This tutorial reflects how to job-related with the data from "check-all-that-apply" multiple choice survey concerns in SPSS Statistics making use of multiple response sets.
Getting Started through SPSSToggle DropdownWorking with DataToggle DropdownExploring DataAssessing DataToggle Dropdvery own

Our tutorials referral a datacollection referred to as "sample" in many kind of examples. If you"d like to download the sample dataset to work with the examples, choose among the records below:


This tutorial is a primer on just how to occupational via information from multiple option, multiple-response (or "inspect all that apply") questions in SPSS Statistics.

You are watching: Which of the following give the nominal value of a variable? check all that apply.

Jump to:

Multiple response sets take place once you have a collection of associated selections or attributes in which a topic or experimental unit deserve to possess one or more of those characteristics. In this tutorial, we will focus on a particular kind of multiple response set: multiple response (or "check-all-that-apply") questionnaire items.

A multiple response question presents a list of possible answer alternatives, and the respondent selects all choices that are true for them. For example, suppose we are interested in surveying a group around what kinds of electronic devices they own, and also intend we are especially interested in the 3 most prevalent forms of mobile computer devices: laptops, phones, and also tablets. We could develop a survey question like this one:

*


If you are provided the alternative between these two structures, the multiple-column system is strongly preferred. If your information is videotaped making use of the single-column structure, you will certainly must "clean up" the data to obtain it into the one-column-per-selection format.


To appropriately analyze multiple response questions in SPSS, your datacollection should have the adhering to structure:

Each row (case) should represent one topic, survey response, or speculative unit.For a given multiple response question, each answer alternative must be represented in a separate column (variable).The information values have to follow one of these 2 schemes:Numeric code (frequently 1) if existing, blank (missing) if not existing.Numeric codes representing present and not present (such as 0=Absent out, 1=Present).

The complying with 2 examples show both schemes making use of the same underlying data. In these examples, the columns reexisting the answers to a check-all-that-apply question, "Which of the following devices perform you own?", with four answer options: lapoptimal, phone, tablet, or "other". In ordinary language, the information provided in both examples:

Cases 1, 2, and also 5 very own a lapheight and a phone and a tablet.Cases 3, 4, 6, and 8 very own a lappeak and a phone.Case 7 owns only a phone.

Numeric code if current, empty if absent

In this coding scheme, we have a distinct numeric code representing the "checked" or "present" state, but usage a lacking worth (blank) to reexisting the "unchecked" or "absent" state.

*

In this instance, 0 denotes "absent" or "not checked", and 1 denotes "present" or "checked". You carry out not necessarily have to use the numbers 0 and also 1, yet you have to use the exact same numeric codes across every one of the columns. Value labels are strongly recommfinished, so that you can remember the definitions of the codes.


In this tutorial, we will certainly be making use of simulated data from a theoretical survey through 2 questions. Neither question was required, so respondents can choose to skip one or both concerns.

*

From this table, we can view that six (6) respondents did not choose any digital gadgets. Keep this number in mind once reviewing the Multiple Response Frequencies output in the following instance.

This table additionally tells us:

Thirty-4 (34) respondents, or 7.8% of the sample, own a single electronic gadget.Two-hundred forty (240) respondents, or 55.2% of the sample, own 2 digital tools.One-hundred forty-3 (143) respondents, or 32.9% of the sample, own three digital gadgets.Twelve (12) respondents, or 2.8% of the sample, own four digital devices (i.e., selected all 4 answer options).

To filter out people who did not answer the multiple response question, use the Select Cases procedure to store instances if schosen > 0 (schosen higher than 0).


SPSS has actually a two-action procedure to use multiple response sets making use of the dialog windows:

Define the multiple response set.Identify the variables representing the worths for that collection.Indicate which number code(s) must be counted as "present".Run the multiple response frequencies or crosstabs steps.

After a multiple response collection is defined, it is only preserved as long as the SPSS session is energetic. Once you close SPSS, the multiple response collection interpretation is erased; the following time you start SPSS, you would certainly must re-define the multiple response set if you wanted to re-run the multiple response frequency tables. (The exception to this is if you have the Custom Tables module <1>, which is not covered below.) The best method to prevent having to re-define your multiple response sets is to save the syntaxation created by the Multiple Response Frequency Tables and also Crosstabs procecdures in a SPSS syntaxes file, because the syntaxation for these measures automatically consists of the interpretations of the response sets. (This is spanned in both examples later in the tutorial.)


