Fields in a Question Record
The organization of the Reference section is subdivided into the following categories.
This section of the manual is organized around the format of the blank Manager question record. This record is shown in the "Question Form" window and can be viewed by selecting the Manager function Design tab and then selecting the Questionnaire button. The question record contains the principal logic of the Manager program and links itself to the default next record (NextQ field) in the questionnaire’s chain of logic. For complete reference information on the five other record types, see the section entitled "Record Types".
Not all fields appear on every record. Depending on the question type chosen in the Type field, different fields will appear. All fields will be discussed here. Some of those not shown in the "Question Form" are HardMax HardMin, SoftMax, and SoftMin.
The question record also contains additional fields used for documentation and data dissemination purposes. These fields are not necessary during the design process and can be filled in during the data collection phase or even later than that.
To access the additional fields, use the Docs tab on the Main Menu, Questionnaire button.
The order of presentation of all the question record fields is in alphabetical order. To locate a field, identify it in the question record then look it up in the "List of Fields in a Question Record" below.
List of Fields in a Question Record
The following list contains all the fields that comprise a question record. Some of these fields are mandatory for every record, some are optional. Those marked with an asterisk are a part of the Documentation Entries portion of the record Depending on the question type some fields are not used with certain question types. Click on the individual question field to jump to its description.
This field is used to group data into categories based on what topic the variable is related to, hence names like Family Background, Marriage, Military, Jobs, etc. may appear in the drop down (
) menu. This field is more an administrative field and has more to do with post-collection data management than with questionnaire design. This field can be used to provide a way to tag variables independent of the section-groupings enforced during questionnaire design. Contents of the field may be added by typing new terms in the blank field. See also, Chapter 5 - Reference, Question Fields, "Areas of Interest Field."
Audio Field
This field links the current question text to an audio file that reproduces the question text. Simply place the corresponding audio file name (the file must be a *.wav file) in this field.
Card Field*
Used with the Col (Column) field to locate an answer in the answer file. See Column field below. The Card and Column fields are just the conventional card-image addresses of the answer. The first, "left most" columns of each "card" may contain specific identification information for each case depending on how the Col (Column) field is configured.
The location of the responses in the answer file may either be specified by the designer or generated by the software. The choice is a matter of preference, but in a complex, lengthy questionnaire, software generation is subject to fewer errors and fewer headaches.
Card The Card field is hierarchical and used to group a set of cards in a hyperlogical order.
Col (Column) The Column field is a subdivision of a Card arranged in two parts:
If the designer is generating the Deck and Column numbers, the length of the field must match the length in the Format field. For program generation this is automatic.
Administrative field that contains the continuos code ranges for responses. Continuous codes are response ranges used to locate ‘unpredictable’ responses such things as salary, pay, dollar amounts, etc. within a range. These are coded after data collection.
The Column field. It is divided into two boxes and used with the Card field to locate an answer in the answer file. (See Card field above.) The Card and Column fields are the conventional computer card-image addresses of the answer. The "left most" columns of each "card" may contain specific identification information for each case. The amount of identification information depends on where the "column" position of where the answer field begins. The left box is used to specify the column position where the answer begins
The locations of the responses in the answer file may either be specified by the designer or generated by the software. The choice is a matter of preference, but in a complex, lengthy questionnaire, software generation is subject to fewer errors and fewer headaches.
Card The Card field is hierarchical and used to group a set of cards in a hyperlogical order. Cards are given numbers that contain a number of columns.
Col The Column field is a subdivision of a Card arranged in two parts:
If the designer is generating the Deck and Column numbers, the length of the field must match the length in the Format field. For program generation this is automatic.
Comment Field*
The designer or archivist can use this field to add comments to individual Manager records. Comments are administrative in nature and will not appear in the Survey program.
