on predictive analytics without proper scrutiny.

Predictive analytics, powered by advanced algorithms and vast datasets, hold the potential to forecast trends, optimize operations, and even mitigate risks. From finance to healthcare, retail to transportation, organizations are increasingly turning to predictive models to gain a competitive edge and enhance efficiency. However, the allure of predictive analytics can sometimes blind decision-makers to the pitfalls that lie beneath the surface.

The journey towards the baddest predict often begins with flawed assumptions or incomplete data. Whether it's overlooking crucial variables, relying on outdated information, or succumbing to biases, even the most sophisticated algorithms can falter when fed faulty inputs. Furthermore, the complexity of real-world systems introduces layers of uncertainty that algorithms may struggle to navigate, leading to erroneous predictions.

One notable example of the baddest predict in action occurred in the financial sector during the 2008 global financial crisis. Banks and financial institutions, armed with complex risk models and algorithms, failed to anticipate the impending collapse of the housing market and the subsequent ripple effects on the economy. The reliance on historical data and oversimplified models obscured the underlying vulnerabilities, resulting in catastrophic losses and widespread economic turmoil.

Similarly, in the realm of healthcare, predictive analytics have been heralded as a game-changer for disease detection and treatment optimization. Yet, instances of misdiagnosis or ineffective treatment plans underscore the limitations of predictive models in the face of evolving medical complexities. Factors such as genetic variability, environmental influences, and individual behaviors can confound predictions, leading to suboptimal outcomes for patients.

Moreover, the deployment of predictive analytics in sensitive domains such as law enforcement and criminal justice has raised concerns about fairness and bias. Biased data inputs or flawed algorithms can perpetuate systemic inequalities, leading to unjust outcomes for marginalized communities. The infamous case of predictive policing algorithms exacerbating racial profiling and discriminatory practices serves as a stark reminder of the ethical dilemmas inherent in algorithmic decision-making.

So, how can organizations guard against the allure of the baddest predict? Firstly, a critical appraisal of data quality and model assumptions is paramount. Robust validation processes, thorough sensitivity analyses, and continuous model monitoring can help identify and mitigate potential sources of error. Additionally, interdisciplinary collaboration between data scientists, domain experts, and ethicists can foster a more holistic understanding of the predictive process and its implications.

Furthermore, transparency and accountability must be prioritized to foster trust and mitigate risks. Clear communication of model limitations, potential biases, and decision-making criteria can empower stakeholders to make informed judgments and challenge flawed assumptions. Moreover, mechanisms for recourse and redress should be established to address instances of algorithmic harm or unintended consequences.

In conclusion, while predictive analytics hold immense promise for driving innovation and enhancing decision-making, the specter of the baddest predict looms large. By acknowledging the inherent uncertainties and limitations of predictive models, and adopting a proactive approach to validation, transparency, and accountability, organizations can navigate the complexities of predictive analytics with greater confidence and integrity. Only then can the true potential of predictive analytics be realized without succumbing to the perils of the baddest predict.

[iambriannelson]

Unveiling the Mysteries of Control Variables in Research blog at Greatassignmenthelper

Unlock the secrets of control variables with Greatassignmenthelper in our latest blog! Understanding the significance of control variables is crucial for robust research methodologies. These variables act as guardians, ensuring that your study remains focused and accurate.

In this insightful read, Greatassignmenthelper sheds light on the role of control variables in research design, explaining how they enhance the reliability of your findings. Learn how to identify and manipulate these variables effectively to eliminate confounding factors and achieve meaningful results.

Dive into the world of research sophistication and elevate your academic prowess. Read the complete blog at Greatassignmenthelp and embark on a journey to master the art of control variables. Your key to unlocking research excellence awaits #ResearchMethods #AcademicExcellence #Greatassignmenthelp

[narwhalhatblues]

Standard and Specialized Treatment of Depression in Autistic Adults Without Intellectual Disability: Literature Review (for school, not published)

Within the last thirty or so years, researchers have learned a lot about Autism Spectrum Disorder. Between 1991 and 2020, the proportion of published research about Autistic adults (as opposed to children) approximately doubled. Still, research about Autistic children has far outpaced the growth of research about Autistic adults (Kirby & McDonald, 2021). Furthermore, many studies regarding Autistic adults demonstrate positive changes in “individual functioning” as Autistic children age into adulthood, but “functioning” is defined as a decrease in the autistic symptoms which make navigating life in an ableist world more difficult, rather than improved overall mental health outcomes (Shattuck et al., 2019).

