Winsteps Help for Rasch Analysis
Contents
| Index
| Search
- Winsteps Program Manual
- Control Variables Index
- Control Variables by function
- Input Files Index
- Output Tables Index
- Output Files Index
- Plots index
- Introduction
- Installation and Execution
- Menus
- Control and Data Files
- * Control Variables
- #1#= user-defined replaceable token
- &END end of control variables
- &INST start of control instructions
- ; starts a comment
- @Field= name for location in label
- AGREEFILE= paired-person agreement counts
- AGREEGROUP= compare with other groups in Table 35 = Yes
- ALPHANUM= alphabetic numbering
- ANCESTIM= alternative method for anchored analyses
- ASCII= characters
- ASYMPTOTE= item upper and lower asymptotes
- BATCH= Batch/script mode analysis
- BOXSHOW= draw boxes around Output Tables
- BYITEM= display graphs for items
- CATREF= reference category for Table 2
- CFILE= scored category label file
- CHART= graphical plots in Tables 10, 13-15
- CHISQUARE= in IFILE= and PFILE=
- CLFILE= codes label file
- CMATRIX= category matrix = Yes
- CMLE= Conditional Maximum Likelihood Estimation = No
- CODERANGE = numerical range of data codes
- CODES= valid data codes
- CONVERGE= select convergence criteria
- CSV= comma-separated values in output files
- CURVES= probability curves for Table 21
- CUTHI= cut off responses with high expectations
- CUTLO= cut off responses with low expectations
- DATA= name of data file
- DATESHOW= show date in Table headings
- DELIMITER= or SEPARATOR= data field delimiters
- DIF= columns within person label for Table 30
- DISCRIMINATION= item discrimination
- DISFILE= category/distractor/option count file
- DISOPTION= distractor file option = C
- DISTRACTOR= output option counts in Tables 10, 13-15
- DPF= columns within item label for Table 31
- EDFILE= edit data file
- EFILE= expected scores on items
- END LABELS or END NAMES
- EQFILE= code equivalences
- ESTMULT= estimate change multiplier
- ETYPE= option for EFILE=
- EXCAT= extreme category adjustment
- EXCELDECIMAL= decimal separator for Excel spreadsheets
- EXCELNUMBER= Excel converts numeric labels into numbers
- EXTRSC= extreme score correction for extreme measures
- FITHIGH= higher bar in charts
- FITI= item misfit criterion
- FITLOW= lower bar in charts
- FITP= person misfit criterion
- FORMAT= reformat data
- FORMFEED= the form feed character
- FRANGE= half-range of fit statistics on plots
- FREQUENT= heading in Tables 1, 12, 16
- FSHOW= show files created from control file
- GRFILE= probability curve coordinate output file
- GROUPS= or ISGROUPS= assigns items to rating scale groupings
- GRPFROM= location of ISGROUPS
- GUFILE= (G0ZONE=, G1ZONE=) Guttmanized response file
- HEADER= display or suppress subtable headings
- HIADJ= correction for top rating scale categories
- HLINES= heading lines in output files
- IADD= add item difficulty to GRFILE= measure
- IAFILE= item anchor file
- IANCHQU= anchor items interactively
- ICOLORFILE= colors for item labels in PKMAPs
- ICORFILE= item residual correlation file
- IDELETE= item one-line item deletion
- IDELQU= delete items interactively
- IDFILE= item deletion file
- IDROPEXTREME= drop items with extreme scores
- IFILE= item output file
- ILFILE= item label file
- IMAP= item label on item maps Tables 1, 12
- INDEXED= summarize response strings for estimation = Yes
- INUMB= label items by sequence numbers
- IOFSFIELDS= field selection for IFILE=
- IPDECIMALS= decimals reported for values in IFILE=, PFILE=
- IPEXTREME= placement of extreme scores on maps
- IPMATRIX= response-level matrix
- IPRFILE= change codes for blocks of persons and items
- IREFER= identifying items for recoding
- ISELECT= item selection criterion
- ISFILE= item structure output file
- ISGROUPS= or GROUPS= assigns items to rating scale groupings
- ISORT= column within item name for alphabetical sort in Table 15
- ISRANGE= category range of a rating scale
- ISUBTOTAL= columns within item label for Table 27
- ITEM= title for item labels
- ITEM1= column number of first response
- ITLEN= maximum length of item label
- IVALUEx= recoding of data
- IWEIGHT= item (variable) weighting
- KEYFROM= location of KEYn
- KEYn= scoring key
- KEYSCR= reassign scoring keys
- LCONV= logit change at convergence
- LESS= heading in Tables 1, 12, 16
- LINELENGTH= length of printed lines
- LOCAL= locally restandardize fit statistics
- LOGFILE= accumulates control files
- LOGLIKE= log-likelihood of response string
- LOWADJ= correction for bottom rating scale categories
- MAKEKEY= construct MCQ key
- MATRIX= correlation output format
- MAXPAGE= the maximum number of lines per page
- MBOTTOM= lowest measure in Table 1 etc.
