Data for analysis =

This tells Facets where to find the data.  Observations must be ordered integers in the range 0-254 for analysis. Facets can recode other data into integers using Rating Scale=


Either in the same file as the specifications:

Data =

followed by the data in the Facets data-format. This must be the last control specification, e.g.,


2,23,6,4 ;(The data)




Or in a separate data file or data files:

Data = diskfile.txt

where diskfile.txt is the name of the file containing the responses in the Facets data-format.


This can be extended to several data files with "+" signs:

Data = diskfile1.txt + diskfile2.txt + diskfile3.txt


or use %include file

Data =





The format of the diskfile.txt is determined by the filename suffix:



Data= Format


text file (MS-DOS or Windows)

.xls .xlsx

Excel workbook: first or only worksheet

.rda .rdata

R data file


SAS data file


SPSS data file


STATA 13 and earlier data file


text file (MS-DOS or Windows)


If the data file is not found, a file dialog box is displayed:



Please Edit Specification file, and Exit then restart the analysis.



Facets data format: Single observations


Facet 1

Facet 2

Facet ...


Element number

or Element label

or Dvalues=

Element number

or Element label

or Dvalues=



or value to recode with Rating-scale=


Facets data format: Multiple observations


Facet 1

Facet 2

Facet ...


Element number

or Element label

or Dvalues=

Element number range

(one facet, any facet)


numbers to match the element number range

or values to recode with Rating-scale=


Every line of data must have the same number of facet references (element numbers or element labels), usually given by Facets=. These facet references always appear in the same order. The element numbers are given in order by facet number, as specified after Labels=, unless overridden by Entered=. The facets references are then followed by one or more observed responses.


Example 1. Here is the flat file of observations:

123 665 25467

452 339 34245

312 034 34452

^   ^   ^



       Items 1 through 5

where format of flat file record is:

columns 1-3, "123" in record 1, are examinee identifiers

columns 5-7, "665" in record 1, are judge identifiers

columns 9-13, "25467" in record 1, are the ratings on items 1 to 5.


Convert this to Facets format. Insert commas and "1-5" added:


1, Examinee





2, Judge





3,  Item



Dvalues = 3, 1-5  ; put 1-5 in third data facet

; Extra blanks in the data file are ignored,  also comments starting ";"


123 665 2 5 4 6 7 ; blanks are separators

452 339 3 4 2 4 5 ; "1-5" added from dvalues=

312 034 3 4 4 5 2 ; 


Example 2: Data references recorded in facet number order with element labels


2, Person ; Facet number 2 in the data



1, Judge ; The facet number, "1", not the order of definition, second, is what matters.



3, Item




Hugo,Fred,4,1 ; Judge 2 (Hugo) rated Person 23 (Fred) on Item 4 (Sincerity) with a 1.


Example 3: Person 13 is rated on Item 4 by Judge 6 with a "3". This is data record is




Example 4: The data are letter grades, which are converted to numeric ratings by a Rating (or partial credit) scale= specification. Student 13's grade by Teacher 4 on Essay 1 is "B+"

Rating (or partial credit) scale=Grades,R9







Example 5: Data references not in facet order. Entry= supplies both the order of the facet references and their identification in the data records:

Facets=4 ; 4 facets defined

Entry = 3,2,1 ; only 3 in data records, in different order

Model = ?,?,?,D ; first ? matches first reference, i.e., facet 3, etc.

Data =

4,23,2,1 ; Judge 2 (Hugo) rated Person 23 (Fred) on Item 4 (Sincerity) with a 1.


Example 6: Element number 0 of each facet is used to indicate that no element of that facet participated, e.g., Judge number 0 is used when no judge intermediated. Keepasnull= enables you to use 0 as an active element number. Thus, if a facet does not take part in an observation, element number 0 is recorded in the position in the data line for that facet, e.g.,

2,0,27,3 ; item 27 (facet 3) did not require a judge from facet 2 to rate it.


Example 7: Some data do not require references to some facets. Person 13 (facet 2) scored a 1 on item 9 (facet 3), which was not judged (facet 1). The judge element number is entered as 0 for that observation in the data:

0,13,9,1 ; element number in facet 1 is 0 (not participating)


Example 8: Multiple observations can be entered on the same data line, provided that the elements are in ascending or descending sequence by element number in one facet, and all other facets are the same. The extreme element numbers of the sequence are given, e.g., items 1 through 10 are indicated by 1-10,  Judge 2 (facet 1) gave person 23 (facet 2) on item 6 (facet 3) a response of 4, then

Judge 2 (facet 1) gave person 24 (facet 2) on item 6 (facet 3) a response of 5.

This can be recorded as



or, this can be recorded in one line as:

2,23-24,6,4,5 or 2,23_24,6,4,5

where the data points are entered to match the elements in the sequence.

The data points are also separated by commas ",".


Example 9: Person 3 (facet 1) took a 10 item test (items 1 through 10 of facet 2). He succeeded, scored 1, on items 1 to 5, and failed (scored 0) on items 6 to 10.

