SPSSX Discussion

Calculate time/event type across cases

Classic

List

Threaded

4 messages Options

Esther Fujiwara

Calculate time/event type across cases

Hi,
I am running into a seemingly simple problem I cannot figure out so far.
In short, my data looks like this:

ID ev time
1 0 0
1 0 10
1 1 20
1 1 30
1 1 40
1 0 50
1 0 60
1 0 70
1 1 80
1 1 90
etc.
2 0 0
2 0 10
2 1 20
2 1 30
2 1 40
2 0 50
etc.

'ID' is a picture ID
'ev' is eyefixation (yes, no)
'time' is a consecutive run time (in ms).

What I need to extract are two tings:

First I need a number that depicts first, second third _consecutive_
eyefixations for each pic ID. That is some sort of counter that goes through
variable 'ev' and creates a new variable that gives a '1' to the first set
of consecutive cases in which ev = 1, a '2' to the second set, and so on.
So, for this example, the counter would be '0' for the first 2 cases of ID
1, '1' for cases 3-5, '0' for cases 6-8, and '2' for cases 9-10 and so on.

If that were possible, I would then like to calculate the fixation times for
each distinct block consecutive fixations. So, how long did people look at
the picture when they first looked, and thereafter. For Id 1, the first
block of time would be 30 ms (cases 2-5), the second block of time would be
20 ms (cases 9-10); for ID it would be 30 ms (cases 3-5).

Does someone have any idea on how/if this is possible..???

Thanks in advance for any help!

~Esther

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

Maguin, Eugene

Re: Calculate time/event type across cases

I think this yields what you want. Does it? Gene Maguin

data list list / ID ev time.
begin data
1 0 0
1 0 10
1 1 20
1 1 30
1 1 40
1 0 50
1 0 60
1 0 70
1 1 80
1 1 90
2 0 0
2 0 10
2 1 20
2 1 30
2 1 40
2 0 50
end data.

Compute efcount=0.
Do if (id eq lag(id)).
If (ev eq 1) efcount=lag(efcount)+1.
If (ev eq 0) efcount=lag(efcount).
End if.
execute.
Aggregate outfile=* mode=addvariables/break= id efcount/fixtime=sum(time).
format ID ev(f1.0) time efcount(f2.0) fixtime(f3.0).
list id ev time efcount fixtime.

ID ev time efcount fixtime

1 0 0 0 10
1 0 10 0 10
1 1 20 1 20
1 1 30 2 30
1 1 40 3 220
1 0 50 3 220
1 0 60 3 220
1 0 70 3 220
1 1 80 4 80
1 1 90 5 90
2 0 0 0 10
2 0 10 0 10
2 1 20 1 20
2 1 30 2 30
2 1 40 3 90
2 0 50 3 90

Number of cases read: 16 Number of cases listed: 16

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Esther Fujiwara
Sent: Saturday, January 25, 2014 2:15 PM
To: [hidden email]
Subject: Calculate time/event type across cases

Hi,
I am running into a seemingly simple problem I cannot figure out so far.
In short, my data looks like this:

ID ev time
1 0 0
1 0 10
1 1 20
1 1 30
1 1 40
1 0 50
1 0 60
1 0 70
1 1 80
1 1 90
etc.
2 0 0
2 0 10
2 1 20
2 1 30
2 1 40
2 0 50
etc.

'ID' is a picture ID
'ev' is eyefixation (yes, no)
'time' is a consecutive run time (in ms).

What I need to extract are two tings:

First I need a number that depicts first, second third _consecutive_ eyefixations for each pic ID. That is some sort of counter that goes through variable 'ev' and creates a new variable that gives a '1' to the first set of consecutive cases in which ev = 1, a '2' to the second set, and so on.
So, for this example, the counter would be '0' for the first 2 cases of ID 1, '1' for cases 3-5, '0' for cases 6-8, and '2' for cases 9-10 and so on.

If that were possible, I would then like to calculate the fixation times for each distinct block consecutive fixations. So, how long did people look at the picture when they first looked, and thereafter. For Id 1, the first block of time would be 30 ms (cases 2-5), the second block of time would be
20 ms (cases 9-10); for ID it would be 30 ms (cases 3-5).

Does someone have any idea on how/if this is possible..???

Thanks in advance for any help!

~Esther

=====================
To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

David Marso

Re: Calculate time/event type across cases

Administrator

In reply to this post by Esther Fujiwara

Since your description is rather inconsistent I leave you to look at the LAG function in COMPUTE and AGGREGATE commands!
If you more carefully describe how you get your durations maybe more to come.
How is it cases 2-5 are in the first and 9-10 in the second.
Is it NECESSARY that the 1's get enumerated and the 0's get set to 0? If so it will be a pain in the ass, if not then utterly trivial!

IF ID NE LAG(ID) OR $CASENUM=1 CountX = 1.
IF ev EQ LAG(EV) CountX = LAG(CountX).
IF MISSING(CountX) CountX = LAG(Countx) + 1.