- MCMLE= maximum number of CMLE iterations = 0, no limit
- MFORMS= reformat input data
- MHSLICE= Mantel-Haenszel slice width
- MHZERO= adjustment to zero cells in Mantel cross-tabulation
- MISSCORE= scoring of missing data codes
- MJMLE= or MUCON= maximum number of JMLE iterations
- MNSQ= show mean-square or standardized fit statistics
- MODELS= assigns model types to items
- MODFROM= location of MODELS
- MORE= heading in Tables 1, 12, 16
- MPROX= maximum number of PROX iterations
- MRANGE= half-range of measures on plots
- MTICK= measure distance between tick marks on plots = 0, auto-size
- MTOP= highest measure in Table 1 etc.
- MUCON= or MJMLE= maximum number of JMLE iterations
- MUNITS= measurement units per line in Table 1 etc.
- NAME1= first column of person label
- NAMLEN= length of person label
- NAMLMP= name length on map for Tables 1, 12, 16
- NEWSCORE= recoding values
- NI= number of items
- NOMEASURE= measure value for deleted and inestimable = 9999
- NORMAL= normal distribution for standardizing fit
- OFILE= report output file
- OSORT= option/distractor sort
- OUTFIT= sort misfits on infit or outfit
- PAFILE= person anchor file
- PAIRED= correction for paired comparison data
- PANCHQU= anchor persons interactively
- PCORFIL= person residual correlation file
- PDELETE= person one-line item deletion
- PDELQU= delete persons interactively
- PDFILE= person deletion file
- PDROPEXTREME= drop persons with extreme scores
- PERSON= title for person labels
- PFILE= person output file
- PKMAP= customize Table 36 diagnostic maps
- PLFILE= person label file
- PMAP= person label on person map: Tables 1, 16
- POFSFIELDS= field selection for PFILE=
- PRCOMP= residual type for principal components analyses in Tables 23, 24
- PSELECT= person selection criterion
- PSORT= column within person label for alphabetical sort in Table 19
- PSUBTOTAL= columns within person label for Tables 28, 35
- PTBISERIAL= point-biserial-type correlation coefficient
- PVALUE= proportion correct or average rating
- PWEIGHT= person (case) weighting
- QCMLE = Quasi-CMLE estimates
- QUOTED= quote-marks around labels
- RARE= heading in Tables 1, 12, 16
- RCONV= score residual at convergence
- REALSE= inflate S.E. for misfit
- RESCORE= response recoding
- RESFROM= location of RESCORE
- RFILE= scored response file
- RMSR= report root-mean-square residuals in measure tables
- ROW1HEADING = heading in first row of output file
- SAFILE= structure-threshold input anchor file
- SAITEM= item numbers in SAFILE with one grouping
- SANCHQ= anchor category structure interactively
- SCOREFILE= person score file
- SCREEN= screen log file
- SDELQU= delete category structure interactively
- SDFILE= category structure deletion file
- SEPARATOR= or DELIMITER= data field delimiters
- SFILE= category structure-threshold output file
- SFUNCTION= function to model Andrich thresholds with polynomials
- SICOMPLETE= simulate complete data
- SIEXTREME= simulate extreme scores
- SIFILE= simulated data file
- SIMEASURE= measure or data
- SINUMBER= number of simulated data files
- SIRESAMPLE= number of persons resampled
- SISEED= simulated data seed
- SORTABLE= webpage has sortable columns
- SPFILE= supplementary control file
- STBIAS= correct for estimation bias
- STEPT3= include category structure summary in Table 3.2 or 21
- STKEEP= keep unobserved intermediate categories in structure
- SUBSETS= perform subset detection
- SVDEPOCHS= maximum epochs for SVD estimation
- SVDFACTORS= singular-value decomposition factors
- SVDFILE= singular-value decomposition file
- SVDMIN= singular-value decomposition minimum improvement
