|
<< Click to Display Table of Contents >> Navigation: User Guide > Project components > Master files |
Master files is the name given to the files in a project that contain all the information about a project. A project folder will also contain many output (report) files.
The master files should be backed up regularly in case of computer failure. See also the Archive section below.
The Companion uses two types of master file:
Every project must have a project file and this is the first master file to be created. This file contains the metadata for the project:
Project global settings include:
•Details about the project such as the job title
•Default settings for this project
•The raw data structure for this project
A project file will contain many entries. An entry is one of the following types:
•Question entry that describes a question to be asked
•Filter entry that describes a selection of respondents based on earlier answers
•Variable entry that combines information from other entries
•Program entry that controls question randomisation
For more detailed information see Entry types.
All entries must have a unique name, except program entries.
The order of entries is important and each entry is allocated a sequence number. The questions will be asked in sequence number order unless randomisation is used. Filters and variables can only refer to entries that precede them in the entry list.
This master file contains the questionnaire data.
The answers given by respondents are stored in a data file and each respondent is allocated a unique serial number.
There are a number of different types of data file. The raw data structure determines which type will be used.
The terms "data" and "raw data" refer to the same information or file. The term "data" is used to refer to the answers to questions. The term "raw data" refers to how those answers are actually stored in a data file.
The Companion uses:
•CSV raw data which means the answers are in spreadsheet form
•"Fixed format" raw data which means that the answers to each question are always stored in the same place for every record.
The places where answers are stored are called data locations. In binary data a data location is also called a column.
The raw data structure can be:
•Character (recommended) - the raw data is a text file with readable numbers and letters
•Binary - the raw data is stored as card columns and each column has 12 codes 1234567890XV
The raw data structure can also be:
•One line per questionnaire (recommended)
•Multiple lines per questionnaire - each line can be up to 99 columns but is usually 80 columns, these lines are sometimes called cards
The Companion usually allocates data locations for you, but you can do this yourself if a particular layout is needed.
IMPORTANT: a data file is no use on its own. The data in it is only accessible when you have a matching project file with entries that describe where, and how, the answers are stored. This will usually be the same project that was used to collect the data.
Whenever a master file is changed a copy is put into a sub-folder called "Archive".
If a mistake is made the previous version of the master file will have been stored in Archive and can be retrieved.
This folder may become cluttered with old versions of files. It should be cleaned out occasionally and can be discarded completely once a project has been completed.