In fact, it is well-established that Autistic people are at a higher risk for depression than non-Autistic, or “allistic” people (Menezes et al., 2020), and co-occurring depression in Autistic people is associated with poor quality-of-life, inpatient hospitalization, and medical illness (Rosen et al, 2018). Unfortunately, meta-analyses of research within the last few years have shown poor or minimal evidence for success (and perhaps even risk of harm) with standard modes of treatment for depression when used with Autistic clients (Linden et al., 2022; Menezes et al., 2020; White et al, 2018). This has led to the recent development of research discerning how treatment may be specialized or adapted specifically for use with Autistic clients (Bal et al., 2022; Brewe et al., 2021; Hume, 2022; Tomaszewski et al, 2022) and research identifying potential barriers to treatment, or confounding variables which may make standard treatments more difficult or complex (Price, 2022; Zheng et al., 2021).

This literature review aims to synthesize available research regarding therapeutic treatment of Major Depressive Disorder in Autistic adults without intellectual disability to discern whether standard treatments are effective, showcase research proposing adaptations to standard treatments for use specifically with Autistic adults, and identify potential barriers to or complicating factors of treatment. The goal of consolidating available research is to provide an overview that is more accessible for anyone who wishes to serve their Autistic clients more effectively, train clinicians on current research, or perhaps even develop a new kind of therapeutic treatment that could incorporate as much available research as possible, rather than focusing on pre-existing modes of treatment. This literature review will not discuss pharmacological interventions for depression in Autistic adults.

Treatment as Usual

Systematic Reviews and Meta-Analyses

In the last few years, meta-analyses have been published studying the efficacy of standard treatments for depression (and other co-occurring conditions) in Autistic adults. The first systemic review in over a decade was published by Menezes et al. (2020). The Menezes study systematically examined several standard depression treatments for depression in Autistic youth and young adults. Twenty psychosocial treatment studies met the criteria of inclusion and studied the effectiveness of cognitive remediation therapy, behavioral therapy, cognitive behavioral therapy, combined psychosocial intervention, mindfulness-based therapy (MBT), and social, academic, and/or vocational skills training. The strength of the evidence in the studies was generally poor, with MBT showing preliminary efficacy, and “inconsistent” evidence (for or against) Cognitive Behavioral Therapy (CBT).

A second meta-analysis was conducted by Linden et al. (2022). This study examined randomized controlled trials studying the treatment of both anxiety and depression in Autistic people. The meta-analysis identified 71 randomized control trials with 3,630 participants and found that the certainty of evidence that studied (standard) interventions could improve the mental health of autistic people was very low (Linden et al., 2022). The review stated that a high number of studies were based on an individualistic model which suggested that Autistic people themselves need to change who they were as people. Many interventions were targeted at reducing core autistic features rather than targeting the mental health conditions at issue. The review declared, “the routine use of interventions to manage core features of autism with a view to improve mental health conditions of Autistic people should be avoided” (Linden et al, 2022). The evidence from the Linden review suggested to the researchers that “some forms” of cognitive behavioral therapy may decrease depression scores in Autistic adults, and mindfulness therapy may decrease depression scores in Autistic adults with previous mental health conditions (Linden et al., 2022).

It is important to note, as an aside, that behavioral interventions based on the standards of Applied Behavioral Analysis did not present any evidence for efficacy for Autistic adults in the Linden et al. meta-analysis. In fact, a paper by Henny Kupferstein (2018) which surveyed 243 adults ages 18-73 found that 46 percent of ABA-exposed respondents met the diagnostic threshold for PTSD, with 47 percent of the affected subgroup experiencing extreme levels of severity.

Assessment

While standard treatments may need adjustment to increase efficacy, a study conducted by Williams et al. (2020) has shown that the primary diagnostic tool for depression, the Beck Depression Inventory-II, is a valid measure of depressive symptoms in autistic adults, and is appropriate for quantifying depression severity in research studies or screening for depressive disorders in clinical setting. In other words, while treatment outcomes between Autistic and non-autistic adults differ (for instance, a 2017 article by David et al. called CBT the “gold standard” of psychotherapy), clinicians and researchers can remain confident that the standard depression measurement tool is not confounded by the presence of Autistic symptoms.

Adaptation of Standard Treatment

Sonny Jane Wise, a self-identified Autistic Bipolar ADHDer, developed the Neurodivergent Friendly Workbook of DBT Skills (2022), but no studies have yet been conducted to find whether this workbook is more effective than standard DBT treatment for Autistic adults.

Therapeutic Alliance

A study published by Brewe et al. (2021) examined the trajectory of therapeutic alliance at four timepoints during a 16-week mindfulness-based training targeting emotion regulation problems in Autistic adolescents and young adults aged 12 to 21 years old. While the exclusion criteria for this study included having a diagnosis of a psychiatric disorder other than Autism, the findings are applicable to this paper’s subject matter as emotion regulation has been implicated as a potential pathway for psychiatric disorders (Brewe et al., 2021). The study found that stronger alliance predicted decreased dysphoria at posttreatment. 

