Vox

Leitura de 23 min

Creating Consensus, Defeating Bias and Getting Better

Summary

This episode of Developer Tea focuses on the critical challenge of cognitive bias in professional work, particularly for developers. Host Jonathan Cottrell argues that biases distort our perception of statistics and data, leading to poor decisions that handicap our careers, relationships, and impact. He establishes that the key to defeating bias is not eliminating it entirely—which is impossible—but developing methods to work around it.

The core solution proposed is creating consensus through objective, replicable measurement processes. By establishing agreed-upon protocols for measurement that multiple people can follow independently to reach the same conclusion, teams can build trust in data and circumvent individual biases. This involves moving away from ego-driven interpretations and toward shared, rational methods of analysis.

Practical strategies include collaborating with people who think differently to balance out biases, creating decision trees before seeing data results to prevent post-hoc rationalization, and critically evaluating measurement methods rather than just the outcomes. The episode emphasizes that bias is a pre-wired human limitation affecting even experts, so the goal is awareness and systematic workarounds, not perfection.

Ultimately, defeating bias leads to more truthful self-assessment, better collaboration, and a stronger connection to one’s career purpose through constant improvement. By trusting rational processes over instinctual reactions, developers can make better decisions, build better products, and have a more positive influence on those around them.

Recommendations

Books

  • Thinking Fast and Slow by Daniel Kahneman — Recommended as the essential introductory book on bias and rationality. Described as a dense read that provides good reason to trust rational decision-making while showing how humans often fail at it due to cognitive biases.

Tools

  • WooCommerce — Sponsored tool presented as a trusted default for developers, powering 30% of online stores. Highlighted as open, hackable, and data-owning—not locking users in. Also mentioned as a marketplace where developers can sell plugins for extra income.
  • Mad Monk Tea — Personal recommendation for loose leaf tea, described as the best brewing process. Mentioned that tea can be re-steeped and a discount code (developer tea) offers 15% off first orders.

Topic Timeline

  • 00:00:00 — Introduction to defeating bias through consensus — Jonathan introduces the episode’s core thesis: the key to defeating bias is developing consensus. He explains that biases change how we view statistics, and the episode will explore methods to create baseline measurements, build trust in data, and work around cognitive limitations. He connects this to the podcast’s purpose of helping driven developers do better work and find career purpose.
  • 00:03:00 — The problem with data perception and misuse — The discussion addresses why data and statistics are often misunderstood or misapplied, especially in small companies or individual contexts. There’s a perception gap where data is seen as separate from useful information. Jonathan argues we need to develop a better understanding that statistics themselves don’t have agendas—only people do—and we must examine the people behind data as closely as the numbers.
  • 00:06:00 — Practical consensus through objective measurement — Jonathan makes the concept practical for developers: you don’t need to know all research methods, but you do need to establish objective measurement processes that others can replicate to reach the same conclusion. This baseline consensus creates reliable, trusted information. He compares it to a protocol that everyone agrees upon, which builds confidence and shared understanding within teams.
  • 00:08:28 — Recommended reading: Thinking Fast and Slow — Jonathan recommends Daniel Kahneman’s ‘Thinking Fast and Slow’ as essential reading on bias and rationality. He acknowledges it’s a dense book but worth it, noting that Derek Sivers might have a summary. The book provides reasons to trust rational decision-making while vividly illustrating how humans frequently fail at it due to inherent biases discussed in previous episodes.
  • 00:12:20 — Accepting and observing bias through mindfulness — The conversation shifts to the reality that everyone operates with bias, even experts like Kahneman. The first step is recognizing and accepting this limitation. Jonathan suggests practices like meditation to observe one’s thoughts from a distance, allowing you to analyze whether you’re being affected by bias in the moment. This meta-cognitive awareness is crucial for taking action.
  • 00:14:54 — Collaboration and decision trees to circumvent bias — Two key strategies are presented: collaborating with people who think differently to balance biases, and creating decision trees before seeing data results. By pre-committing to actions based on data outcomes (if-this-then-that), you avoid negotiating with or explaining away data that contradicts your preferences. This structured approach prioritizes method evaluation over result judgment.
  • 00:19:11 — Applying lessons to developer career improvement — Jonathan brings the discussion back to practical developer work, using examples like tracking how long projects take to inform future estimates. This process of using past information to predict and improve future performance is framed as fundamental to career growth. By being more truthful with oneself and others about cognitive limitations, developers can enter a constant improvement cycle that fuels purpose.

