When does bias begin? This is a question that is pretty difficult to answer, unfortunately, but we know that bias doesn't begin once your research is already underway. Bias very likely begins well in advance of this. That's what we're talking about in today's episode of Developer Tea. My name is Jonathan Cutrell and you're listening to Developer Tea. My goal on this show is to help driven developers like you connect to your career purpose and not only to connect to that at a personal level, but for that to affect the work that you do for you to do better work as a result of that connection. And that's what we're talking about in every episode and one way or another, we're getting back to that core purpose of this show. And we hope to do this in a quick format. That's the whole point of this show going live three times a week rather than having one long show and discussing this research bias in one episode. We want to discuss it in these more digestible formats. So hopefully that provides more insight to you. Hopefully, it allows you to listen to these episodes in smaller spaces in your day and day and day. So we're talking about research bias for all three episodes this week and really we're talking more about bias than we are just research bias, but this applies really directly to user research. When you're doing user experience research or customer research or whatever title you want to put on this, it's you understanding something about the thing that you're working on through the lens of another person. So we've already talked about some of the types of bias that can come up. For example, acquiescence bias. If you want to learn more about that, make sure you listen to the previous episode of Developer Teafrom Monday. But in today's episode, we're talking more about, you know, what is this first early bias? What are some of the early biases that we can fall prey to? And we're going to talk about two of them, but there are certainly much more and we may allude to a few of those as we go through this discussion today. But I want to take a moment and return to one of the kind of examples that we talked about in the last episode. We talked about this idea that a person can go to their friends, to their family, to people around them and ask them to give them their opinion about an idea or their opinion about a feature. And that based on this kind of quick research that a lot of times we end up making bad decisions because as we already mentioned, the acquiescence bias, people tend to, you know, agree with us. They don't want to create friction. They don't have a reason to create friction. And so a lot of the time they're going to agree with you, especially if you present this as if it's something that you believe is a good idea. So when does this actually start? When does the bias actually occur? We'll discuss one of the very earliest moments of this bias is kind of birth right after we talk about today is excellent sponsor, Linode. We're highlighting some of the things that Linode provides in each episode of Developer Tea for the last couple of episodes. And I want to take a moment to highlight Linode's high memory plans. So Linode has a plan that's going to suit pretty much every person who can listen to this show. So you are just starting out as a developer, of course, Linode's plan started $5 a month. Let's go through some of the numbers here. With $5 a month, you get a gigabyte of RAM. You get 20 gigabytes of SSD storage. You get a terabyte of transfer. You have an internal network that is 40 gigabits per second fast. That's incredibly fast. And the outgoing network is a thousand megabits per second. Again, that's $5 a month or $0.0075 an hour. That's something like.7.5 cents per hour that you're running the server. So that is the entry level plan that Linode provides. But they also have something on the totally opposite end. Their high memory plans. These plans are designed for much more memory intensive applications. So for example, you can get a 60 gigabyte of RAM. Let me say that again. 60 gigs of RAM on four cores. The storage for this plan is 90 gigabytes and that's SSD storage. The transfer for this plan, the transfer, the data transfer is seven terabytes. The internal network once again, you have the same internal network, 40 gigabits per second. But now you have an upgraded outgoing network, this is 3,000 megabits per second. This is only $240 a month. That's 36 cents an hour for that incredibly powerful box. All of their plans are hourly. So you can actually go and use this for a limited period of time. Go and check out what Linode has to offer. They're going to have something to offer you. By the way, that's not even their highest plan. They have a 200 gigabyte plan as well that gives you 16 cores. By the way, that one gives you 10,000 megabits per second outgoing. So again, go and check out what Linode has to offer to you. Just in terms of hardware and in terms of the numbers, they're going to blow you away. But well beyond that, we've talked about so many other features that Linode provides. We can't go through them all today. I encourage you to go and check it out. Instead of M-slash Linode, use the code Developer Tea 2018 at checkout and you'll get $20 worth of credit to use on any of Linode services. Respect out of M-slash Linode. Thank you again to Linode for sponsoring today's episode of DeveloperT. So when does this bias begin? It starts before you ever ask a question. Because you have to recognize that when you have an idea or when you have a plan or you want to validate something, you're already coming in with your own opinions. And this is kind of the genesis of your bias. Is that you have an opinion and you have a desire and unfortunately it's very difficult for you to believe or to accept that your opinion may be wrong. It's very difficult to act in a way that is unbiased towards your own opinion. So even unconsciously or even when you're trying to avoid this, you're going to prefer a setup that validates your opinion. This leads to confirmation bias in the end. But what it starts with before you begin your research, it starts with a few other types of biases. We're going to talk about a few of those today. The first one is selection bias. Imagine that you have a handful of friends and out of that handful of friends, you have one friend who is extremely frugal. They save all of their money. They cut coupons. They never buy anything brand new. They're always looking for second hand, even tech gear, their laptops or even their clothes. They're always looking to save a little bit of money. And then on the other end of the spectrum, you have a friend who is relatively liberal with how they spend their money. They're not really saving a lot. They're kind of carefree and they seek after experiences and they're willing to take on a credit card or two. Now imagine that you want to make a big purchase. Let's say something like a car. And you are trying to pick between a used car, an old car, and a brand new Tesla. You're considering buying a brand new Tesla. You know that you're going to go into debt and you're going to have to have a car loan for a little while if you buy the Tesla. But you really want to spring for it. If you want to gather opinions on the subject, you're much more likely to go to the person who would do that thing as well, who you predict would agree with you in the first place. You're much more likely to select them to gather an opinion what they think about this particular decision. This is kind of an active selective bias. This is you deciding that you