A separate, thoroughly sourced study written by Romy Hume (2022) discusses how texts relevant to therapeutic alliance caution therapists that the task of developing a supportive, empathic relationship between therapist and client may be particularly challenging, blaming such difficulty on the language and social skill deficits of Autistic clients. Some texts even argue that investing energy in building a therapeutic alliance may not be prudent, presuming a low chance of success (Hume, 2022). In disagreement, Hume points to research that demonstrates higher affective empathy and desire for social relationships in Autistic people, stating that hesitance to engage in new relationships may be the result of repeated interpersonal trauma. Hume argues that epistemic injustice, diagnostic overshadowing, and the double empathy problem may be at the root of the discrepancy between the viewpoints of therapists and Autistic clients. Examining available research, Hume suggests that relationship-building with Autistic adults may require adjustment of pace to reduce cognitive overload, using clear, unambiguous, literal language, refraining from correcting the client or enforcing an agenda, and creating a sense of predictability in the timing, location, and procedure of therapy. However, the research conducted on therapist alliance in Autistic clients before Hume’s article involved Autistic children and adolescents, not adults.

Hume interviewed 17 Autistic adults and 3 non-Autistic mental health professionals about their experiences and recommendations for improved relationship building. This research found that direct, explicit, practical unconditional positive regard was of heightened importance with Autistic adult clients, challenging general recommendations that unconditional positive regard should be expressed indirectly (Hume, 2022). Additionally, Hume found that Autistic clients could become confused by perceived inauthentic behavior from their therapists, decreasing levels of trust. Some indirect behaviors that are often considered socially ideal, such as mirroring physical posture, felt like emotional manipulation to Autistic clients (Hume, 2022). Hume also noted that offers of food and drink made a profoundly positive impact on Autistic clients, positively shifting the therapy atmosphere and associated expectations, challenging existing research that shows that sharing food with clients is not a common practice among psychologists (Hume, 2022). Another difference between Autistic and non-Autistic clients is the subject matter of discussion with therapists. Autistic clients wanted their therapists to ask about and remember their enthusiasms, likes and dislikes, and their hopes and dreams, rather than solely focusing on problems (Hume, 2022).

Goal Development, Self-Determination, and Guided Self-Help

A few studies have examined specific therapeutic techniques which may be adapted for Autistic clients. A study on adapting behavioral activation conducted by Bal et al. (2022) found that the construction of goals for Autistic people may be different from typically developing populations. Similarly, Tomazewski et al. (2022) emphasized the importance and complexity of self-determination in transition-aged Autistic adults, noting a lack of support for specific skills associated with a successful transition to adulthood. Self-determination refers to a set of beliefs, knowledge and skills such as self-awareness, decision making, and goal setting that enable someone to engage in self-directed behavior and pursue goals and desires in areas a person feels important to them (Wehmeyer, 1998). Participants in the Tomazewski study included 237 autistic transition-aged youth and a subsample of their 198 caregivers who completed self-determination measures. This study found that increased levels of depression and executive functioning difficulties were associated with decreased capacity for self-determination (Tomazweski, 2022). Another adaptation to assist Autistic adults with self-help issues was developed by Russell et al. (2019) in Winchester, England. Researchers developed a low-intensity guided self-help model of CBT and recruited 70 participants with a mean age of 38 years, all of whom were Autistic and met diagnostic criteria for depression. Five therapist coaches and 21 trial participants were interviewed. This qualitative study found some evidence that guided self-help intervention may be effective in reducing depressive symptoms (Russell, 2019).

Barriers to Treatment and Confounding Variables

It makes sense that goal development, self-determination and guided self-help may be especially potent areas for therapists to target with Autistic clients, and lack of knowledge about the specific challenges of Autistic people may prevent therapists from adequately understanding how to assist clients with the development of these skills. A study conducted by Zheng et al. (2021) collected information from 315 Autistic young adults diagnosed in childhood, approximately two-thirds of whom were also diagnosed with depression. Depressed females were about 3.5 times more likely than depressed males to have a depression diagnosis, and the likelihood of receiving depression treatment was higher among those with a formal depression diagnosis and with higher levels of education. Among currently depressed adults, nearly one-half reported having barriers to depression-related services. The most frequently mentioned concern was financial issues and insurance coverage, with more than half struggling to pay for the services they needed (Zheng et al, 2021). Participants also brought up the issue of limited access to appropriate care. One respondent stated, “Most therapy is geared for neuro-normative people. Therapists struggle [to] understand that I am Autistic and what being Autistic means.” Other barriers included the logistics of finding transportation and managing schedules, difficulties describing their feelings to professionals and labeling their emotions (e.g. alexithymia), other co-occurring symptoms (e.g. fatigue, anxiety, and fear) preventing them from seeking care, treatment side effects, and family members’ lack of understanding.