Episode Info

  • Podcast: Developer Tea
  • Author: Jonathan Cutrell
  • Category: Technology Business Careers Society & Culture
  • Published: 2018-04-06T14:00:00Z
  • Duration: 00:21:41

References

Podcast Info

Transcript

[00:00:00] So the key to defeating bias is to develop a consensus. This is what we’ve established.

[00:00:12] And biases are going to change the way that you view statistics. So we’re going to talk about

[00:00:18] how to develop that consensus, how to create that baseline measurement expectation and those

[00:00:25] methods. And we’re going to talk about ways that you can ultimately begin to trust those statistics

[00:00:32] and defeat or at least work around some of your biases. My name is Jonathan Cottrell and you’re

[00:00:38] listening to Developer Tea. My goal in the show is to help you, hopefully you’re a driven developer,

[00:00:44] help you as a driven developer connect to your career purpose and do better work so you can have

[00:00:51] a positive influence on the people around you. And this episode is dedicated to you.

[00:00:55] It’s dedicated to doing better work. That’s the side of the purpose of this podcast that we’re

[00:01:03] focusing on when we talk about bias. But the other kind of side effect of this is that as you begin

[00:01:09] to work in ways that are seeking truth above ego, as you begin to work in ways that allow you to

[00:01:19] see your bias for the first time, maybe ever, and defeat it,

[00:01:25] work beyond that ego, work beyond that bias, you start to uncover a better sense of purpose.

[00:01:35] Because here’s the reality. Most people are handicapping their work, they’re handicapping

[00:01:41] their careers, their relationships by allowing bias and allowing their ego to get in the way.

[00:01:48] We would be much happier, we would be much more focused, much more productive,

[00:01:55] and ultimately we would work better together collaboratively we would build better things

[00:02:02] we’d have a better impact on the world if we could recognize our faults first if we recognize

[00:02:10] our ego if we recognize our bias and then we do something about it so this this show is about

[00:02:19] doing better work this episode is about doing better work but it’s also in a in a kind of a

[00:02:24] more fundamental way this is going to help you seek after a more purpose-driven career as well

[00:02:31] so that’s the reason we talk about these kinds of things about uncovering and fixing

[00:02:39] or at least being aware and finding a way to work around these failures of being a human

[00:02:47] and we’ve already established why why bias is important why it can have a negative effect

[00:02:53] on the work you do

[00:02:54] do. We’ve established that statistics and data and all of these things that have been traditionally

[00:03:00] kind of seen as cold or calculated measures and often are not really respected in a small

[00:03:09] data scenario, right? In a small company or at an individual level, these things are discarded

[00:03:16] quickly or otherwise misused or misunderstood. They aren’t applied appropriately. They aren’t

[00:03:25] usually learned from correctly because we have this perception that data is something different

[00:03:31] from information. Data is something different from a description or a measurement, right? So

[00:03:40] we need to unravel this, first of all, in ourselves. We need to develop

[00:03:46] data. We need to develop data. We need to develop data. We need to develop data. We need to develop

[00:03:46] a better understanding of how statistics matter, how we can think rationally better with

[00:03:55] statistics, and how statistics don’t really have an agenda. Only people have agendas. And so

[00:04:03] when we’re trying to understand how statistics apply or if we believe that something is skewed,

[00:04:10] we need to be looking at the people as closely, at least, as we are looking,

[00:04:16] at the data itself as we’re looking at the numbers. Now, we have learned about some methods,

[00:04:22] some ways of eliminating variability or eliminating skews in your data, right? So,

[00:04:30] for example, redoing a given experiment so that you have, essentially, the more times that you try

[00:04:42] an experiment and get the same result, the more…

[00:04:45] or authority that result has. And this is called replication. There are tons of research methods

[00:04:53] and specific ways of going about research, more in the academic field, that can help eliminate

[00:05:02] bias as well. Of course, things like double-blind studies and having a good control set and having

[00:05:09] a significant enough sample of data in order to analyze it properly,

[00:05:14] having learned… if you’re doing some kind of machine learning to analyze your information,

[00:05:20] then you’ll have different segments of data. For example, your training set versus your test set

[00:05:29] versus the holdout. And all of those things have different meanings. We aren’t really going to cover

[00:05:33] all of this on the show, but suffice it to say that data and statistics and all of these things

[00:05:39] that we collect, that we measure, we need to do so carefully. And as we’re looking at the data, we’re

[00:05:44] going to be looking at the data, we’re going to be looking at the data, we’re going to be looking at the data,

[00:05:44] as we go about collecting information in a way that we can use it rationally, as we do that,

[00:05:52] if we are practicing these more careful behaviors, especially if we know why those behaviors are in