Cond. Jump Field
The Conditional Jump field is composed of several fields that work together to perform complex skip instructions in the logical flow of the interview. These fields can be used with the Translate Answer, Evaluate Expression, Select Option, and various Keyword Code functions that evaluate conditions.
|
Maximum: |
Numeric or ‘symbolic’ value which delimits the upper range of a condition for an answer. Answers that equal this value will trigger the jump. |
|
Max: |
If checked, indicates that the Maximum value is dynamic, that is, the value expressed in the Maximum field is a %symbol% that contains a computed numeric value. This capability allows the designer to compute and specify dynamic range delimiters. |
|
Minimum: |
Numeric or ‘symbolic’ value which delimits the lower range of a condition for an answer. Answers that equal this value will trigger the jump. |
|
Min: |
If checked, indicates that the Minimum value is dynamic, that is, the value expressed in the Maximum field is a %symbol% that contains a computed numeric value. This capability allows the designer to compute and specify dynamic range delimiters. |
|
Branch to: |
Question name to which the jump is made when the answer satisfies this Maximum/Minimum condition. |
|
Value: |
Value or the response placed in the answer file if this branch is taken. |
DCode or Distribution Code Field
See Pick List Field.
Free text field for any kind of notes or documentation that appears in the edit-design record. May be used as a note from one designer to his or her successor.
This field should no longer appear in Manager records beginning with version 2.1.53. The function of this field has been taken over by the FIELDS=() statement in the text of a Roster Edit question type. Formerly, in Roster Edit question types, the Field field contains the name of the field in the roster to be edited. If a field name was specified, the interviewer was be able to edit that field, only. The FIELDS=() statement is more flexible and allows for multiple fields to be edited.
|
This field is one of the primary fields in the record. This field determines the format of the response to the question. It anticipates the form of the datum and how it is stored. Different question types use different formats. To specify a data format in this field, the user must pick from a list of predefined formats. This list, the Format table, is displayed to the right. The Format field contains the text that appears in the drop-down menu of the Manager record, the ID is the assigned value of the text and used by the Manager and Survey programs.
FormTxt field __??The Format field may be used in conjunction with the Length field in the following question types: Evaluate Expression (Integer, 2), Fixed Answer (obsolete, Integer, 2), Keyword Codes (Integer, 2), Numeric Entry, Select Option, and Text Entry.
The Length field does not appear in the question record for the following question types: Date Entry, Dummy Record, Multiple Fields, Time Entry, Translate Answer (obsolete), Roster Edit, and Special Codes. The reasons are that Date and Time define their lengths by the format, rosters and multiple fields can collect data of various lengths, the former Translate Answer evluated a condition to 'true' or 'false', and no one knows the special requirements of the Special Codes question type.
The Length Explanation descriptions below attempt to explain the relationship between Format of the datum and its required Length. (See also the Length field.)
The HardMax and HardMin (hard maximum and hard minimum) fields define the range of acceptable answers for certain questions. The maximum defines the upper limit, and minimum the lower one. The range is valid only for questions that require a numeric answer or a date. Ranges are never specified for an alphabetic answer. These fields can be used in conjunction with SoftMax and SoftMin fields. (See SoftMax and SoftMin fields.)
Any question type that uses a Format specification of Integer, Long Integer, Floating Point, Date or Time is subject to range constraints. Several question types may include numeric answers:
Two special responses, Don’t Know (value -2) and Refusal (value -1), return integer values, but they override all range constraints for any type of question
Range specificationsin HardMax and HardMin fields and the SoftMax and SoftMin fields can be either static range checks or dynamic range checks. The two forms are distinguished below:
|
Static |
or absolute, are expressed in absolute terms such as integers, dates, times; |
|
Dynamic |
or relative, are expressed in terms of the contents of a symbol. (See SaveSym field.) This capability permits the computation and specification of valid ranges during execution of the survey, e.g., date-of-birth, refers to the value that is stored in the symbol %date-of-birth%. The fixed-symbol 'curdate' without the quotes is used by the Manager to refer to the current date and always contains the current date. When refering to symbols in the text of questions, the symbol's names must be enclosed in percent signs (% %), e.g., %date of birth% or %curdate%). |
Help Field
This field is for context-sensitive help on-line. Its purpose is to make help available to interviewers for any question that may require it. To make help available, place the name of a help record in this field so that the interviewer can have ready access by pressing <F1> Help. The context-sensitive help record must reside in the Help file that is associated with the keyword HELP= in the Vbimport.ini file. To insert a record in the Help field, identify it by highlighting its name in the drop down menu and click on the highlighted name or type in a valid help record name.
