D4 Master Partial Interval, Whole Interval, & Momentary Time Sampling

Hi everyone and welcome back. We're

going to talk about discontinuous

measurement. And so this

is different than continuous in that

we're sort of sampling the behavior.

We're not seeing all instances of

behavior. So what is it? It's measuring

a couple instances of behavior. If you

break the observation into intervals or

moments in time, we want to use this

when continuous measurement is not

practical and you might have time

constraints. I often ask teachers if

they can try this because it's much

easier to do other things while you do

this. Initially, it's pretty difficult,

I would say, but once you learn how to

do it well, it becomes a lot easier.

So this is when you have limited time or

attention for observation, multiple

clients at once, behavior occurs at a

high frequency. So it would be difficult

to take a frequency count. So we use

discontinuous or interval recording. And

so we have three types and I'm going to

talk about each type

individually. This is an interval

recording sheet that I created.

The first thing you need to do is pick

your time intervals. And this is going

to be individualized to the

student or the behavior that you're

working with. And these cannot change.

So when you're in baseline from

intervention, you cannot change this.

You must collect the data using the same

type of interval recording and the same

intervals. This specific example has

10second intervals. You can do 30

seconds, you can do 20 seconds, you can

do a minute, you can do five minutes. It

just really depends. The smaller they

are, the better the picture of behavior.

The bigger they are, the easier it is to

do and the less data you're collecting.

You're going to look at that interval

from from the from the 0 to 10 seconds

and you're going to mark did behavior

occur or did it not? And there's some

rules with that with partial, whole, and

momentary time sampling. And we'll talk

about each one of those

individually. But you mark down and then

once you mark it down, you're done. And

you wait for the next interval. And from

10 to 20 or 19, did a behavior occur?

And so you can set your watch so it

every 10 seconds it buzzes. Or you can

do different things with timers to help

you. Or you can just take a watch or

start my stopwatch and put it next to my

data sheet. So I'm just watching the

time so I constantly know when we're

moving into a new interval. So 10

seconds is really the smallest anyone

ever does. I've never seen someone do

five. So 10 is really small, very time

inensive, almost as good as taking

continuous measurement. And then I've

seen I've done 10 minute intervals

before. So you can make the intervals

really big if you need things to be

easier. So most data sheets for this

interval recording is going to be long w

like it goes this way and a length

landscape. So I you can use either

nothing changes about this but

it is just what you're going to see if

someone hands you a data sheet. So you

write down from 0 to 10 seconds did it

occur from 10 to 20 did it occur? And

all you're marking is a yes or no.

You're not marking anything else. You're

not marking a frequency. You're not

marking a duration. You're not talking

about latencies. Just yes or no, X or

check, zero or one, Y or N. People use

different symbols to do

this. And so that's part of the reason

why it's a lot

easier. So why we use this? We use this

because we're we only want to measure

some instances of behavior. So, we're

not measuring all instances, but it is

just a lot easier. It breaks our

observations into moments or time

segments. We're looking at whether

behavior occurred within that small time

interval or not. This is used when

continuous measurement is not practical.

Really high frequency behavior, you have

to do other things. You want to observe

many people at once. A lot of times I'll

introduce a partial for therapy. So when

the behavior techs are doing therapy,

they'll also take interval recording

every 5 minutes. It helps us track

behavior throughout the session, but

isn't so time inensive to do frequency

or duration on a behavior. If it's

frequent, if it's not frequent, they can

probably do that during therapy tech.

But if it's frequent behavior, it's

almost impossible to run therapy and

take the data at the same time. So then

we introduce a five minute interval

system for our data

collection. We have three different

types, partial, whole, and momentary

time

sampling. So we use it when there's

limited time or attention for

observation, when there's multiple

clients at once. So you need to when

behaviors occur at a high frequency, we

need a general trend. the exact count is

not as important. Partial interval

recording is the one you'll use most

often. Most people are taking partial

when they say, "I'm doing an interval

recording." Okay. So, you get to mark

yes if the behavior occurs at any point

in the interval. It doesn't matter if

it's a second, if it's the whole

interval, if it's half the interval.

It's always going to be yes if you see

the behavior. So the second you're in

your interval, let's just say we have

these five minute intervals and within 3

seconds you see behavior, you mark yes

and you're essentially done for the rest

of the interval. You don't need to

record anything cuz it's already a yes.

Whether they behave more or not isn't

going to change anything in that given

situation.

What partial unfortunately does is it

overestimates the frequency of the

behavior because you're marking yes in

this interval even if you see 1 second

of behavior. You get more yeses than you

would get nos and so it's an

overestimation of behavior. Keep that in

mind when you're choosing to use it.