- SVDTYPE= singular-value decomposition residual type
- T1CAT= show category number
- T1I#= number of items summarized by "#" symbol in Table 1
- T1MWE= mean and S.D. with extreme measures in Tables 1, 12, 16
- T1P#= number of persons summarized by "#" symbol in Table 1
- T1SCORE= show raw score in Tables 1,12,16
- T1WEIGHT= show items and persons by weight in Tables 1, 12, 16
- T2SELECT= selects Table 2 subtables
- T30PAIRSMAX= maximum pairs to display in Tables 30.1, 31.1
- T45OPTIONS= field choices for Table 45
- T7OPTIONS = selects Table 7.1 and 11.1 detail lines
- TABLES= output tables
- TARGET= estimate using information-weighting
- TCCFILE= Test Characteristic Curve and Test Information Function
- TCCHIGH= highest value of TCC
- TCCINCR= increment on TCC measurement scale
- TCCLOW= lowest value of TCC
- TFILE= input file listing tables to be output
- TITLE= title for output listing
- TOTALSCORE= show total scores with extreme observations
- TRPOFILE= transposing the data matrix
- TRPOTYPE= transposing specifications
- UAMOVE= move all anchor values a pre-set amount
- UANCHOR= anchor values supplied in user-scaled units
- UASCALE= anchor user-scale value of 1 logit
- UCOUNT= number of unexpected responses: Tables 6, 10
- UDECIMALS= number of decimal places reported
- UEXTREME= include extreme scores for UIMEAN= or UPMEAN=
- UIMEAN=or UMEAN= the mean or center of the item difficulties
- UMEAN=or UIMEAN= the mean or center of the item difficulties
- UPMEAN= the mean or center of the person abilities
- USCALE= the user-scaled value of 1 logit
- UTF8ENC= UTF-8 encoding
- UTF8SUB= substitute character for multibyte UTF-8
- W300= Output files in Winsteps 3.00 format
- WEBFONT= font of webpages
- WHEXACT= Wilson-Hilferty exact normalization
- WMLE = Warm's Mean Likelihood Estimation
- XFILE= analyzed response file
- XWIDE= columns per response
- Examples of Control and Data Files
- Analysis Window
- Output Tables
- Charts
- Graphs
- Plots - Excel
- Plots - R Statistics
- Plots - Webpage Browser
- Output dialogs
- Special Topics
- Advice to novice analysts
- Anchored estimation
- Annotating graphics and tables
- Arithmetical expressions
- Automating file selection
- Average measures, distractors and rating scales
- Batch mode example: Score Tables
- Bayesian prior distributions
- Between-group and Within-group fit statistics
- Big / Huge / Maximum Datasets
- Biserial correlation
- Bonferroni - Multiple t-tests
- Bradley-Massof Successive (Consecutive) Polytomous Pseudo-Rasch Model
- Bradley-Terry model and paired comparisons
- Category boundaries and thresholds
- Category mean-square fit statistics
- Chi-squares, d.f., probabilities and Mean-squares
- CMLE, MMLE, PMLE: Using estimates from other software in Winsteps
- Column and classification selection and transformation
- Combining items into superitems and testlets
- Comparing estimates with other Rasch software
- Comparing tests or instruments
- Computer-Adaptive Testing (CAT)
- ConQuest Rasch analysis software, Quest and Winsteps
- Convergence considerations
- Correlations: point-biserial, point-measure, residual
- Decimal, percentage and continuous data
- Decision consistency
- Demographic codes - changing
- Dependency and unidimensionality - Repeated measures - Longitudinal studies
- Dichotomous mean-square fit statistics
- DIF - DPF - bias - interactions concepts
- DIF for two person groups
- Dimensionality
- Dimensionality and Structural Validity investigation - an example
- Dimensionality: PCAR contrasts & variances
- Dimensionality: when is a test multidimensional?
- Disjoint strings of responses
- Disordered Andrich Thresholds and disordered rating categories
- Displacement measures
- Edit taskbar caption
- Equating and linking tests
- Estimation bias correction - warnings
- Estimation methods: JMLE, PROX, WMLE, CMLE, PMLE, AMLE
- Exact Match: OBS% and EXP%
- Excel plotting
- Excel-formatted output Tables
- Extra Specifications prompt
- Extreme scores: what happens
- Faster analysis: How long will an analysis take?
- Fit diagnosis: infit outfit mean-square standardized
- Global fit statistics
- Glossary
- Google Map of Winsteps statistics
- Guttman errors, Guttman reversals
- Guttman parameterization of rating scales
- Half-rounding
- Inestimable measures
- Information - item and test
- Installing multiple versions of Winsteps
- IRT to Rasch conversion
- Item difficulties from score-to-measure table
- Item difficulty: definition
- Item discrimination or slope estimation
- Item map in Excel or Word
- Item Maps - Combining
- Iterations - PROX & JMLE
- jMetrik Rasch analysis software and Winsteps - useful features
- JMLE is inconsistent!
- Local Dependence
- Logit and probit
- Mantel and Mantel-Haenszel DIF statistics
- Misfit diagnosis: infit outfit mean-square standardized
- Missing data
- Mixed or Mixture Models and Saltus models
- Müller's continuous rating scale model CRSM
- Multidimensional Model - Multidimensionality
- Multiple copy-and-paste or other repetitive actions
- Multiple Response items - MR questions
- Non-uniform DIF tables
- Null or unobserved categories: structural and incidental zeroes
- One item test - One observation per respondent
- Order of elements in Control file
- Partial Credit model
- Person Response Functions (PRF)
- Person-free and Item-free measurement
- Pivot anchoring
- Plausible values
- Poisson counts
- Polytomous mean-square fit statistics
- Printing pretty tables and graphs
- Prior distributions
- Probabilities from measures
- Quality-control misfit selection criteria
- R Statistics Rasch analysis software and Winsteps
- Random number generator
- Rank order data
- Rasch - Why use Rasch methodology?
- Rasch Model
- Rasch Software
- Rating scale conceptualization: Andrich, Thurstonian, half-point thresholds
- Ratio scale of measures: twice as much
- Raw scores as measures
- RColorBrewer palettes
- Rectangular copying
- Registering .dll or .ocx files
- Reinstalling Winsteps
- Reliability and separation of measures
- Response-pattern frequency
- Rules for assigning values to control variables (key-words)
- RUMM2030+ Rasch analysis software and Winsteps
- Scoring items with multiple responses
- Shortcut Keys
- Simulating data
- Smartphone - running Winsteps on your PC
- Specifying how data are to be recoded
- Speeded tests
- Split items - procedure
- Standard errors: model and real
- Standard setting
- Starting Winsteps from the DOS prompt
- Subsets and connection ambiguities
- Subtest scoring and weighting
- Super-items
- TAM R Statistics Rasch analysis software and Winsteps
- Teaching Rasch measurement, Rasch analysis_2
- Testlets
- Theta or B? parameterizing ability
- t-statistics
- Unidimensionality
- Unobserved and dropped categories
- User-friendly rescaling: zero point and unit
- Using a word processor or text editor
- UTF-8 Chinese character set
- Validity investigation
- Videos and Tutorials
- Weighting items and persons
- Winsteps: history and "steps"
- Wright maps in Word, Excel, R, Notepad++
- Problems and Error Messages
www.winsteps.com