When this field is checked the answer the respondent gave will be hidden to the interviewer if the question is revisited either by going back to a previous question or by "sliding" back more than one answer and then moving forward again and revisiting the question.
Check box that when checked indicates that the question or roster item is an information sheet item. If checked, it will then become an information sheet item in the next round, if any. Primarily useful for longitudinal studies.
Jumps Field
See Cond. Junp Field.
This 40 character field is used to specify a SAS label for the variable collected. The default is the first 40 characters of the text field.
Length Field
This field is used to specify the length of the data type indicated in the Format field. (See Format Field). Depending on the format of the data type indicated in the Format field, the length of the data may vary. The Length field does not appear in the question record for the following question types: Date Entry, Dummy Record, Multiple Fields, Time Entry, Translate Answer (obsolete), Roster Edit, and Special Codes. The reasons are that Date and Time define their lengths by the format, rosters and multiple fields can collect data of various lengths, the former Translate Answer evluated a condition to 'true' or 'false', and no one knows the special requirements of the Special Codes question type.
The Length Explanation descriptions below attempt to explain the relationship between Format of the datum and its required Length. (See also the Length field.)
This field links the current record to the original record from which the current record was derived. This pertains mostly to longitudinal studies, where the same or a very similar question is posed in subsequent rounds.
Used with the User Guide Entries field to create a hyperlink to the manual's explanation.
Special feature used with the Select Option question type and in combination with the Pick List field. This feature is used to limit the focus of very long pick lists, e.g., cities within a state, province, or county; names of organizations within a category.
Normally the name of a pick list file is placed in Pick List field and that file is displayed during the interview from which the interviewer must select one or more choices. When this field is used, the name of a pick list file is placed in the Lookup field and a symbol is placed in the Pick List field. The contents of the symbol is a variable that is used to limit the contents of the pick list. For example, the Lookup field contains a the name of a file that contains all the cities in the U.S organized by state, e.g., CITIES. The Pick List field contains the symbol %STATE% that was used in a previous question to collect the name of the state of the respondent. The subsequent file that will be displayed during the interview will contain only the list of the cities from that state.
The Mark field is used to signal that one or more responses should be chosen from a "pick list". This field is used only with Select Option question type and in conjunction with the Pick List field. See Pick List Field. This field is restricted the following specifications that must be selected from the drop down menu:
|
Save Text |
Response will be considered to have textual or character value. A number will not be considered to have numeric magnitude. |
|
Save Value |
Response will be considered to have numeric magnitude and not a character value. |
|
Mark All that Apply |
Select all applicable responses from a pick list and save their values. |
|
Verbatim, Save Text |
Record verbatim response and select one response from a pick list whose text will be saved. |
|
Verbatim, Save Value |
Record verbatim response and select one response from a pick list whose value will be saved. |
|
Verbatim Mark All |
Record verbatim response and select all applicable responses from a pick list and save their values. |
When one of the Verbatim options is selected the Survey program will display a verbatim window along with the Pick list.
See Cond. Jump Field.
See HardMax and HardMin Field and SoftMax and SoftMin.
Contains the name of the default next question record in the logical chain. This question will be the next in the chain, if no conditions and conditional skips are specified in the Cond. Jump Field field. If there is a value of NIL is in the NextQ field, it indicates the last question record in a chain.
Depending on the question type chosen in the QType field, there may be conditions when the default next question specified in the NextQ field is overridden. In the Select Option, Translate Answer, Evaluate Expression, and certain Keyword Code question types, conditions and skip instructions may be stated in the Cond. Jump field that, if fulfilled, will cause the logic to branch to a question other than the one stated in the NextQ field.
