In this video, I explain two well-known examples of observer bias: the case of Clever Hans, and Rosenthal and Fode’s experiment with “bright” and “dull” rats. Observer bias (or experimenter expectancy effect) is considered, along with an explanation for how to reduce this bias.
Don’t forget to subscribe to the channel to see future videos! Have questions or topics you’d like to see covered in a future video? Let me know by commenting or sending me an email!
Need more explanation? Check out my full psychology guide: Master Introductory Psychology: http://amzn.to/2eTqm5s
Video transcript:
Hi, I’m Michael Corayer and this is Psych Exam Review.
In this video I’d like to talk about observer bias, this is where the beliefs and expectations of an observer can influence the data that’s collected.
We’re going to start with an example from Germany in the early 20th century and this example actually involves a horse. This was a horse named Clever Hans and Hans was believed to be clever because people thought he could do arithmetic; he could add, subtract, multiply, divide work with fractions, he could also tell the time of day, he could tell the date on a calendar and he could even supposedly read and understand German.
Now Clever Hans was owned and trained by a guy named Wilhelm von Osten and here’s a picture of Wilhelm Von Osten with Clever Hans standing behind him and von Osten took Clever Hans all around Germany giving performances and demonstrations and many people were convinced that Hans really was this clever. In fact in 1904 13 people put together a Hans Commission to investigate and they determined that von Osten wasn’t using any trickery. They really thought than Hans was doing these things.
But another psychologist was a little more skeptical, this was a guy named Oskar Pfungst and Pfungst decided to do some investigating of his own and he did things like test Clever Hans with other questioners rather than just von Osten and he found that actually Clever Hans was still pretty good. He was able to answer questions if other people asked him. In fact, even if Pfungst himself asked, Hans was often able to get the correct answer.
So it didn’t seem that it was something specific Von Osten was doing. Then Pfungst thought about it more and he realized that there must be some way that Hans is getting the correct answer. So he ran some tests where the questioner didn’t know the answer. So Hans could see the numbers but the questioner couldn’t. What he found in this case was that Hans’ ability dropped off dramatically.
So when Von Osten knew the answer to a question that he asked, Hans was right about 89% of the time, if Von Osten asked a question but wasn’t able to know the answer in advance Hans was only correct 6% of the time.
So what was going on here? How was the answer being communicated? Well, what Pfungst figured out is that Clever Hans was picking up on unconscious cues. So the way that Hans answered was by tapping his hoof. And what Pfungst realized was that when people asked a question they would then wait and so Hans would start tapping his hoof and as he tapped tension built up, the questioner was sort of anticipating as he get closer and closer to the correct number of taps and this tension would reveal itself in the questioner’s posture. So Pfungst realized that they would ask questions and sort of lean back sort of anticipating as he got closer and closer when they got to the correct number of taps people would show this tension or they would even maybe sort of release as he did the final tap and Hans would notice this and stop tapping and therefore he did get the correct answer.
So it turns out that Clever Hans was not clever because he could do math or read, he was clever in that he was able to pick up on these unconscious cues.
I’d like to look at a second example and this is a study by Robert Rosenthal and Kermit Fode this was a study they published in 1963 and what Rosenthal and Fode did was they gave senior psychology students rats to run in some experiments. So they had 12 students and each student was given five rats to run in some tests and they were told that they had rats that were either
“bright” rats so some students were told “the rats you’re getting have been bred to be particularly good at solving mazes”
or they were told that they had maze “dull” rats, they said “the rats that you’re running well, not so smart, they’ve been bred from rats that are really bad at mazes. So these are maze dull rats.” And the students were then asked to run some experiments with the rats and collect their data. The experiments were pretty simple, the rats were placed in what’s called a T-maze.
Here’s our rat here. And the rat’s job is to run down this hallway and when he gets here he has to make a choice to go to left or to the right. So it looks like a T. And in this case one of the hallways was painted white at the end and the other hallway was painted gray. Of course they would switch which side was which. The rat’s job was to always run to the gray side. So he would come down here, he’d have to see which one was the gray side and hopefully go to that side. If he went to the gray side he got a reward if he went to the white side then he didn’t.
The students kept track of the data from their rats, so they kept track of how many times was the rat correct, did he choose the gray side or the white side and they also measured how long did it take him to make a choice to get to the end of the maze. And they did 10 times per day over the course of five days with all of these rats and it turns out that the students found that the maze bright rats were significantly better in both correct completions and in speed than the maze dull rats
Now this doesn’t sound too surprising but of course, as you might have guessed, there’s a twist in this study and the twist was that there were no maze bright or maze dull rats. In fact all of the rats came from the same breed and they were randomly divided and given to the students. So instead of bright or dull rats we actually have random rats.
And this means that the two groups of students should have gotten basically the same results for all the rats the correct completions and speed should have averaged out to be about the same between the two groups.
So why wasn’t it? Well Rosenthal and Fode said that this showed there was observer bias. It was the students that were causing the difference, not the rats. So we call this observer bias, or if you prefer alliteration you can call this the experimenter expectancy effect.
So this refers to the idea that the students had expectations they thought that the rats were either going to do well or they thought the rats were going to do poorly and this expectation ended up causing the result that they thought they going to have.
How did this happen? Is it that the rats self-esteem was being affected because their handlers thought that they were smart or dumb? Well, probably not. Rosenthal and Fod concluded that it was probably a number of very small unconscious cues. The students weren’t aware that they were doing this. What they were doing is they were treating the rats slightly differently.
So it might have been in their handling maybe they were a little bit gentler with these smart “prized” rats here, “we’ve got to be careful with these ones” and maybe they were a little bit rougher with the stupid rats, the dull rats, and maybe this treatment changed the stress levels of the rats and that influenced their performance. Or maybe when the students were measuring the time with their stopwatch maybe they were just a little bit faster with a smart rats or they were paying a little bit less attention with the dull rats. And these small differences that the students didn’t even realize that they were doing add up and give us a significant effect, a significant difference between these two groups.
This is why it’s important that we have some systems and some standards for how we collect data because we want to avoid these effects. We want to avoid any potential for observer bias.
And in fact, we’ve already seen the best way to avoid observer bias. And this is, make sure that the observer doesn’t have expectations. So in the case of Clever Hans this meant if the questioner doesn’t know the answer then the questioner can’t somehow communicate that answer to Hans and in the case of Rosenthal and Fode’s study if the students weren’t told that they had bright or dull rats then we probably wouldn’t have found a difference between these two groups.
So we want our observers to be “blind”, we want them to not have expectations about what’s happening. They shouldn’t know which group they’re testing. If we do this, it can help to reduce our observer bias.
Hope you found this helpful, if so please like the video and subscribe to the channel for more. Thanks for watching!