[00:06:00] place, then we’re much more likely to trust those statistics. And in order to make this practical

[00:06:07] for you as a developer, you don’t have to go and read about all of the various research methods

[00:06:12] that eliminate variability or eliminate bias. You don’t have to go and read about all of the various

[00:06:14] research methods that eliminate bias or eliminate bias or skew. You don’t have to know about

[00:06:17] everything. What you do need to understand is a way to measure that you and someone else can both

[00:06:25] come to the same conclusion, both come to the same measurement when looking at a given subject,

[00:06:32] right? So you need to be able to measure something with an objective process that

[00:06:37] someone else can also measure that same thing with the same process, having not seen your measurement

[00:06:44] and arrive at roughly the same conclusion. This baseline consensus, right? This is you developing

[00:06:53] a consensus that allows you to have a reliable measurement. By doing this in this particular

[00:07:01] manner, you’re going to trust the information, but beyond you, the other people you work with

[00:07:07] and collaborate with, they will trust it as well because the method is shared. The method is agreed

[00:07:13] upon. It’s not just a matter of how you do it. It’s a matter of how you do it. It’s a matter of how you do it.

[00:07:14] It’s a matter of how you do it. It’s a matter of how you do it. It’s a matter of how you do it.

[00:07:14] In a way, it’s like a protocol, right? You have a way of doing this particular measurement and

[00:07:21] it’s something that everyone agrees upon. In addition to having this kind of defined process

[00:07:28] of measurement to increase your confidence, you also need to go through the process of understanding

[00:07:37] statistics and why they matter. And that’s kind of what we’ve been doing on the last couple of

[00:07:42] episodes, try to introduce you to the process of understanding statistics and why they matter.

[00:07:44] Some of these pieces of information about data and about statistics and about measurement

[00:07:51] and about quantification so that people can parse them and hopefully come to better conclusions

[00:07:57] about what this stuff means to them, right? But it does help for you to develop a trust

[00:08:05] in that rationality. And this may take looking through information for yourself. It may take

[00:08:13] reading a book. It may take reading a book. It may take reading a book. It may take reading a book.

[00:08:14] It may take reading a book. It may take reading a book. It may take reading a book. It may take

[00:08:14] case studies, reading books about data-driven decision-making, for example. These are all

[00:08:20] useful ideas. Certainly the book that we always recommend when discussing bias and rationality

[00:08:28] and statistics is Thinking Fast and Slow by Daniel Kahneman. This is kind of the introductory book.

[00:08:33] It’s a dense read. At the very least, go and read. Maybe Derek Sivers has a summary of this book,

[00:08:43] but it is certainly worth the read. But essentially what you’ll find in this book

[00:08:48] is good reason to trust rational decision-making. And you’ll also, and perhaps more importantly

[00:08:57] and more pronounced, you’ll find that humans very often fail at making rational decisions.

[00:09:05] This is because of the thing we discussed in the last episode, the biases that we fall prey to.

[00:09:11] So we’re going to talk about,

[00:09:13] today’s sponsor, and then we’re going to talk about ways that you can circumvent some of these

[00:09:17] biases. Today’s episode is sponsored by WooCommerce. You need good defaults. This is

[00:09:25] something every developer needs. It’s something we’ve talked about on the show before. Things

[00:09:30] that you can trust, rather than having to pick your tools for every single project,

[00:09:36] and rather than having to go and vet the code yourself, you need tools that you can trust that

[00:09:42] have been tested by thousands of other people. These are useful defaults. And WooCommerce is

[00:09:50] exactly that. In fact, if you’ve ever shopped online more than about three times, you’ve probably

[00:09:56] used WooCommerce because WooCommerce powers roughly 30% of all online stores. WooCommerce is built on

[00:10:04] top of WordPress, which is an open platform. And this also means that it’s totally hackable.

[00:10:10] If you have the chops,

[00:10:12] you have the skills, you can go in and customize WooCommerce and integrate with anything that you

[00:10:18] know how to integrate with. Because again, it’s totally open for your control. You also own your

[00:10:24] data forever. That’s a unique thing about WooCommerce. You own your data. If you want to

[00:10:30] take it and go elsewhere, if you’re totally unhappy for whatever reason, WooCommerce is

[00:10:34] not going to lock you in and try to keep you as a customer. They want you to be happy with their

[00:10:39] product. And so they’re going to give you your data.

[00:10:42] Now, let’s say you’re just a programmer and you don’t really need to create an online store,

[00:10:50] but you’re still interested in hacking e-commerce stuff. Well, as it turns out,

[00:10:56] WooCommerce has a marketplace that you can develop for and you can sell your plugin.