Pick List Field (a.k.a. DCode Field)
The Pick List field contains the record names of distribution code records in the Pick List. These records are linked to the question record via this field. The appropriate record name or a %symbol% containing the DCODE record name is placed in this field. The Pick List field is used in conjunction with the Mark Type field. (See Mark Field.)
Distribution code records and continuos code records may be viewed from any question record at any time.
The complete record will appear in a window on the screen with its numerical values, textual representations, audio file links, and sequential ordering. See the figure below.
To see the complete record, use the scroll bar, if applicable.
To browse among the records in the file, use the arrow buttons.
button.
To exit and return to the question record display, press the <X> button in the upper right corner.
To insert a record in the Pick List field,
Distribution values that become part of the answer file, can be used to determine skips in Evaluate Expression and Translate Answers, used as conditions in some Keyword Codes, and used as range restrictions. These values can also be used to drive value labels for SAS or SPSS. Other uses for distribution codes is at the discretion of the designer.
The format of the expected Pick List value must be compatible with the data in the Format and Length and Mark fields.
This field sets a flag, usually temporary, that when checked, the data dissemination software will use to create a starting set of questions for a codebook.
This parameter, when checked pulls an answer into the answer field from the Save Array location specified in the SaveSym (save symbol) field. The desired element must have already been stored in the Save Array location. The field is used to supply an answer for verification, review, or updating purposes. Thus if the interviewer asked the age of a household member and the age were pre-loaded from the information sheet to the Save Array, the interviewer would just have ask if the information was correct, correct it if necessary, and press <Enter>to store the information both as answer to the question and update the Save Array. The program would then go on to the next question.
This field is used to assign each question in the survey a unique name that is used to identify the question. The unique name serves as a reference label to the question in branching and skip instructions elsewhere in the questionnaire. Indices are built from the question name. The question name must be assigned by the designer and has few restrictions. However, a simple naming convention for this field can save problems later on when identifying questions, sections of the questionnaire, iterations of the question, and subject matter.
Divide the name into parts. For example, a section or subject name, followed by an ordinal number. Each iteration of repetitive questions may be added to the name using a hyphen (-), underscore (_) and an ordinal number. It helps the designer if the section name can be mnemonic in some way. Question names are not case sensitive. Some examples are given below. Do not use a period (.) in the question name.
| Sample Format | [Section_name#]-[question#].[iteration#1].[iteration#2].[iteration#n] |
| Examples | QSTART |
| Q-END1 | |
| Q4-8a question 8a in section Q4. | |
| HHI1-1question number 1 in section HH1(household roster) no. 1. | |
| BIOCHLD1-1 question number 1 in biological child roster no.1. | |
| ES3-2a question 2a in section ES3 or the 3rd Employer Supplement. |
Restrictions Names cannot begin with: %, or $. These character are reserved for symbols and substitutions respectively. Names cannot have a leading or imbedded space.
This field contains a list of the various question types that can be specified in the Manager and are recognized by the Survey program. Each question type has a specific format. The following types may be specified: Calendar, Date Entry, Dummy Record, Evaluate Expression, Fixed Answer (becoming obsolete), Keyword Codes, Multiple Fields, Numeric Entry, Roster Edit, Select Option, Text Entry, Time Entry, Translate Answer (becoming obsolete), and Special Codes. Each of these question types is discussed in detail in the section entitled "Question Types."
Record Position field. This field sequentially locates the position of the question record in the file. The Question Name field identifies a record uniquely, this field locates it sequential position.
R Number field. This field contains the unique number for every variable in a file. Can be used to link variables in previous rounds of a longitudinal survey for final documentation and data retrieval programs.
In Roster Edit question types, the Roster field contains the name of a roster to be displayed for editing in the Survey program. The roster name must be one of those specified in the Roster database.
|
To select a roster and place it in the Roster field, use the drop down menu, highlight the roster name, and click on the roster name or type in a valid name .
To view the roster record named in the Roster field, right-click on the name in the field and a dialogue box will appear requesting whether the user wants to:
Choose Display and a window similar to the one "Roster Record" window will appear.