It's really good for monitoring high

frequency behaviors to decrease. When

you have high frequency behaviors, you

probably want to mark partial. And so

you mark yes during partial. If you see

behavior at any point in the interval.

If you have a whole interval where you

see zero behavior, then you would mark a

no or an X. Check for yes, X for no.

Sometimes it's zero for no, and one for

yes. And sometimes it's just yes or no.

The strengths and weaknesses of this.

It's used for decreasing high frequency

behaviors. Interval recording is

efficient for capturing high frequency

behaviors like hitting or shouting. It

captures all occurrences within an

interval. Even if there's multiple

behaviors happening in one interval, the

interval recording will mark as one

occurrence. So that's how the

overestimation occurs and it's less

precise. If you took frequency, you

would capture that. It only happened

once, not within the 5 minute period.

You lose that data of once in a five

minute period with this. So that's how

it's overestimating and that's how

you're losing data. That's why this is

discontinuous. Well, just doing a

frequency count would be considered

continuous. So here's some good

behaviors to use for partial interval.

It's really good to take tantrums can

really work on this one. Self-injury,

aggression, property destruction,

disruptive vocalizations, those can be

very high frequency behaviors. Whole

interval recording. So whole interval,

it's the same data sheet. Those data

sheets I showed you, you could use it

for any of these. You divide your

intervals into some sort of standard

time period. 10 seconds, 20 seconds, 30

seconds, 1 minute, 5 minutes, 10

minutes.

And then you're going to mark yes if the

behavior occurs through the entire

interval. If the behavior stops at any

moment, it's a no. This underestimates

behavior. An example of this, if the

behavior occurs through the whole

interval, you mark yes. And if it stops,

so if you're watching the behavior for a

couple, it's a 5-minute interval and you

the behavior is going going and you're

waiting, is this hole? Is this a yes? Is

this a yes? and then it stops for 1

second, you write no and you're done.

Even if the behavior starts up again,

you're still done. So that's how it

could save time with this one. You want

to mark yes if it occurs the whole

interval without stopping. And then if

it stops at any moment or if you see no

behavior, it's a no. The second you see

the behavior stop, you're done with that

interval. Essentially, if you're, you

know, at the very start of the interval,

it's a no. even if the behavior starts 2

seconds

after. So that's how it underestimates

behavior. So you're even though you

might most of the interval they engaged

in behavior and the two last seconds

they did not engage in behavior, you're

going to mark that as a no, it's an

underestimation of how much the behavior

is occurring.

So whole

interval strengths and weaknesses when

you want to increase behaviors it's the

off can be the better choice. It's

really good for behaviors you want to

see very long like staying on task. It

shows sustained engagement. It can show

participation in positive behaviors.

It's a focus on the duration. Partial is

more of a focus on frequency and it's

suited for really long behaviors.

So you might these are things that you

might want to use whole for sitting in

sitting in seat during class. So staying

focused on the assigned work on task

behavior. Use this as on task behavior.

Group participation staying attentant

and participating in group activities.

And then walking in line appropriately

lining up and walking with the class in

a quiet and orderly manner. Momentary

time sampling is our last one. It's my

favorite. It is also the least precise,

but it is my favorite. This one's the

easiest for me. When I do school

psychology assessment, part of the law

for assessment is that no matter what,

even if you're looking at a learning

disability, you still have to observe

the child. Learning disabilities, we

determine a kid has a learning

disability through mainly large scale

uh measures. Normative testing is when

you're comparing them to a group

representative of the whole United

States. So you look at their processing,

their cognitive ability, different

things, and you say yes, their

processing is low enough to qualify for

a learning disability. So though the

classroom observation is really

important for

any child in psychoed assessment because

it helps us develop goals and shows us

where they're functioning in the

classroom. it's not crucial to

eligibility for a learning disability,

but so I use this because this gives me

the best information about the child and

I usually do on task behavior. So it's

kind of allows me to rule out ADHD

because ADHD and learning disabilities

are comorbid. They often are tied

together. But you can use it for lots of

things. I also use it when I need group

data. You pick a certain time within the

interval and you don't change that. For

example, if you're doing a 10 second

interval, you might choose the last two

seconds. It gets that seconds gets

bigger as an interval gets bigger. So 10

seconds, I would look at the last 2

seconds of the 10-second interval. For

if I was looking at 30 seconds, I would

look at 3 seconds. If I was looking at a

minute, I might look at the last 4

seconds. So you look at the set time.