Using Dialog Windows

Step 1: Define Multiple Response Set

To specify a multiple response collection with the dialog windows, clickAnalyze > Multiple Response > Define Variable Sets.

*

*

A Variables in Set: The variables from the datacollection that compose the multiple response set. For surveys, this is commonly the collection of columns equivalent to the "selectable" options for a single survey question.

B Variables Are Coded As: The information values used to suggest that the category was existing.

Dichotomies: Use if a single numeric value was supplied throughout every one of the variables to show if the category was "present".Categories: Use if there was a range of number codes used to indicate if the category was existing, or if tright here is more than one category that will be counted as "present".

Note that it is just possible to select among these schemes. If your data does not enhance among these schemes, you may require compute recoded versions of the variables utilizing the Recode into Different Variable procedure.

C Name and Label: The name (required) and also label (optional) of the multiple response collection. The naming rules for multiple response set names are the very same as the normal variable naming rules in SPSS (no spaces, must start with a letter).

D Multiple Response Sets: List of all response sets that have been characterized in the current SPSS session. This panel will certainly be empty if no response sets are characterized.


Step 2: Multiple Response Frequencies

After establishing up a multiple response set, you will certainly be able to access the Multiple Response Frequencies choice through the menus. To carry out this, click Analyze > Multiple Response > Frequencies.

*

All multiple response sets you've characterized during the current SPSS session will certainly show up on the left.

The two alternatives in the Missing Values section regulate exactly how instances with lacking values need to be treated. These settings will have different effects, relying on whether you usage blanks versus numeric codes to recurrent unschosen options, and also whether you specified a dichotomy or a variety of category codes in the previous step:

The Exclude situations listwise within dichotomies choice will certainly treat cases via any kind of absent worths as fully absent. If you coded your selected worths as 1 and empty if not selected, this certain option will only count instances where all values were current. This specific option need to only be provided if you coded selected worths as 1 and unschosen values as 0 (or some other nonlacking numeric code).The Exclude situations listwise within categories choice will certainly just think about a case as "missing" if it does not have actually at leastern one variable with the stated number code.

To avoid having actually to re-define the exact same response set, we recommend utilizing the Paste button (rather of the OK button) to geneprice the command syntax code for the multiple response frequency table or crosstab. This is bereason the syntaxes command for multiple response sets, MULT RESPONSE, contains the meaning of the collection in the command also. (This will be shown in the example listed below.) Using the Paste button will certainly create the syntax regulates to the syntaxation home window, which you have the right to then use to execute the analysis without needing to go via the dialog home windows.


Step 3: Multiple Response Crosstabs

After setting up a multiple response collection, you will certainly have the ability to accessibility the Multiple Response Crosstabs alternative through the menus. To carry out this, click Analyze > Multiple Response > Crosstabs.

*

A Variable list: The variables in the current dataset. Categorical variables in this list can be offered as Row, Shaft, or Layer variables. For each variable in this list that you usage in the table, you will should usage the Define Ranges button to tell SPSS which number categories you want to be contained in the table.

B Multiple Response Sets: The multiple response sets that have been characterized throughout the existing session. These variables can be used as Row, Tower, or Layer variables.

C Rows: The variable(s) you desire to be supplied as the rows in the crosstab.

D Columns: The variable(s) you want to be offered as the columns in the crosstab.

E Layers: The variable(s) you desire to be used as the "layer" variable in the crosstab. The categories of the layer variable will appear on the outera lot of edge of the table.

Sketch of the positions of row, column, and also layer variables in an SPSS crosstab.
Col 1Col 2
Layer 1Row 1
Row 2
Layer 2Row 1
Row 2

If multiple variables are gone into in the Row, Pillar, and/or Layer boxes, there will certainly be a sepaprice table for each unique combination of the row*column*layer variables.