The study by Zheng et al. (2021) noted that a growing body of research suggests that gender identity is related to a greater risk for mental health problems, and Autistic people are more likely than the general population to be gender diverse. The issue of gender diversity in Autistic people is incredibly complex and an area of emerging advocacy and research (Adams, 2022). An examination of available autobiographical literature by Noah Adams (2022) found that transgender Autistic people cite experiences of Autism diagnosis, community, coming out as trans, and gender as meaningful in their lives.

Masking

Dr. Devon Price’s book Unmasking Autism (2022) provides extensive scholarly coverage of the concept of Autistic masking, and how masking both harms and helps Autistic mental health. Masking is defined by Price as a common coping mechanism in which Autistic people hide their identifiably Autistic traits in order to fit in with societal norms, adopting a superficial personality at the expense of their mental health. Price also provides practical guidance and exercises which may assist in the development of self-determination skills, with the aim of decreasing the harmful effects of masking and increasing understanding of oneself.

Diagnosis

The issue of professional diagnosis is one that is fraught within the adult Autistic community for a plethora of reasons, including financial woes and poor clinician understanding of high-masking Autism (Huang, 2020). Furthermore, symptoms of Autism may appear similar to depression, especially in high-masking adults (Williams et al., 2020). A review of co-occurring psychiatric conditions with autism noted that Autistic people may present with symptoms of depression that are distinct from non-autistic depressed people. For instance, depression may manifest as an increase in behaviors identifiable as symptoms of autism, including stimming and focusing on special interests, the latter of which may be morbid in nature (Rosen et al., 2018). That same review identified that an increase in Autistic symptoms may include decreased ability to communicate effectively, which some researchers call “regression,” and an increase in self-injurious behaviors.

Autistic Burnout

Autistic burnout is not a DSM diagnosis, but further research about this unique phenomenon may inform the diagnosis and treatment of depression in Autistic adults (Higgins et al., 2021). A study by Raymaker et al. (2020) suggests that autistic burnout appears to be distinct from occupational burnout or clinical depression. The primary characteristics of autistic burnout are described as chronic exhaustion, loss of skills, and reduced tolerance to stimulus. A second study conducted by Higgins et al. (2021) came to a similar conclusion, and added that behavioral activation treatments may be contra-indicated for clients experiencing autistic burnout.

Conclusion

Autistic culture is extremely complex. As such, this literature review remains insufficient as a total encapsulation of the issues surrounding treatment of depression in Autistic adults. For instance, this review fails to discuss or examine issues of racialized Autism (Brown et al., 2017) and other intersectional concerns other than gender identity.

A study by Benevides et al. (2020) found that Autistic people would like future research to prioritize improving quality of life and social well-being outcomes. Respondents to the survey also indicated they would like further research on societal barriers such as stigma and discrimination against Autistic people that leads to exclusion, bullying and other forms of trauma. Ultimately, this literature review concludes that because standard treatments for depression in Autistic people have been shown to provide inefficient evidence for efficacy, expertise in Autistic issues, adaptation of current treatments, and further research are necessary in order to more adequately assist Autistic clients with depression.

References

Adams, N. (2022). Autistics Never Arrive: A Mixed Methods Content Analysis of 

Transgender and Autistic Autobiography. Bulletin of Applied Transgender Studies 1 

(1-2): 145-161. https://doi.org/10.57814/de1e-gj97

Bal, V. H., Wilkinson, E., Glascock, V., Hastings, R. P., & Jahoda, A. (2022). Mechanisms of 

Change in Behavioral Activation: Adapting Depression Treatment for Autistic People. 

Cognitive and Behavioral Practice. https://doi.org/10.1016/j.cbpra.2022.03.006

Benevides, T. W., Shore, S. M., Palmer, K., Duncan, P., Plank, A., Andresen, L., Caplan, R., 

Cook, B., Gassner, D., Hector, B. L., Morgan, L., Nebeker, L., Purkis, Y., Rankowski, B., 

Wittig, K., & Coughlin, S. S. (2020). Listening to the autistic voice: Mental health 

priorities to guide research and practice in autism from a stakeholder-driven project. Autism, 24(4), 822-833. https://doi.org/10.1177/1362361320908410

Brewe, A.M., Mazefsky, C.A., & White, S.W. (2021). Therapeutic Alliance Formation for 

Adolescents and Young Adults with Autism: Relation to Treatment Outcomes and Client 

Characteristics. Journal of Autism and Developmental Disorders 51, 1446–1457. https://doi.org/10.1007/s10803-020-04623-z

Brown, L.X.Z., Ashkenazy, E., Onaiwu, M.G. (2017). All the Weight of Our Dreams: On Living 

Racialized Autism. DragonBee Press.

David, D., Cristea, I., & Hofmann, S. G. (2017). Why Cognitive Behavioral Therapy is the 

Current Gold Standard of Psychotherapy. Frontiers in Psychiatry, 9(4). 