So, you don't have to do anything until

that set time. So, you're watching your

clock and you get down to you're doing

10-second interval between um 7 and 9 or

8 and 10. It just depends on how you're

going to write your intervals. You look

up at the child or the person you're

observing. If they're engaged in the

behavior, it's a yes. If they're not

engaged, it's a no. So, you do not have

to look at them for the rest of the

intervals. When I need teachers to take

data, this is what I give them. And I

give them like a 5-minute interval. So

every 5 minutes their phone goes off and

they look at the child and mark yes or

no for behavior occurring. It's very

easy for them to do an hour a day or so

and it gives me a sampling of what's

going on. Mark yes if the behavior is

happening and no if it's not within that

um specific moment. This can under or

overestimate. It just depends. You're

like sampling the smallest amount of

time for whether behavior is occurring.

So, it's the least precise of all the

different types of data collection we're

talking about. This is so easy to use

while multitasking. I use I often for on

task, off task, there's a whole app

called Boss that has this all built in

and then it makes you a fun little chart

and stuff. And they also have sensory

behaviors in there. It's a really cool

app that I highly encourage you to look

at. It does cost I think like $25, but

that's lifetime. So you get it forever.

It's so much less time consuming

whenever I have a lot of stuff going on.

Doesn't require watching the entire

interval. So you can do other things

during the intervals. You know, with a

10-second interval, it is easier than

doing other types. But um you can't do a

lot. Oh, you can take ABC data really

easily. Even if you have a 10-second

interval, if you have a long interval,

you can definitely do lots of other

things while waiting for your timer to

go off. Works well in really busy

environments where there's lots of

people, lots of things going on. And

then the best part about this is you can

do group. So, especially with research,

this is how I take group data. I do a

lot of on task behavior for a full

class. I've done up to 15 kids. It takes

some time to be that good at this, but

essentially just give yourself two

seconds for each child. And so if you

want to look at 10 10 kids, then make a

20 second interval. So you just label

them however they're sitting. I try to

put them in that row in that space. You

can probably do five if they're moving

around. You just go like student A, like

Emily. Emily, is she engaging or not?

Yes or no? And then you move to Tom. Is

he engaging or not? Yes or no? Every two

seconds you take data on each student.

Then it recycles. Is Emily engaging in

behavior? Is Tom? It's perfect for

looking at onas behavior for a whole

class. And so that's how you do group.

You could do if they're moving around

during recess, you could probably

do like maybe five or four, but you have

to go find the kid. You have to track

the kid with your eyes each time. So if

they're moving around, you need to look

at less kids.

So, you want to use it for on task work

during independent work. Participation

in group activities is great. Engagement

with the materials is great when they're

using iPads or learning resources. This

is a sample of an actual data sheet for

interval recording. So, it's on task

behavior. We're looking at teacher.

These are the behaviors that you're

saying yes to. Looking at teacher while

she's talking, talking to teacher or

looking at assignments. So, anything

else would be a no. That's clear. They

didn't put whether it's partial, whole,

or momentary here, but you would write

that down. Like, I'm taking partial or

whole. You're doing a 1 minute interval,

and you have 10 minutes total. So, 1 2 3

4 5 6 7 8 9 10. And this is how we

calculate the data. So, you can see that

first interval, we'll just say, let's

say she took partial. The behavior did

not occur during the entire interval.

Now, in the second interval, she saw

behavior at some point in that one

minute period. We don't know if she saw

it most of the 1 minute period, half of

the one minute period or what, but we

know she saw something. The observer.

Third minute, we saw something. Fourth

minute, we saw some behavior. Five,

there was no behavior. Six, we saw some

behavior. Seven, we saw no behavior.

Eight, we saw some behavior. Again,

could be a second, could be the whole

interval. Nine, we saw some behavior.

and 10 we saw no behavior. How you

calculate or talk about these interval

systems is you count how many pluses you

have over the total intervals. We have

six pluses. So six times we saw a

behavior during an interval and we have

10 intervals. So it's 6 / 10 and you

multiply that by 100 to get your

percentage. So the student was on task

six out of 10 intervals. So 60% of the

time when you come back tomorrow if you

start your intervention you do exactly

this and we can compare that 60% to the

new percentage to see if the

intervention is

working. This is a chart that breaks it

down for you. Poll is record if behavior

occurs at any time for the entire

interval. It's great for increasing

behavior show substain substained

engagement. So if you want to see really

long engagement, you might choose this.

It underestimates behavior and it's not

ideal for brief or infrequent behaviors.

And then partial record if behavior

occurs at any point during the interval.

It's good for reducing behaviors,

efficient for high frequency behaviors,

and does not show duration at all.

Momentary time sampling. Record if the

behavior occurs at any moment at the end

of the interval or throughout. Easy

while multitasking. less time consuming

and great for groups and it may miss

behavior between checks and can over or

underestimate behavior. Okay, thank you.

I'll see you next video.