help at square one

questions concerning analysis/theory using program MARK

help at square one

Postby gwilson » Mon Aug 04, 2003 8:30 am

I am just starting to use Mark, and have failed to get past the first step - getting Mark to read my data. I keep getting the error message

'encounter history was too short for the
number of occasions specified'

As far as I am aware I have specified the correct number of occasions. My data is held in MS excel - I've been saving it as a text file then renaming it as an INP file. Are there any specific tricks to converting Excel data into the correct format?

Gavin Wilson
gwilson
 
Posts: 1
Joined: Mon Aug 04, 2003 7:57 am

Re: help at square one

Postby brant » Mon Aug 25, 2003 6:10 pm

gwilson wrote:I am just starting to use Mark, and have failed to get past the first step - getting Mark to read my data. I keep getting the error message

'encounter history was too short for the
number of occasions specified'

As far as I am aware I have specified the correct number of occasions. My data is held in MS excel - I've been saving it as a text file then renaming it as an INP file. Are there any specific tricks to converting Excel data into the correct format?

Gavin Wilson


Gavin,

There are several things that could be going wrong. Most of these depend on the type of modelling you are attempting.

First, (although i am sure you have done this) make sure that your encounter periods (weeks, days, months, etc.) actually match what you input to the data importation screen of Mark (e.g. make sure you have 14 encounter occasions if you input 14).

Also, make sure your data are formatted correctly for the model you choose (e.g. LDLD format for Known-Fates (Kaplan-Meier), LLL for CJS model, etc.)

Another thing to do is to check the Mark "Book" under the Data Formatting Chapter (Chapter 2). You will first want to make sure that your .inp (text) files look *exactly* like those in the examples.

Another GOOD example of working data files is found here.

For example, if you code known fates data from a period as such:

/*001180*/ 101010101011 1.67 2.43 18.5
/*001179*/ 000010101100 2.85 2.62 18.7

/*CritterID*/ EncounterHx Covar1 Covar2 Covar3

This would be wrong. The reason this would be is because you are lacking the "1" follwing the encounter history data that "shows" Mark that the data is on an individual basis. The correct format would be:

/*001180*/ 101010101011 1 1.67 2.43 18.5
/*001179*/ 000010101100 1 2.85 2.62 18.7

/*CritterID*/ EncounterHx "indiv. code" Covar1 Covar2 Covar3

Missing little things like this can lead to the nasty (and uncool) 'encounter history was too short for thenumber of occasions specified' or any number of other strange happenings.

-brant
brant
 
Posts: 5
Joined: Wed May 28, 2003 7:54 am
Location: Athens, GA


Return to analysis help

Who is online

Users browsing this forum: No registered users and 1 guest