Higgins, J. M., Arnold, S. R., Weise, J., Pellicano, E., & Trollor, J. N. (2021). Defining autistic 

burnout through experts by lived experience: Grounded Delphi method investigating 

#AutisticBurnout. Autism. https://doi.org/10.1177/13623613211019858

Hume, R. (2022). Show Me the Real You: Enhanced Expression of Rogerian Conditions in 

Therapeutic Relationship Building with Autistic Adults. Autism in Adulthood, 4(2), 

151-163. http://doi.org/10.1089/aut.2021.0065

Huang, Y., Arnold, S. R., Foley, K. R., & Trollor, J. N. (2020). Diagnosis of autism in adulthood: 

A scoping review. Autism, 24(6), 1311-1327. http://doi.org/10.1177/1362361320903128

Linden, A., Best, L., Elise, F., Roberts, D., Branagan, A., Tay, Y. B. E., Crane, L., Cusack, J., 

Davidson, B., Davidson, I., Hearst, C., Mandy, W., Rai, D., Smith, E., & Gurusamy, K. 

(2022). Benefits and harms of interventions to improve anxiety, depression, and other 

mental health outcomes for autistic people: A systematic review and network 

meta-analysis of randomized controlled trials. Autism. 

Kirby, A. V., & McDonald, K. E. (2021). The State of the Science on Autism in Adulthood: Building an Evidence Base for Change. Autism in Adulthood: Challenges and Management, 3(1), 2-4. https://doi.org/10.1089/aut.2020.29018.avk

Kupferstein, H. (2018). Evidence of increased PTSD symptoms in Autistics exposed to applied behavior analysis. Advances in Autism 4(3). http://doi.org/10.1108/AIA-08-2017-0016

Menezes, M., Harkins, C., Robinson, M. F., & Mazurek, M. O. (2020). Treatment of Depression in Individuals with Autism Spectrum Disorder: A Systematic Review. Research in Autism Spectrum Disorders, 78, 101639. https://doi.org/10.1016/j.rasd.2020.101639

Price, D. (2022). Unmasking autism: discovering the new faces of neurodiversity (First edition.). Harmony Books.

Raymaker, D.M., Teo, A.R, Steckler, N.A, Lentz, B., Scharer, M, Santos, A.D., Kapp, S.K., Hunter, M., Joyce, A., & Nicolaidis, C. (2020). “Having All of Your Internal Resources Exhausted Beyond Measure and Being Left with No Clean-Up Crew”: Defining Autistic Burnout. Autism in Adulthood, 2(2), 132-143. http://doi.org/10.1089/aut.2019.0079

Rosen, T. E., Mazefsky, C. A., Vasa, R. A., & Lerner, M. D. (2018). Co-occurring psychiatric conditions in autism spectrum disorder. International Review of Psychiatry, 30(1), 40–61. http://doi.org/10.1080/09540261.2018.1450229

Russell, A., Gaunt, D., Cooper, K., Horwood, J., Barton, S., Ensum, I., Ingham, B., Parr, J., Metcalfe, C., Rai, D., Kessler, D., & Wiles, N. (2019). Guided self-help for depression in autistic adults: the ADEPT feasibility RCT. Health technology assessment (Winchester, England), 23(68), 1–94. https://doi.org/10.3310/hta23680

Tomaszewski, B., Klinger, L.G., & Pugliese, C.E. (2022). Self-Determination in Autistic 

Transition-Aged Youth without Intellectual Disability. Journal of Autism and 

Developmental Disorders 52, 4067–4078. https://doi.org/10.1007/s10803-021-05280-6

Shattuck, P. T., Garfield, T., Roux, A. M., Rast, J. E., Anderson, K., Hassrick, E. M., & Kuo, A. (2019). Services for Adults With Autism Spectrum Disorder: A Systems Perspective. Current Psychiatry Reports, 22(3). https://doi.org/10.1007/s11920-020-1136-7

Wehmeyer, M. L. (1998). Self-Determination and Individuals with Significant Disabilities: Examining Meanings and Misinterpretations. Journal of the Association for Persons with Severe Handicaps. https://doi.org/10.2511/rpsd.23.1.5

White, S.W., Simmons, G.L., Gotham, K.O., Connor, C.M., Smith, I.C., Beck, K.B., & 

Mazefsky, C.A. (2018). Psychosocial Treatments Targeting Anxiety and Depression in 

Adolescents and Adults on the Autism Spectrum: Review of the Latest Research and 

Recommended Future Directions. Current Psychiatry Reports, 20(82). 

Williams, Z. J., Everaert, J., & Gotham, K. O. (2020). Measuring depression in Autistic adults: 

Psychometric validation of the Beck Depression Inventory–II. Assessment, 28(3). 