F Define Range: Opens the Define Range prompt. This option becomes obtainable when you've included a continual variable to the Row, Pillar, or Layer box, and have actually clicked on the variable so that it's highlighted.

*

Note that this means that you cannot use string variables in these tables, and the numeric category codes you desire to encompass in the table have to be sequential (i.e., if you had actually categories 1=Disagree, 2=Neutral, 3=Agree, and also you only wanted to include categories 1 and also 3 in the table, you would certainly should recode the variable so that Neutral is not within the range.) If tbelow are numbers in between the minimum and also maximum that are not represented in the oboffered data, those numbers will certainly be ignored.

G Options: Opens the Options window:

*

Cell Percentages: Add percentperiods to the table cells (in addition to reflecting the counts).Row: Percenteras will certainly be based on the row complete.Column: Percenteras will certainly be based on the column complete.Total: Percenteras will certainly be based upon the table full.Percentperiods Based on: If one or more of the "cell percentages" options are schosen, controls the worths supplied in the denominators of those calculations.Cases: The marginal totals reexisting the number of cases in that team. When summing over the categories of a multiple response collection, the amount of the cells may not equal the marginal total.Responses: The marginal totals equal the amount of the cells in the table.Missing Values: Change how missing values are handled in the table. By default, a situation must have actually absent worths on all response set variables to be counted as absent. Applying among these settings will instead usage listwise lacking information handling; i.e., if a instance contends leastern one lacking value among the response collection variables, it will certainly be treated as missing. If your data uses the numeric/blank encoding plan, applying these settings will certainly primarily result in empty or mostly-empty tables.Exclude instances listwise within dichotomies: Applies just once the multiple response set meaning supplied dichotomies.Exclude situations listwise within categories: Applies only once the multiple response collection definition offered category code ranges.

<1> How do I conserve multiple response sets defined with the food selection system? - IBM

<2> IBM SPSS Statistics Knowledge Base.https://www.ibm.com/support/knowledgecenter/en/SSLVMB_26.0.0/statistics_mainhelp_ddita/spss/base/idh_mulc_opt.html


Problem Statement

Suppose we desire to know what types of electronic tools (laptops, smartphones, and also tablets) college students generally own. Our desired summary would certainly look something favor this:

Outline of desired summary table for a multiple-response question.Devices ownedn% of respondents (n=??)
Laptop
Phone
Tablet
Other

If we were to try to usage the continuous Frequencies procedure on this information (Analyze > Descriptives > Frequencies), the resulting tables would not be succinct:

*

The initially table reflects the number of valid and also missing responses for each variable. Notice the variety of absent responses for each variable: Due to the fact that we are using the scheme of 1=checked, missing=not checked, the lacking worths here actually reexisting the number of human being that did not select that alternative. It does not necessarily suppose that they did not answer the question! We have to only consider people that left all 4 options empty as skipping the question. It's not feasible to recognize exactly how many people left all four options blank from the standard Frequencies procedure.

In the individual frequency tables, we view the variety of people who checked that choice (in the rows labeled "Valid - 1"). The Percent column represents the propercent of the complete sample that checked that choice. Since this procedure can't determine if tright here were people who did not answer the question, we don't recognize for particular if we should use the complete sample dimension as the denominator to compute the percentperiods.

Instead, we need to usage the Multiple Response Frequencies procedure, which can address all of these problems, and create a table structured prefer the above.

Sample Dataset for This Example


Running the Procedure

Using the Dialog Windows

If utilizing the dialog home windows, we should do this in 2 steps: initially, making use of the Define Multiple Response home window, and also then making use of the Multiple Response Frequencies home window.

Open the Define Response Set window (Analyze > Multiple Response > Define Set).Highlight the four multiple response variables (click variable owns_lappeak, then host dvery own Shift and also click variable owns_other) in the left column. Then click the arrow switch to move them to the Variables in Set box.In the Variables Coded As area, in the box labeled Counted Value, form 1.In the Name box, kind a brand-new name for the set; in this instance, we'll use the collection name tools. In the Label box, form a descriptive label; in this instance, we'll use "Electronic gadgets owned".
*
Click Add to save your new collection.When finiburned, click Close.