[00:11:03] This is a great way to make extra money as a developer on the side.

[00:11:07] Go and check out what WooCommerce has to offer to you as a developer. Head over to

[00:11:11] WooCommerce.com.

[00:11:12] slash developer T to get started today. That’s WooCommerce.com slash developer T.

[00:11:18] Thank you again to WooCommerce for sponsoring today’s episode.

[00:11:22] By the way, just before this episode, I had some excellent tea. This was a slightly sweet tea,

[00:11:30] and it was made by this incredible company called Mad Monk Tea. Now, if you’ve never had

[00:11:36] loose leaf tea, you’ve heard about it on the show, hopefully before. This is the way that tea

[00:11:42] is meant to be consumed. It’s the best brewing process for tea that I’ve ever had.

[00:11:49] Another cool thing about this is that the tea itself, when you make a little pot of tea with

[00:11:55] loose leaf tea, you can re-steep it again. Mad Monk Tea is one of my picks. In fact,

[00:12:01] they’re the only pick that I have as of right now. I encourage you to try it out if you’ve

[00:12:06] been interested in loose leaf tea before. Head over to MadMonkTea.com. If you use the code to

[00:12:12] developer tea, you’ll get 15% off your first order. We’re talking about how to defeat bias

[00:12:20] on today’s episode. This is a huge topic, and it’s something that researchers spend their entire

[00:12:27] careers focusing on, even single biases. We’re not going to cover all of the information that

[00:12:34] you need to know in order to defeat every bias and become a mental superpower wielder. That’s

[00:12:41] not going to happen.

[00:12:43] What can happen is you can start to change the way that you interface with your own thoughts.

[00:12:52] What does that mean? Well, if you’ve practiced meditation, for example, you know that you can

[00:12:58] observe your mind working. It’s kind of a strange phenomenon, but you do not necessarily have to

[00:13:07] get caught up in the thoughts that you have or even in the beliefs that you have.

[00:13:12] You can take a step back and watch those thoughts occur. And to our current conversation,

[00:13:20] you can take a step back and you can analyze if maybe right now you are being affected by a bias.

[00:13:29] And the reality is that we’re all affected by biases, no matter how aware we are of those

[00:13:35] biases. Because this is not something that is a behavioral challenge. It’s not a habitual challenge.

[00:13:42] This is something that is kind of pre-wired into our brains and different biases for different

[00:13:48] people. But even the most educated and most aware individuals, even the most awarded economists,

[00:13:58] even Daniel Kahneman himself, who wrote the book on this subject,

[00:14:03] accepts the reality that they’re going to operate under some kind of biased mentality,

[00:14:10] under some kind of bias mentality. And so, I think that’s going to be a big challenge.

[00:14:11] Some kind of biased behavior, right? And so, what we have to do is first recognize that that

[00:14:17] exists. First recognize that we’re going to operate with our bias intact. And sometimes,

[00:14:25] we won’t even be able to see it. In fact, perhaps most often, we won’t even be able to see it.

[00:14:30] So, it’s very important that, first of all, we recognize so that we can accept this reality.

[00:14:35] Because if we accept the reality, then we can take action steps to find ways of working around it.

[00:14:41] If we don’t accept it, then we’re very unlikely to take any kind of action to circumvent any of our

[00:14:48] biases. So, one action that you could take, for example, and this is a very common way of dealing

[00:14:54] with bias, is to have other people who think differently than you to collaborate with.

[00:15:00] This is kind of a way of evening out your biases. Other people are not going to have less or

[00:15:06] necessarily more biases than you, but they’re going to see

[00:15:11] your interactions differently than they see their own. They’re going to see

[00:15:15] your statements, your information, your beliefs. They’re going to see it differently. In fact,

[00:15:20] they’re going to observe the entire world differently than you observe it.

[00:15:24] And so, when you react out of bias, perhaps their bias wasn’t necessarily triggered.

[00:15:29] Part of this is because bias is very deeply a part of our own experiences and often is informed

[00:15:36] by things like trying to protect ourselves from loss. So, if I’m acting,

[00:15:41] in a biased way, because I’m trying to protect myself from personal loss, then perhaps another

[00:15:47] person who doesn’t have the threat of the same loss can watch that behavior and identify it.