Wise, S. J. (2022). The Neurodivergent Friendly Book of DBT Skills. Independently published.

Zheng, S., Adams, R., Taylor, J. L., Pezzimenti, F., & Bishop, S. L. (2021). Depression in 

independent young adults on the autism spectrum: Demographic characteristics, service use, and barriers. Autism. https://doi.org/10.1177/13623613211008276

[scripttorture]

In what instances would ICURE fail to change someone's beliefs? Would access to outside information or (not very good at their jobs) guards discussing events without tact help someone realise something fishy was going on?

Very broad question without clear, satisfying answers Anon.

 Basically: there is no guaranteed way to change a person’s mind. There are strategies we know can’t work. But everything that has a chance of success also has a chance of failure. And it isn’t a clear cut thing that I can give you a clean list of factors for.

 Humans are difficult creatures to study because they come with a lot of inbuilt confounding factors and individual variation. This makes it very difficult to identify clear reasons why something didn’t work. Because we have to assume that multiple factors are at work and the interaction of those factors may be as important as each factor individually.

 Even if you write your villains performing ICURE ‘perfectly’ there is still a chance of failure. And therefore it is realistic for you to decide it fails for this character.

 We can’t really study which ‘bits’ of ICURE are most effective. Partly because of that little thing called ethics but partly because setting up a study would be incredibly difficult. It’s hard enough to measure belief. Finding a large enough sample size, controlling for every possible confounding factor or variable and studying people for the years it would take to get any answers… It’s a big ask. It’s probably never going to happen.

 So with the caveat that we can’t tell if any of the parts of ICURE are more important let’s talk about how they can break down.

 ICURE, for everyone who hasn’t heard me talk about it before, is a set of techniques which can (sometimes) be used to manipulate a person into changing their views. They take months or years to have any real effect and as mentioned they’re not always successful.

 The acronym stands for Isolate, Control information, create Uncertainty, Repetition and Emotive responses. And if you’re writing a story where villains are trying to apply this but not doing it well it can break down at literally any one of these points.

 I would say based on what I’ve read that different groups focus more heavily on different aspects depending on their setting and strategy. Groups that are straight up kidnapping or imprisoning people often seem to focus more heavily on isolation and controlling information but often fall down on the other three. Whereas the impression I get of cults and some extremist political groups is that they focus more on creating uncertainty and emotive responses, which they can then use to further isolate members from family and friends.

 Controlling information is a common place for ICURE to break down nowadays. The rise of the internet and the decreasing size of devices has made it easier for victims to access unauthorised sources even when imprisoned.

 But repetition is also a very common place for ICURE to break down because in large groups not every individual is going to follow the same script perfectly. Group members can also undermine ICURE by lashing out, physically or verbally, driving their target away.

 Creating uncertainty doesn’t always work. Sometimes victims straight up do not believe what they’re told. Some attempts to create uncertainty around core beliefs lead to a knee-jerk rejection of what’s being said. Sometimes targets know more about a given subject then the person trying to create uncertainty and as a result the attempt is absurdly obvious.

 Emotive responses are similarly… charged. Attempting to instil a sense of disgust or rejection of something an individual supports won’t always work. Over a long period of time it can. But I can think of a lot of cases where it has instead taught individuals to lie to the group, hide their beliefs or activities and served to drive them away from the group.

 Isolation is either difficult or easy depending on the context of the story. A character who is in a literal prison can easily be isolated from anyone but vetted individuals. A character who has been targetted by a cult, but is still going about normal day to day business, is a lot harder to isolate completely.

 Cults and extremist groups tend to rely on uncertainty, repetition and emotive responses because they know that if they can shift a target’s beliefs the target will isolate themselves.

 Let me give you an example to illustrate this. Imagine a country where there’s a big, culturally important celebration that involves eating candied orange peel and wearing red. Now imagine a cult within the country that rejects candy as sinful and wearing red as a sign of bad character.

 A character targetted by this cult might feel increasingly uncomfortable with this festival. May be at first they go with their friends and family, wear red but don’t eat the candied orange peel. May be the year after they decide not to go, missing a chance to spend time with their friends and family. May be a few years later their rejection of the festival is so deep they try to persuade their friends and family not to go.

 This leads to a big argument. They and their friends/family say things in the heat of the moment. Now all sides are upset and communication becomes harder.

 These kinds of patterns of behaviour lead to the target isolating themselves from friends and family, as their views become more extreme and drive away people who aren’t members of the cult.

 But crucially they can still choose to socialise with people outside of the cult. This will probably be met with social censure from the cult, making it difficult and painful. It is still possible. And outside friendships or activities can help a person to break free or resist ICURE techniques.

 All of this basically means you have a lot of options for your story because there are plenty of things you can weave in that would undermine ICURE.

 Your character is in prison, so breaking isolation is more difficult. But if the prison is overcrowded or there’s a sudden influx of people being transferred between facilities the character might end up with a… poor choice of cell mate from the guard’s perspective. Some one with beliefs radically opposed to the guards or someone who could support and shore up the character’s old beliefs.

 There may also be opportunities for covert communication and bonding within the prison. Perhaps prisoners can gather during breaks and have worked out a cant or code to talk about beliefs the guards are trying to stamp out.

 Control of information can break down because isolation has broken down, with prisoners trading information. It can also happen through the prisoner trading for a phone or a similar item allowing them to access forbidden information. Or it can happen through things like guards inadvertently giving out information.

 Uncertainty is difficult to create around core beliefs. The impression I get from anecdotal accounts is that pushing too hard at core beliefs too early often causes targets to withdraw from the people attempting ICURE. It can also lead to targets doubling down on their beliefs.

 People attempting ICURE can also mess up on creating uncertainty, as described above.

 Repetition can break down because guards don’t all do or say the same things consistently. They could contradict each other. Or they might just not repeat the same thing very often.

 Emotive responses can break down in much the same way creating uncertainty does. Not everyone responds emotionally to the same things or in the same way. Once again different guards can undermine the desired response. The character might dig in to their original position, they might withdraw from the people attempting ICURE. They might just learn to lie to them.

 I think as a writer the best approach to this is to use a mix of internal and external factors effecting multiple parts of ICURE. Just because I think that would create a better story.

 The readers can see the internal struggle and resistance in the character. They can also see the guards messing up and how that impacts the character. May be the importance of support from other people, fellow prisoners or cleaning staff or doctors or anyone else that fits with the setting.

 Basically including multiple elements will give you a more fleshed out story with more emotional depth and impact. That’s a good narrative reason to include it.

 I hope that helps. :)

Available on Wordpress.

Disclaimer

[mahek320]

Depression To Dementia?

With the Covid pandemic encompassing the entire world these past two years, it has become very clear that it has had many negative impacts on everyone's lives. Besides the very obvious ones,  one of the key effects of it has been a significant increase in cases of depression. In today's day and age, it seems like depression has almost been normalized. However, what many people do not realize is that depression often comes hand in hand with other mental illnesses, whether it be a symptom of them or simply accompany them. One of these mental illnesses may very well be dementia, a little nerve-wracking right?

But why? Well, there are a few possibilities as to how both depression and dementia may be linked. The first possibility is that depression is just simply a risk in developing dementia, just as certain foods and toxins create a risk of developing cancer. Another possibility is that depression is an actual symptom of dementia, pointing to its eventual onset. And the last possibility is that it is purely coincidental. To help support the thesis that depression may lead to dementia, a study was conducted over several years with Danish men with or without depression to truly understand if there really is a link between the two.

The way the study went about it was first to collect history of the patient's depression up until age 55. It was collected through primary and secondary diagnoses (ie. a diagnosis by a doctor vs. data on antidepressants). There were 3 main categories of depression within the patients: No history of depression, depression identified via medication, and depression identified via hospitalization. The severity was determined through the amount of the prescriptions along with the number of hospitalizations. After this process, the data on the men with dementia was collected. Again, the data was collected through primary and secondary diagnoses just as with depression. With all of this data, the study also identified potential confounders that could explain the link between depression and dementia, which they made sure to include in the analysis.

So what actually was the analysis and did it identify a true correlation? Well, four models of Cox proportional hazards regression, which is a “method for investigating the effect of several variables upon the time a specified event takes to happen”, in which the first model had no adjustments, the second model adjusted for social factors,, the third model adjusted for diseases, and the fourth model adjusted for addiction. Along with this, five supplementary analyses were conducted. The first took into account height and weight, the second was a sensitivity analysis which had more restrictions in reference to the temporality between measures of depression and dementia, the third only accounted to men with a psychiatric diagnosis, the fourth excluded men who were diagnosed/identified via medication, and the fifth used a propensity score to even the sample sizes of men with no depression vs men with hospitalization due to depression.

The results supported the link. Men without depression only had 1.2 cases of  dementia per 1000 people, 2.1 for men on medication, and 3.6 for men who had been hospitalized. So, this means that the risk of dementia was highest among men with the highest level of depression. In each of the four models with their adjustments, it was found that men who were hospitalized had higher cases of dementia within the population than men who were simply on anti-depressants, and both were higher than men without depression. However, there was one key difference between both of these situations. Although those who were hospitalized had a higher risk of developing dementia, the amount of times they were hospitalized had no significant link to it. On the other hand, those who took more prescribed medications did show a higher association to developing dementia. So, although hospitalizations are more closely linked with the possibility of developing dementia, the number of times had no significance, whereas, although people who were on medications were less likely to develop dementia, the amount of medications they took did affect the risk level.

