Public speaking training often teaches us to remove every filler word from our speech. Podcasters frequently ask in podcasting communities whether to delete every "um" or leave them in.

Psychology offers a different perspective for us on removing all these filler words through the "Pratfall Effect". This concept suggests that highly competent people become more relatable and likable when they make small mistakes because perfection creates something like a wall between you and your audience.

In this micro-episode:

1. How over-editing can lead to this "Uncanny Valley" effect
2. Why filler words can serve as processing signals for your listener
3. The risk of sounding too much like AI-generated content, because it's 2026

Find more episodes and subscribe at stereoforest.com/minute.

About and Support

Written, edited, and hosted by Jen deHaan.

About Jen at https://jendehaan.com

Support the Show

• Like this episode or show and want more? Support us with a one-time tip: https://minute.captivate.fm
• We love our podcast host Capitvate.fm! Contact to ask me anything, anytime. You can support the shows by signing up with Captivate here: https://www.captivate.fm/signup?ref=yzjiytz
• We have our newsletters on Kit.com. We also have our tip form with them, and sell products on their platform. Easy, and they don't take a cut! Check Kit out and support the show using this: https://partners.kit.com/ijdkivtf8ndd
• Transcriptions by MacWhisper. I use and love the Pro version (subscription free!) - you can get it too using this link: https://gumroad.com/a/20303251/ivpqk
• Schedule posts? We use Metricool (reasonable for multiple accounts/brands/shows). Support us using our link: https://f.mtr.cool/VZBOZR

About Jen

Host: Jen deHaan has a background of almost 30 years in tech, education, & instructional design and 10 years in improv and performance.

Jen's website: https://jendehaan.com

Made and produced by Jen deHaan in British Columbia, Canada.

WEBVTT

So if you've taken Toastmasters or really any kind of formalized or organized public speaking lessons or training,

you've probably come across the idea of cutting out filler words. Like these are the ums or the pauses, anything along

those lines. And this question about what we do with them comes up all the time in podcasting communities. Do we delete

them? Do we leave them in? Is it natural? Is it too much?

So let's think about two effects, two concepts from psychology. The first one is called the pratfall effect. This is

when someone who's really accomplished or respected, they make a mistake in front of the group. Now after you've

established that authority kind of in a community, these sort of little mistakes are going to signal that you're

relatable. You're just like one of us. Audiences like that.

And audiences tend to not like perfection because it creates a lot of a distance or a wall between you and them. You're

not just like one of us. Now, it's really important with this concept to know that if the person has not established that

authority in a community, making mistakes is not going to be received the same way. It could be received in a negative

light. So you do need to think about that.

But if you are going out there and establishing that authority, you are probably someone who can be more relatable if

you do make mistakes. You're not perfect. You're not fully edited. So let's also think of this concept from psychology

called the uncanny valley effect. And this is where over editing a video or your audio can make it feel off. So uncanny

valley is when, like imagine a robot,

and the robot looks super human. It's really advanced, but you can tell it's not actually human. It's just something is

just ever so slightly off. That's the uncanny valley effect. So if you do cut up your audio a lot and the cadence goes off,

it could kind of have an effect a little bit like that. And audiences now are very sensitive to that because of all of the

AI

content that's out there. It's just slightly off and people are kind of knowing to look for that sort of thing and it

makes you feel a little bit uncomfortable because it's uncanny. Now I have seen comments out there from people saying

they're going to leave in filler words or add more filler words even more to make all those mistakes so people know it's

human and it's not AI generated. However, AI generation is adding those filler words and those mistakes into

AI generated content to make it really human. So your mileage might vary a little bit if you take that angle. So filler

words aren't all bad. Like um and ah, they actually serve as functions in conversation. Like they're signaling to your

listener that you are processing, say an idea, you're searching for that word, you're thinking in real time.

words are going to give your audience some processing time. So this is another advantage because if the subject matter

is difficult or that word is big, those pauses are going to help people digest what you just said before you move on. And

another thing it can do is it can signal to your listener that you are not finished your thought yet, so they shouldn't

jump in. So if you're doing an interview show or you have a co-host, it can be useful for that.

Now, your audience generally is not going to notice things like this. However, they are going to notice if you feel

anxious or they're going to notice if you're using too many filler words. And that might be because you're anxious.

That will shift the energy. It's not going to be a positive anymore. They might sense that you're not confident in the

material if you're using too many filler words. Something you can practice is pausing when you need to think and focus

on not adding

any filler words in those places. Because too many fillers can be distracting and then you don't need to cut them out. So

just work on finding that balance. Find out how many filler words or pauses represents your natural speech and then go

with that. I'm Jen DeHaan. This is the Credibility Minute. Find more episodes and get in touch with me and subscribe at

stereoforest.com slash minute.