After clicking Close, nothing will appear to happen; this is normal. To actually produce the table, we now run the Multiple Response Frequencies procedure:

Click Analyze > Multiple Response > Frequencies.In the left box, double-click the brand-new variable collection, devices, to move it to the appropriate box.Click OK.Using Syntax

Using syntaxes for multiple response frequency tables is much simpler: the interpretation of the collection and the command also to develop the frequency table are done in the exact same command:

MULT RESPONSE GROUPS=$tools 'Electronic tools owned' (owns_laptop owns_phone owns_tablet owns_various other (1)) /FREQUENCIES=$devices.

Output

Running the over procedures or syntaxes produces the complying with output:

*

The first table, Case Summary, counts the number of situations with valid and also "true" nonlacking values -- i.e., situations that did not have any type of of the choices checked. We view that only 6 situations did not choose any kind of of the answer choices. This matches what we saw from the Count Values Within Cases procedure (above).

The second table, $gadgets Frequencies, is the frequency table of interest. From left to right, the columns of this table show:

First column: The name or label of the multiple response set.2nd column: The variable names or variable labels (if assigned) of the variables in the multiple response set.N: The variety of instances who schosen that response alternative. Notice that these worths match the valid values in the frequency tables from the "basic" Frequencies procedure. The number in the Total row is the total number of selections.Percent: The propercentage of selections accounted for by this category. This column will always amount to 100%. You can confirm the worths in this column by splitting the N of that row by the Total N from the last row of the table (991).Percent of Cases: The propercentage of the cases (i.e., survey respondents) accounted for by this category. This column's sum will be greater than 100%, however the individual prosections can be taken as the pervasiveness of that alternative among the survey sample. This is frequently even more systematic than the Percent worth. You deserve to confirm the worths in this column by dividing the N of that row by the Valid N from the Case Rundown table (429).

Interpretation

Using the values of N and also Percentage of Cases from the multiple response frequency table, we have the right to fill in the table from the start of this example:

Completed summary table: electronic gadgets owned by college students.Devices ownedn% of respondents (n=429)
Laptop39792.5%
Phone38690.0%
Tablet16839.2%
Other409.3%

This table tells us that:

There were 429 students that responded to the question, i.e. selected at least one of the four tool form options.The large majority of the respondents owned a laptop (92.5% or 397/429)The substantial majority of the respondents owned a phone (90.0% or 386/429)Less than fifty percent of the respondents owned a tablet computer (39.2% or 168/429)Less than 10% shelp they owned some other type of electronic gadget (9.3% or 40/429)

Limitations

We observed that the Multiple Response Frequencies procedure will certainly treat an individual as "missing" (i.e. did not answer the question) if the individual had actually missing worths for all variables in the set. However before, our survey question just had actually four options -- lappeak, phone, tablet, and also "other". All of these options assume that the respondent owns an electronic device. If someone does not have an digital tool, the just means they deserve to accurately respond is to not choose any choices! This means that we can't differentiate between civilization who don't very own any type of digital gadgets and also people that skipped the question. Given our original research question, this would certainly be particularly problematic: if we are interested in understanding the digital tools that college students very own, we need to be particular around what proportion of students carry out not very own any kind of devices, since that can influence students' access to digital course products.

How can we prevent this trouble when creating future surveys? One alternative is to include an answer option that would certainly especially accommoday people who don't own an digital gadget. If we carry out this, we will must take an additional step to proccasion respondents from providing inconsistent answers: for example, we don't want to allow the option for someone to answer "I very own a phone and I don't own any electronic devices". Some online survey platforms (such as Qualtrics) allow the survey designer to designate certain answer choices as "exclusive". Answers noted as "exclusive" will be "either-or": you deserve to choose any kind of and every one of the non-exclusive options, or you deserve to choose the exclusive option, but not both simultaneously.

See more: 4 Of The Best Places To Live In Dc For College Grads, 2021 Best Neighborhoods For Young Professionals

Remember: a great multiple alternative question will certainly have answers that expectations the full variety of possible answers. (This is true for both single-choice and also check-all-that-use question types!) Consider your research study question, and also usage it to overview whether you need to incorporate an option prefer "other", "not applicable", or "namong these".