The roster record may be edited from this window. A new roster record may also be created from this window.
To edit a Roster field, place the cursor in the field and click the left mouse button.
To create a roster, press the button. A blank roster record template will appear.
Save Symbol field. This field is used to indicate that the answer to the current question is to be saved in the Save Array at the symbolic location specified by the name in the SaveSym field. Usually an answer saved in this manner is referred to later in the text field of a subsequent question.
To view symbols in the symbol list, go to the Main Menu screen, select the Edit tab, and press the Symbols button.
Using Symbols
Symbols allow the user to refer to variables dynamically. The Manager-Survey allows robust use of symbols. A datum can be collected or created and stored at a symbolic location to be used later as a substitution, reference, counter, and/or index. Symbols and loops may be combined to create an indexed substitution or reference to variables in nested loops. The only limitations are the designer’s proficiency and the variables’ referenced in nested loops can only go five deep.
Simple
A symbol may be used as a simple text substitution in a question: i.e., the name of the respondent’s employer.
"You said you worked for ‘employer’s-name’ last year."
The symbol’s syntax requires that it be enclosed in percent signs, % -% , e.g.:
"You said you worked for %employer’s_name% last year."
Indexed
Symbols may be indexed for textual substitution: i.e., the names of members in a household.
"You said ‘household-member’s-name-n’ highest completed level of schooling was…"
Sample symbol’s syntax:
"You said %HHI.name()%’s highest level of schooling was…"
The parentheses, (), imply that this symbol is referenced in a loop that will repeat the question and substitute each of the qualified member’s names in the household. The current value of the control variable (e.g., loop counter) is used to differentiate among the names, e.g., %HHI.name(1)%, %HHI.name(2)%, %HHI.name(3)%,…%HHI.name(n)%
Imbedded
Symbols may be imbedded within symbols to index substitutions in nested loops, to refer to symbols created in other loops, and to refer to different iterations of nested loops.
A symbol is imbedded within another symbol by using double percent signs, %% - %%, around the outer symbol’s name. One instance might be if the question is part of a nested loop. The outer loop is processing household members. The inner loop is asking for employers for each person in a household. One person may have (had) multiple employers. To reference the person's name in the question text of the INNER loop, the syntax for a substitution would look like:
Was %%HHI.name(%loop-1%)%% working for %employer()% last year?
|
%HHI.name(%loop-1%)% |
is an indexed symbol for each member in a household. The names for each individual are stored as %hhi.name(1)%, %hhi.name(2)%, %hhi.name(3)%, … %hhi.name(n)%, etc. |
|
(%loop-1%) |
is a symbol that indexes each member in the household (a.k.a. loop counter). The loop counters can be called anything at all. This one might also be called the %name_loop_counter% for clarity. |
|
% % |
the extra % % surrounding the symbol indicate an imbedded symbol. |
It is important to know WHERE a symbol was created and WHICH index to use or an error message will occur in the Survey during execution.
Special Feature for Symbols
Imbedding within nested loops may be carried out to five nested loops. The use of a colon, ':' , in the index allows the user to index/reference different iterations of nested loops i.e. symbols created in outer loops...
Imbedded symbols from different iterations of nested loops may be referenced by using double percent signs, %% -%%, and a colon, ":", between elements of the index. For example, if it is necessary to refer to the 2nd person's 4th employer, or %employer(2:4)%, the general syntax might look like:
%%symbol1(%loop-counter-#1%:%loop-counter-#2%)%%;
The specific syntax to pull in a particular loop's value would make reference to:
%%employer(%loop-1%:%loop-2%)%%
It is also permissible to refer to a symbol with the straight numbers of the indices, e.g., %employer(2:4)%. The (loop#1:loop#2:loop#n) reference may be carried out to five deep, i.e., (loop1:loop2:loop3:loop4:loop5).
With this feature it is extremely important to know WHERE a symbol was created and WHETHER it has one index, or two, or more. If the Survey cannot find the symbol, and an error message will appear during execution.