want to go to this particular person, whether consciously or unconsciously because they're going to affirm your decision. They're going to affirm you in the direction that you hope for them to affirm you in. This is kind of confirmation bias. Before you get that confirmation bias, you're also kind of going through this active selective bias. But then there's kind of an inactive selective bias. This is really, if you've ever heard the term convenience sample, then this is what it's talking about. Selection of the people that you're going to talk to, most of the time we talk to people that we already have contact with. So we're going to talk to our friends or our coworkers or neighbors or someone that we can see, someone who is nearby in some way, whether spatially or relationally, someone who is nearby. And while for decisions like maybe buying a car, this may not necessarily have a huge effect, it will have a pretty large effect if you have a user base that is much more diverse than your convenience sample. So for example, if you run a company and you live in San Francisco and you're only testing your products with people who live in San Francisco, but you have a substantial user base that lives around the world, then you're probably going to run into some issues. And what can arise as a result of this is kind of dependent on what binds those people together in your convenience. In this case, if we're talking about San Francisco, then we know that as a general rule, San Francisco is going to have more early adopters than the average American city. So if you are going to test something out with early adopters, you're going to have a much different rate or a much different response, then you would, if you were to test it out with people who live in a rural area, somewhere much more distant away from a tech hub. So this is one example and you can think of a thousand other ways that convenience samples can cause issues, whether it's socioeconomic status, maybe it's something that binds them together culturally, maybe it's something that binds them together from a political standpoint or an educational standpoint. There's so many things that can cause that to go awry. And very often we discount that idea because we believe that our immediate surrounding group of people that the diversity they represent is enough to approximate a much larger group. So it's important to recognize both of these types of selective selection bias. The first one being obviously seeking out people who you feel are going to be amenable to your requests, but the second is only selecting people who are within your grasp, within your reach. So selection bias is something that is very difficult to defeat because in some way everyone has something in common with another person. And finding a truly diverse group of people can be very difficult. So in any kind of research you must recognize what pool of people are you pulling from? Where did they come from? How did you find them? What is consistent between them and what additional effect could that have on the research itself? What kind of bias that happens before you even start actually doing this research is called design bias. So I'll give you an exaggerated example of design bias so you can grasp what it is and then we can scale it back and see how it applies in user research. So let's imagine that you've created a podcast listening application and you have two people that are going to perform some kind of user research for you. The first one is really happy with their job and they love this application. They love listening to podcasts. The second one is unmotivated and they don't really care about the application. And in fact in some ways they wish it would fail so they could work on something different. So the first person decides to rent out a studio. And the studio is treated with acoustic materials so that everything is pristine and clear. And this researcher also goes out and buys the highest caliber of noise canceling headphones and they make sure that all of the phones that are running the application are the latest phones. They clean the phones and they even reach out to the people and ask them what their favorite podcasts are ahead of time before the research begins. A second person decides to run this research using an old phone that they found in the back of their closet and they grab their cheapest plastic headphones from the bottom of their old computer bag and they go out to a busy street where there's a lot of noise pollution and people kind of shoving past you and they get people to kind of stop as they're going from one place to another and they're already in a hurry and they ask them to listen to a test audio file that they created in Garage Band a year ago. Now obviously these are both very exaggerated examples. Hopefully it would be obvious to you as the researcher that one is going to come out with much different results than the other. But this kind of thing happens all the time. Anything experiments so that you eliminate bias as a result of, for example, the environment or as a result of the behavior of the researcher themselves, this is a very difficult thing to do. So if we return to our example of asking our friends and family questions about something that we want to pursue, an idea that we think is really good and we want some validation from our friends and family to move forward or not, it's very easy for us to word our questions, for example, in a way that kind of gives them no option other than to agree with us. We want to cast a positive light on this idea that we have rather than casting no light at all because it's easy to believe that even the smallest wavering of positivity could lead to a bad outcome, could lead to somebody saying that it's not really a good idea at all. And this is the kind of thing that we see in working environments all the time. And unfortunately, this often ends up being kind of the way that we go about making decisions. We have, for example, one person who's trying to get a raise, they may be a proponent for an idea that they don't really think is a good idea, but they push for it because they believe it's going to buy them relational capital. And unfortunately, again, this leads to a lot of money being wasted, a lot of bad decisions being made, and ultimately a lot of people who are not being honest with each other. So I'm encouraging you to, as you begin to seek out various types of research, even if it's very simple research, if you're trying to make a decision and you're getting someone else's opinion, if you're seeking this out to be aware of these biases, the ones that we talked about on the last episode, as well as the ones that we talked about in today's episode, the selection bias and the design bias. These are biases that happen at the very beginning before you even really start your research. Biases that happen when you begin to prepare for your research, at the very beginning of this process. Thank you so much for listening to today's episode of Developer Tea. I hope this has been challenging. I hope it's been convicting for some of you. And I hope it's been encouraging for all of you. Thank you so much for listening. Again, thank you to Leno for sponsoring today's episode. We couldn't do what we do on the show without our excellent sponsors. And Leno has been such an incredible sponsor. They create a great product for developers. I encourage you to go check it out. Spec it out of M-slash. Leno and use the code Developer Tea2018 at checkout. Thank you again for listening. And until next time, enjoy your tea.