---

Esther Fujiwara wrote

Hi,
I am running into a seemingly simple problem I cannot figure out so far.
In short, my data looks like this:

ID ev time
1 0 0
1 0 10
1 1 20
1 1 30
1 1 40
1 0 50
1 0 60
1 0 70
1 1 80
1 1 90
etc.
2 0 0
2 0 10
2 1 20
2 1 30
2 1 40
2 0 50
etc.

'ID' is a picture ID
'ev' is eyefixation (yes, no)
'time' is a consecutive run time (in ms).

What I need to extract are two tings:

First I need a number that depicts first, second third _consecutive_
eyefixations for each pic ID. That is some sort of counter that goes through
variable 'ev' and creates a new variable that gives a '1' to the first set
of consecutive cases in which ev = 1, a '2' to the second set, and so on.
So, for this example, the counter would be '0' for the first 2 cases of ID
1, '1' for cases 3-5, '0' for cases 6-8, and '2' for cases 9-10 and so on.

If that were possible, I would then like to calculate the fixation times for
each distinct block consecutive fixations. So, how long did people look at
the picture when they first looked, and thereafter. For Id 1, the first
block of time would be 30 ms (cases 2-5), the second block of time would be
20 ms (cases 9-10); for ID it would be 30 ms (cases 3-5).

Does someone have any idea on how/if this is possible..???

Thanks in advance for any help!

~Esther

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"

Esther Fujiwara

Re: Calculate time/event type across cases

In reply to this post by Esther Fujiwara

On Mon, 27 Jan 2014 15:30:30 +0000, Maguin, Eugene <[hidden email]>
wrote:

>I think this yields what you want. Does it? Gene Maguin
>
>data list list / ID ev time.
>begin data
>1 0 0
>1 0 10
>1 1 20
>1 1 30
>1 1 40
>1 0 50
>1 0 60
>1 0 70
>1 1 80
>1 1 90
>2 0 0
>2 0 10
>2 1 20
>2 1 30
>2 1 40
>2 0 50
>end data.
>
>Compute efcount=0.
>Do if (id eq lag(id)).
>If (ev eq 1) efcount=lag(efcount)+1.
>If (ev eq 0) efcount=lag(efcount).
>End if.
>execute.
>Aggregate outfile=* mode=addvariables/break= id efcount/fixtime=sum(time).
>format ID ev(f1.0) time efcount(f2.0) fixtime(f3.0).
>list id ev time efcount fixtime.
>
>ID ev time efcount fixtime
>
> 1 0 0 0 10
> 1 0 10 0 10
> 1 1 20 1 20
> 1 1 30 2 30
> 1 1 40 3 220
> 1 0 50 3 220
> 1 0 60 3 220
> 1 0 70 3 220
> 1 1 80 4 80
> 1 1 90 5 90
> 2 0 0 0 10
> 2 0 10 0 10
> 2 1 20 1 20
> 2 1 30 2 30
> 2 1 40 3 90
> 2 0 50 3 90
>
>Number of cases read: 16 Number of cases listed: 16
>
>-----Original Message-----
>From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of

Esther Fujiwara

>Sent: Saturday, January 25, 2014 2:15 PM
>To: [hidden email]
>Subject: Calculate time/event type across cases
>
>Hi,
>I am running into a seemingly simple problem I cannot figure out so far.
>In short, my data looks like this:
>
>ID ev time
>1 0 0
>1 0 10
>1 1 20
>1 1 30
>1 1 40
>1 0 50
>1 0 60
>1 0 70
>1 1 80
>1 1 90
>etc.
>2 0 0
>2 0 10
>2 1 20
>2 1 30
>2 1 40
>2 0 50
>etc.
>
>'ID' is a picture ID
>'ev' is eyefixation (yes, no)
>'time' is a consecutive run time (in ms).
>
>What I need to extract are two tings:
>
>First I need a number that depicts first, second third _consecutive_

eyefixations for each pic ID. That is some sort of counter that goes through
variable 'ev' and creates a new variable that gives a '1' to the first set
of consecutive cases in which ev = 1, a '2' to the second set, and so on.
>So, for this example, the counter would be '0' for the first 2 cases of ID
1, '1' for cases 3-5, '0' for cases 6-8, and '2' for cases 9-10 and so on.
>
>If that were possible, I would then like to calculate the fixation times
for each distinct block consecutive fixations. So, how long did people look
at the picture when they first looked, and thereafter. For Id 1, the first
block of time would be 30 ms (cases 2-5), the second block of time would be

>20 ms (cases 9-10); for ID it would be 30 ms (cases 3-5).
>
>Does someone have any idea on how/if this is possible..???
>
>Thanks in advance for any help!
>
>~Esther
>
>=====================
>To manage your subscription to SPSSX-L, send a message to

[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command SIGNOFF SPSSX-L For a list of
commands to manage subscriptions, send the command INFO REFCARD
>
>=====================
>To manage your subscription to SPSSX-L, send a message to
>[hidden email] (not to SPSSX-L), with no body text except the
>command. To leave the list, send the command
>SIGNOFF SPSSX-L
>For a list of commands to manage subscriptions, send the command
>INFO REFCARD

Beautiful!!! This is exactly what I needed - did not occur to me to use a
counter this way.

Thank you!

Esther

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD