Classic users

<< Click to Display Table of Contents >>

Navigation:  Reference Manual > Release information >

Classic users

This section is intended for users of the QPSMR Classic programs.

Your existing licence includes the equivalent Companion version.

All files produced from Classic programs can be used with Companion.

Important: Companion does a more thorough check when a project is opened and there are settings in Project global to ignore unwanted warnings:

Allow empty (blank) entry texts

Ignore carry forward in text

Suppress ambiguous response text warnings

Major reasons to upgrade to Companion from Classic Insight or Classic CATI

Companion is a Unicode program and can handle any languages.

Questionnaire design

Response settings can be applied to many responses simultaneously

Drag and drop filters to apply them

Filters can be applied to many entries at once

There is an undo facility when updating entries

Entries can be put into coloured groups for faster access

Extended and CATI versions have block updates to add summary rows, reject responses, score values, refuse input, and ranking

Data

CSV (spreadsheet) data files can be used.

You can view and update data files in grid form

Drill Down facility for data exploration

Distinction analysis (in Drill Down)

File encoding check

Raw data files can be viewed

More control over imports and exports

Combine and convert data files

Verbatim files can be included when viewing data

Tables production

Quick tables improvements

Built in weighting procedure

Tables produced as formatted Excel with a table of contents

Drag and drop table definition

Table settings can be applied to many tables simultaneously

Up to 4 character column identifiers

Grid tables produced automatically using appropriately names entries

Default formats can be set for a project.

Other reasons to switch to Companion

Easy upgrade to new versions. No need for a new licence when a new version is installed, only needed when the licence expires.

ClickOnce install for automatic upgrade to new releases.

Column widths on the main window can be set for all projects.

Shorthand way to add lots of value entries together.

Shorthand way to create filters from responses.

Shorthand way to create variables (breakdowns) from multiple entries.

Decimal places can be set for value entries.

Named random entries so that the same randomisation can be repeated for later entries.

Automatic allocation of column identifiers.

Android input version for CAPI surveys.

GPS coordinates can be collected for CAPI surveys.

Improved triple-s import.

Input shows more than one question on the screen at a time.

Improved data merge facilities.

Paste special in tables setup.

Reflect export from tables setup applies any global filters to the export.

Background

QPSMR Insight and QPSMR CATI have been replaced by the QPSMR Companion which should be used instead.  The licences for these cover the same functionality in the QPSMR Companion.

QPSMR Reflect has been replaced with Companion Reflect which should be used instead.

Companion Input can be used without a licence on an Input file (.qpsmrinp) and should be used instead of QPSMR Input for paper and CAPI projects.

The QPSMR Companion, Companion Reflect, and Companion Input should be used instead of the Classic programs for everything except:

QPSMR Telin to input CATI data

Command language scripts can also be run using the CL menu in the QPSMR Companion if you have an extended licence.

The same files can be used with the Companion programs that were used with the Classic programs:

Project files (.qdf)

Tables files (.qtf)

Raw data files (.asc or .cba)

Portable data files (.csv)

Sample files (.asc and .qsp)

Quota files (.csv)

Appointments files (.asc)

IMPORTANT: Grid questions in Classic will be converted into separate questions by Companion.  Grid tables can be created from the separate questions in Companion and Companion Reflect.

Classic grid tables will run unchanged in Companion.

If you install a new version of the Companion program, in Classic use [Options] [QPSMR Companion program] to point to the newly installed Companion program.

Companion should always be used to import and export a triple-s XML file.

There are many additional modules in the Companion that are not available in the Classic programs.

Tips for projects and entries

Whenever a project is opened with Companion it is thoroughly checked.  This may produce warnings of problems that are ignored by the Classic programs.  There are settings in Project global to suppress these warnings.

Opening and saving a project with the Companion tidies up the project.

The entry update screens are very different to those in Classic.  All changes are done immediately; use undo to revert to the original.

In Classic a single-coded variable with an R (reject) category will be selected if more than one of the other responses is true.  In Companion the first true response will be selected; the R (reject) category is only selected if all the other responses are false.

If you want to use an R (reject) response in a single-coded variable:

make a multi-coded variable (with the R) instead

copy this multi-coded variable to a single-coded variable

When using the Companion:

The entry list on the Main window can be sorted by column, an additional column contains a sequence number that can be used to revert to the natural order of questions

A right mouse click can be used to get a context menu

The Project global settings also contains the tables default settings

There is an undo (and redo) button in the Entry details window for entry updates

New entries are inserted after the highlighted entry unless "insert before" is used

You can copy an entry using the Entry details window, this is easier than using copy and paste

You can paste replacement entries that overwrite the originals

You can make a variable from an entry using the Entry details window

You can copy question responses and paste into a variable

You can make a breakdown variable using Paste combine

There is a separate menu option on the Main window to paste as variables (not paste special)

You can make CT and SCORE function variables using the Entry details window menu

Variable responses block insert allows negative references and priority control, see Variable responses add

Response settings can be applied to many responses simultaneously using the tool windows

You can start response key numbers with 0 instead of 1

Float entries can have a fixed number of decimal places

You globally ask for response keys to use the full data width

You can alter the data locations of clipboard entries and remove references when using Paste special

You can use regular expressions when selecting by names or texts

Data locations are allocated in an intelligent manner regardless of the setting of the Locations fixed switch

You can apply filters by dragging them from the main window and dropping them in the appropriate place

Because Classic uses only upper case for names, you can set the Project global option to upper case all new entry names.  If you need to do this later then use [Edit] [Select all], then [Edit] [Rename] and [All uppercase] to make all the existing entry names upper case.

A "Responses filter" is called a "Response restrictor" in the Companion.

Tips for data

Companion ignores the "Zero or Undefined" global switches and assumes they are both set to "Undefined".

Companion can handle fixed format data with data locations as byte positions (ASC) or character positions (UNI).

Companion Input (Windows and Android) puts more than one question on each screen.

If you select a block of entries on the main window you can choose to only test those entries along with any entries that they depend on (that they refer to)

Every time a data file is saved it is sorted and compacted and all filters and variables are recalculated and reapplied.

When checking data, Companion will add supplemental reject responses (user defined rejects) where appropriate.  This may affect subsequent filters or variables; a question will never be blank (empty) if it has a supplemental reject.

You can request Companion to correct errors as it scans a records.  If you do this and an error is corrected then all subsequent filters and variables will use the corrected answer instead of the original contents.  Here is an example that uses response filters:

With these 3 questions that all have the same response list:

Q1

Q2 with response restrictor [Q1] applied

Q3 with response restrictor [Q2] applied

If Q1 has responses 1 and 3, Q2 has response 2, and Q3 has response 2 normally only Q2 will be flagged

With corrected errors both Q2 and Q3 will be flagged because Q2 will be empty after cleaning

Alternatively, with the same questions but different response filters:

Q1

Q2 with response restrictor [Q1] applied

Q3 with negative response restrictor NOT [Q2] applied

If Q1 has responses 1 and 3, Q2 has response 2, and Q3 has response 2 normally only Q2 and Q3 will be flagged

With corrected errors, only Q2 will be flagged because Q2 will again be empty after cleaning

When checked data is viewed, any faulty questions are highlighted.  Records with at least one error are also highlighted.

The ability to list unused data is a good check on a QDF.  For example, if a question or response is missing from the QDF this will show as unused data in the data file.

Any data in the file that has not been used will be listed, with one exception: if a code 1 is checked for an answer to a multi-coded question in character data (ASC or UNI), then the code 0 will also be assumed to have been used.  This prevents errors being generated when a multi-coded question is stored as a series of 0 or 1 codes.  If Companion saves the data, then the 0s will not be output; only the 1s will be output.

In Companion, data is viewed in a "data grid" in the Data view window.  Each row is one questionnaire and each column is an entry.  There are options to restrict the rows and columns visible in the display.

Whenever a raw data file is saved, it is cleaned, sorted and compacted.

IMPORTANT: make sure that no questions or responses have been "falsed out" because any filtered data will be made empty in a saved file. Use "Suspend" instead.

[Data] [Export data] replaces Classic data lists, data hand-out, and data export to CSV:

You can save the settings to an Export details file for reuse later

You can include verbatim data from the separate verbatim file

These can be run from a command (BAT) file

[Data] [Merge data files] replaces Classic data hand-in of portable or forcing files and updating coded information:

Excel files can be used directly

Merge files can contain response texts instead of response numbers

Any type of entry can be used for forcing data (matching the first column)

You can use a filter definition to locate records in forcing merge files

[Data] [Weight data file] replaces Classic WGT files and STP rim weighting files:

The effective sample sizes (efficiencies) are shown

The calculated weights are stored in the data file

You can save the settings to a Weighting file for reuse later

There is no practical limit on the number of entries or responses that can be used as targets

Multi-coded entry responses can be used for targets

Integer and float entries can be used for targets

You can set each target to one of the following:

oWeighted target value

oPercentage of respondents

oPercentage of the base for this entry

You can set priorities to increase efficiency for some targets over others

Weights can be constrained to a range of values

A comprehensive weighting report is produced which can be saved to a CSV file.