So now all that is left to identify is why depression and dementia are actually linked. Again, there could be a few possible explanations as stated before. One reason could be that both depression and dementia share the same developmental traits (pathogenesis), such as inflammation. Because they share several of these traits, it is a strong possibility that they have a close link to each other. Along with this, another factor might be telomere length. A telomere is a DNA sequence located at the end of a chromosome and a study found that shorter telomeres may be a risk factor for developing depression as well as a significant risk of developing dementia as well. Besides these two possibilities, there may very well be different correlations between depression and dementia such as smoking, obesity, poor health, etc.

So where do we go from here? This study is only the beginning of our knowledge of the link between depression and dementia. Although it did not dive into it, it is also highly possible that depression could contribute to different subcategories of dementia, however, this requires conducting more studies and research. Similarly, it is also possible that different categories of depression have a different contribution to the risk of dementia within a person. For example, atypical depression has a close link to people with obesity, higher inflammation, and other health complications, therefore it could be predicted that this specific type of depression could lead to vascular dementia versus other forms of dementia. There is still a lot that is unknown about this topic as the correlation was discovered recently but as time goes on, more studies and research will help to make the how and why more clear. 

At the end of the day, it is absolutely true that depression is a very clear risk factor in the development of dementia. The details may be unknown, but an earlier diagnosis of depression may be key to reducing the risk of one getting dementia. There is plenty more research to be done but that does not mean action cannot be taken to help with these mental illnesses.

[mostlysignssomeportents]

America's pandemic spiral

More than 200k Americans have died of covid -  about 70 9/11s, with no end in sight. Indeed, things are getting worse, as the US enters a "Pandemic Spiral," as Ed Yong writes in The Atlantic. Yong identifies 9 factors driving the spiral:

https://www.theatlantic.com/health/archive/2020/09/pandemic-intuition-nightmare-spiral-winter/616204/

I. Serial Monogamy of Solutions: we only pay attention to one thing at a time: isolating, masks, plasma. Some of that is driven by Trump's short attention span and addiction to distraction tactics, but it's also science's methodological isolation of one variable at a time.

We especially struggle with "necessary but insufficient." Masks aren't effective - on their own. Neither is distancing. Neither is ventilation. All three? Pretty good, actually.

II. False Dichotomies: "We save lives or the economy," "It's like a flu, no it's like a plague." Actually, we CAN partially reopen the economy (most retail, with precautions, but not, say, nightclubs), and it IS a mild flu for some, and a death-sentence for others.

III. The Comfort of Theatricality: Hygiene theater (like sanitizing surfaces) provides the appearance of diligence and the comfort of DOING SOMETHING, but it distracts from taking steps that address the most recent science, like mitigating aerosol spread.

IV. Personal Blame Over Systemic Fixes: You can't recycle your way out of climate change, you can't shop your way out of monopoly capitalism, and your personal health strategies won't stop the systemic problems exacerbating the pandemic.

Without sick leave, workplace safety, child care  and transit, people will do things that put themselves and others at risk. Americans love to moralize, but they're terrible at systems thinking.

V. The Normality Trap: We want things back the way they were, and this can overpower our commonsense: we want to re-open tattoo parlors or movie theaters because that tells us it's finally over.

VI. Magical Thinking: Remember Trump's "Maybe this goes away with heat and light?" I confess that I get up every morning, make a cup of coffee and think, "Maybe today's the day this ends." It's impossible not to have these daydreams - but in America, they become policy.

VII. The Complacency of Inexperience: If you come from a privileged group with few cases and few comorbidities, you assume that if we just "let nature take its course," things won't be so bad.

Countries that have had recent experience with epidemics did SO MUCH BETTER than the US. That's why poor African countries - who survived ebola - are kicking America's ass when it comes to addressing the virus.

VIII. A Reactive Rut: We suck at understanding exponential growth, and this deficit is worsened by the time-gap between infection and symptoms, which makes it hard to emotionally grasp the connection between "superspreader" events and outbreaks weeks later.

This confounds our ability to do long-term planning, as we just keep expecting things will be OK in a month or two, and we don't need to (for example) figure out how schooling will work when the virus is still raging.

IX: The Habituation of Horror: Remember the movement to "not normalize Trump?" It was always doomed to fail. Human stimulus response always regresses to the mean - that is, if you're exposed to the same thing all the time, no matter how terrible, you get used to it.

Just ask children of abusive parents, or prisoners in solitary, or Auschwitz survivors. EVERYTHING becomes normal over time.

Yong: "The U.S. might stop treating the pandemic as the emergency that it is. Daily tragedy might become ambient noise. The desire for normality might render the unthinkable normal. Like poverty, racism, school shootings, police brutality, mass incarceration, sexual harassment, widespread extinctions and changing climate,  covid might become yet another unacceptable thing the US accepts."