[00:15:55] Now, this requires an immense amount of trust between you and the people that you collaborate

[00:15:59] with on a day-to-day basis. It also requires transparency and, as we’ve already said before,

[00:16:05] identifying that your ego has no place in your work. Lastly, it’s very important that,

[00:16:11] as you begin to use data or measured information to make better choices and make better decisions

[00:16:20] together in your work or even in your personal life, as you begin to use this information to

[00:16:26] make more rational, more informed decisions, it’s important to remember that your bias is

[00:16:33] going to try to change the way you view that information. You are more likely to discredit

[00:16:41] information, for example, that you didn’t already agree with, and you’re more likely to

[00:16:47] really latch on to information that you already did agree with. When possible, make your decisions

[00:16:55] before you know the information that will determine which route you take. In other words,

[00:17:03] create a decision tree. Sometimes it’s only two branches large, but an if-this-then-that

[00:17:11] kind of thing.

[00:17:11] So if you have data that leads you in one direction, then you know you’re going to take

[00:17:18] that direction. And if you have data that leads you in another direction, then you know you’re

[00:17:22] going to take that direction. Create this more structured way of making decisions rather than

[00:17:30] negotiating with the data. This is something that happens so often with data-driven decision

[00:17:38] making, is negotiating with the data and calling that analysis.

[00:17:41] Trying to explain the data rather than allowing the data to explain itself.

[00:17:49] So determine before seeing the results or before seeing any kind of quantification,

[00:17:55] determine how much you trust the information that you’ve received,

[00:17:59] not based on the results, but rather based on the methods used.

[00:18:04] Determine if you believe that the methods used were well thought out, and determine if they

[00:18:10] are based on the results. And if you believe that the methods used were well thought out,

[00:18:11] based on a consensus, determine if they are replicable, determine if your sample size was

[00:18:17] large enough. And hopefully what this will allow you to do is create a more informed way of

[00:18:24] believing before you see the results, rather than judging the results of that information

[00:18:30] ahead of time. Now, as researchers have found out, sometimes this even is too difficult to do

[00:18:38] because how do you evaluate just how good the results are? And how do you evaluate the results?

[00:18:41] Once again, back in the world of academia, we’ve found ways of deciding just how good

[00:18:49] data is. For example, correlation scores, things like p-value, and those kinds of things. But

[00:18:57] for most developers and most data-driven decision-making scenarios, you’re not going

[00:19:03] to be dealing with that level of information. Really, what you’re going to be dealing with

[00:19:07] is very simple things. Like, for example, a moving average.

[00:19:11] How long did it take you to complete project X? So that when project Y comes along, and it looks

[00:19:18] a whole lot like project X, you can use information, things that you learned in the process

[00:19:24] of completing project X, to inform how you’re going to behave during project Y.

[00:19:31] This is very important to do. This is really kind of the fundamental core tenets of learning,

[00:19:36] taking past information and using it in order to predict

[00:19:40] future. This is such a key part of your career as a developer, and it’s how you’re going to

[00:19:47] improve, become better as a human. Thank you so much for listening to today’s episode. I hope you

[00:19:54] will take all of this to heart as a developer, because these kinds of things are going to make

[00:20:00] or break your effectiveness. But even more than that, if you can find a way to be more truthful

[00:20:07] with yourself, with your co-workers, with your colleagues, with your colleagues, with your

[00:20:10] friends, be more honest about what you think, and to be more truthful to your brain about how

[00:20:18] wrong it can be, you’re going to see your career kind of light up. The purpose-driven kind of

[00:20:26] undercurrent, that fuel that we talk about, you’re going to find more of that because you’re going to

[00:20:32] click into this constant improvement cycle. Thanks again for listening. Thank you again to

[00:20:38] WooCommerce for sponsoring today’s episode.

[00:20:41] Head over to WooCommerce.com slash developer T to get started today. If you have enjoyed what

[00:20:47] you heard on today’s episode, then I can almost guarantee you that you’re going to enjoy future

[00:20:51] episodes, because this is kind of a bread and butter talk, the types of information that we

[00:20:58] share on the show all the time. Go ahead and subscribe in whatever podcasting app you use.

[00:21:03] Now, if you have found developer T to be valuable, then I request something from you. I don’t request

[00:21:10] donation. I don’t request donation. I don’t request donation. I don’t request donation. I don’t request

[00:21:10] donations. I don’t request any of those things. Of course, going in and engaging with our sponsors

[00:21:15] is always a good thing. But beyond any of that, what I request of developer T listeners is that

[00:21:21] you take just a few minutes to leave a review and a rating in iTunes. This is the best way to help

[00:21:28] other developers find developer T. Thank you so much for listening. And until next time, enjoy your

[00:21:34] tea.

[00:21:40] Transcription by CastingWords

Visão de gráfico