3,1-10,1,1,1,1,1,0,0,0,0,0 ; items ascending order


3,10-1,0,0,0,0,0,1,1,1,1,1 ; items descending order


3,1-4,0,0,0,0 ; some items ascending order

3,5-10,0,1,1,1,1,1 ; more items ascending order


Example 10: Missing observations: Person 13 is rated by Judge 8 on item 2 with a "3", item 3 with a "4", and item 6 with a "2". This is entered in the data file as

13,2-3,8,3,4 ; two observations in sequence

13,6,8,2 ; item not in sequence


13,2-6,8,3,4,,,2 ; items 4 and 5 entered as missing


Example 11: Dvalues= simplifies data formatting. The data are in a fixed format:

Person  Judge  Responses to items 1 to 6

13        4       134243

5       17       243223

These can be converted to Facets format by a word processor (Facets ignores blanks):

13  ,     4, 1-6, 1,3,4,2,4,3

5  ,    17, 1-6, 2,4,3,2,2,3



Dvalues = 3, 1-6  ; data facet 3 is 1-6


13  4 1 3 4 2 4 3 ; blanks are separators

 5 17 2 4 3 2 2 3 ; data facet 3 is 1-6


Example 12: Replication (weighting) of a data point can be specified by R (or another replication character) and the number of replications, for instance:

R3,2,23,6,4 means that the value of 4 was observed in this context 3 times.

Fractional replication permits flexible observation-weighting:

R3.5,2,23,6,4 means that the value of 4 was observed in this context 3.5 times.


Example 13: Organizing the ratings. There are three facets: examinees, tasks, raters. Each examinee performs two tasks that are rated by two raters:


examinee, task 1, rater A, rating

examinee, task 1, rater B, rating

examinee, task 2, rater A, rating

examinee, task 2, rater B, rating


or if the task element numbers are adjacent, then:

examinee, task 1-2, rater A, rating for task 1, rating for task 2

examinee, task 1-2, rater B, rating for task 1, rating for task 2


or if the rater element numbers are adjacent:

examinee, task A, rater 1-2, rating for rater 1, rating for rater 2

examinee, task B, rater 1-2, rating for rater 1, rating for rater 2


but NOT:

examinee, task 1-2, rater 1-2, rating, rating, rating, rating


Help for Facets Rasch Measurement Software: Author: John Michael Linacre.

The Languages of Love: draw a map of yours!

For more information, contact or use the Contact Form

Facets Rasch measurement software. Buy for $149. & site licenses. Freeware student/evaluation download
Winsteps Rasch measurement software. Buy for $149. & site licenses. Freeware student/evaluation download

State-of-the-art : single-user and site licenses : free student/evaluation versions : download immediately : instructional PDFs : user forum : assistance by email : bugs fixed fast : free update eligibility : backwards compatible : money back if not satisfied
Rasch, Winsteps, Facets online Tutorials


Forum Rasch Measurement Forum to discuss any Rasch-related topic

Click here to add your email address to the Winsteps and Facets email list for notifications.

Click here to ask a question or make a suggestion about Winsteps and Facets software.

Rasch Publications
Rasch Measurement Transactions (free, online) Rasch Measurement research papers (free, online) Probabilistic Models for Some Intelligence and Attainment Tests, Georg Rasch Applying the Rasch Model 3rd. Ed., Bond & Fox Best Test Design, Wright & Stone
Rating Scale Analysis, Wright & Masters Introduction to Rasch Measurement, E. Smith & R. Smith Introduction to Many-Facet Rasch Measurement, Thomas Eckes Invariant Measurement with Raters and Rating Scales: Rasch Models for Rater-Mediated Assessments, George Engelhard, Jr. & Stefanie Wind Statistical Analyses for Language Testers, Rita Green
Rasch Models: Foundations, Recent Developments, and Applications, Fischer & Molenaar Journal of Applied Measurement Rasch models for measurement, David Andrich Constructing Measures, Mark Wilson Rasch Analysis in the Human Sciences, Boone, Stave, Yale
in Spanish: Análisis de Rasch para todos, Agustín Tristán Mediciones, Posicionamientos y Diagnósticos Competitivos, Juan Ramón Oreja Rodríguez
Winsteps Tutorials Facets Tutorials Rasch Discussion Groups


Coming Winsteps & Facets Events
May 22 - 24, 2018, Tues.-Thur. EALTA 2018 pre-conference workshop (Introduction to Rasch measurement using WINSTEPS and FACETS, Thomas Eckes & Frank Weiss-Motz),
May 25 - June 22, 2018, Fri.-Fri. On-line workshop: Practical Rasch Measurement - Core Topics (E. Smith, Winsteps),
June 27 - 29, 2018, Wed.-Fri. Measurement at the Crossroads: History, philosophy and sociology of measurement, Paris, France.,
June 29 - July 27, 2018, Fri.-Fri. On-line workshop: Practical Rasch Measurement - Further Topics (E. Smith, Winsteps),
July 25 - July 27, 2018, Wed.-Fri. Pacific-Rim Objective Measurement Symposium (PROMS), (Preconference workshops July 23-24, 2018) Fudan University, Shanghai, China "Applying Rasch Measurement in Language Assessment and across the Human Sciences"
Aug. 10 - Sept. 7, 2018, Fri.-Fri. On-line workshop: Many-Facet Rasch Measurement (E. Smith, Facets),
Oct. 12 - Nov. 9, 2018, Fri.-Fri. On-line workshop: Practical Rasch Measurement - Core Topics (E. Smith, Winsteps),


Our current URL is

Winsteps® is a registered trademark

John "Mike" L.'s Wellness Report: I'm 72, take no medications and, March 2018, my doctor is annoyed with me - I'm too healthy!
According to Wikipedia, the human body requires about 30 minerals, maybe more. There are 60 naturally-occurring minerals in the liquid Mineral Supplement which I take daily.