This field allows the user to form groups of questions within an instrument. There may be more than one section or module in an instrument.
The HardMax and HardMin fields cover two absolute limits whose range normally cannot be exceeded. The SoftMax and SoftMin fields cover two relative limits whose range contains more "reasonable" bounds within the absolute limits. Values of all maximum and minimum fields may be expressed integers or symbolic references.
The SoftMax and SoftMin fields are used to check these uncommon answers and when the check fall between the uncommon yet reasonable, a message appears on the interviewer’s screen to verify the answer of the respondent. At this point the interviewer may change the response or verify that it was, in fact, the correct response.
The use of these fields depends upon the latitudes of the designer. The following example demonstrates the use of these fields.
The absolute range would be intact, but the answer would be less than the soft minimum, so a message would appear for the interviewer to verify the response.
This field when checked determines whether the "Fast Forward" feature of the Survey program will stop at this question record even if it has already been answered. The Fast Forward feature is used to quickly step through the questionnaire after an interviewer has gone back to change an answer to a previous question. In order to preserve the integrity of the branching patterns every question between the question that was changed and the spot where the interview was interrupted must be reviewed. All of these questions already may have been answered but because one question was changed a whole new branching pattern may be implied. In "fast forward" mode the program flashes questions on the screen very quickly, but some questions are crucial, needing the interviewer’s attention. Such questions are roster recapitulations and certain check questions in the employer names section. Whenever a question record needs interviewer attention even while running through the survey a second time, the Stop Forward box on the record must be checked. To continue past this question the interviewer will be required to press <F9>to resume fast forward feature.
The field of the question record is never identified by a visible field label as the other fields are. This field contains the text of the question and the documentation to the question. The largest block of white space is devoted to the text and documentation of the question.
The Text field has several uses in a survey. In most cases it contains the question itself, but depending on the question type selected in the Qtype field, the Text field can be used in a number of ways. It can be used to substitute text into a question, to pass an answer on, to instruct the Survey program to carry out a computation or interviewer to carry out a directive, or to simply document the record. For more details on the question types, see "QType Field".
Requirements of the Text field are discussed in this manual under each of the specific question types in the section entitled "Question Types". If a question requires textual variations within its wording, such as the substitution of a spouse’s or child’s name, an answer to a previous question, or a number of weeks worked in a given time period, the Text field can be used to carry them out.
Examples
Several examples of substitution in the question Text are given below. Suppose a question asks "Did your (husband/wife) do any work last year?" The respondent is a female, so "husband" would be the appropriate substitution. The question text in the Manager program for this substitution would then look like:
Did your %husband/wife% do any work last year?
The percent signs (% _ %) delineate a substitution parameter whose value would come from a symbol in the Save Array. In the survey instrument the text for the interviewer would read either
"Did your husband do any work last week?" or
"Did your wife do any work last week?"
The designer may also want to substitute an answer to a question "How many weeks did %he/she% work last year?", in a subsequent question. The subsequent question in its Manager record would then read:
"You said %he/she% worked %Q5-7a% weeks last year."
The substitution parameter, %Q5-7a%, indicates a symbol in the Save Array where the response to question Q5-7a was stored. To the interviewer the statement would now read something like
"You said she worked 26 weeks last year." or
"You said he worked 26 weeks last year."
This 80 character field is used to assign the variable a key-word-in-context (KWIC) title based on a pre-defined keyword list. The KWIC titles are used by data retrieval software as access points to variables. CHRRDBA and CHRR NLS Investigator are, respectively, DOS-based and windows-based data-retrieval programs that provide this access point to the researcher.
This field can contain a reference to a user guide entry for this variable or particular question. The user guide may be a manual that introduces researchers and users to core sets of variables or contains detailed information on components used in multiple rounds of a survey. This is used with the Links field to create a hyperlink to the manual's explanation.
The Varname field is a suggested name to use when creating the SPSS cards.
This field may contain a reference to a year in which the variable appears.
Next Steps
See Question Types, as specified in the QType field above.