Multiple responses -single response was expected

[Salbod]

Good Afternoon,

A few respondents (n = 4 out of 82) circled (on  12 item questionnaire) adjacent response choices, such as “3. Agree” and “4. Strongly Agree”. What code should be used 3?, 4?, 3.5? or a missing value code? Does anyone know a good reference for this situation?

Any help will be greatly appreciated.

Thank you,

Stephen Salbod, Pace University, NYC

 

[Art Kendall]

The response of these 4 people underlines the discussion on this list this morning about jittering when graphing variables with a few ordered or interval categories.  Respondents often think of the measure as continuous.

What I would suggest doing would depend.

Are you summing or averaging the items into a summative scale?

Did these 4 people give any other responses to some items?

Were the item stems balanced for direction?


I don't know of a reference, but have been working with Likert scale since 1971.

On 1/30/2013 12:20 PM, Salbod, Mr. Stephen wrote:

Good Afternoon,

A few respondents (n = 4 out of 82) circled (on  12 item questionnaire) adjacent response choices, such as “3. Agree” and “4. Strongly Agree”. What code should be used 3?, 4?, 3.5? or a missing value code? Does anyone know a good reference for this situation?

Any help will be greatly appreciated.

Thank you,

Stephen Salbod, Pace University, NYC

 

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

[Salbod]

Thanks Art,

They’re being averaged.

They completed the 11 other items with single responses.

The items weren’t balanced.

When I did coding in a Market Research house I remember alternating between the numbers. I might have flipped a coin J --Steve

 

From: Art Kendall [mailto:[hidden email]]
Sent: Wednesday, January 30, 2013 12:53 PM
To: Salbod, Mr. Stephen
Cc: [hidden email]
Subject: Re: [SPSSX-L] Multiple responses -single response was expected

 

The response of these 4 people underlines the discussion on this list this morning about jittering when graphing variables with a few ordered or interval categories.  Respondents often think of the measure as continuous.

What I would suggest doing would depend.

Are you summing or averaging the items into a summative scale?

Did these 4 people give any other responses to some items?

Were the item stems balanced for direction?


I don't know of a reference, but have been working with Likert scale since 1971.

Art Kendall

Social Research Consultants

On 1/30/2013 12:20 PM, Salbod, Mr. Stephen wrote:

Good Afternoon,

A few respondents (n = 4 out of 82) circled (on  12 item questionnaire) adjacent response choices, such as “3. Agree” and “4. Strongly Agree”. What code should be used 3?, 4?, 3.5? or a missing value code? Does anyone know a good reference for this situation?

Any help will be greatly appreciated.

Thank you,

Stephen Salbod, Pace University, NYC

 

 

[John F Hall]

Give them an out of range code (say 35) and declare it as missing.  Later on you can assign them to an actual code (modal value for all their individual ratings?).  Alternatively allocate two to 3 and the other two to 4 (using random choices) or treat all four cases as illiterate and delete them altogether.  Why waste your time, and ours?

John F Hall (Mr)

[retired academic survey researcher]

 

Email:     [hidden email]

Website: www.surveyresearch.weebly.com

 

 

 

 

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Salbod, Mr. Stephen
Sent: 30 January 2013 18:20
To: [hidden email]
Subject: Multiple responses -single response was expected

 

Good Afternoon,

A few respondents (n = 4 out of 82) circled (on  12 item questionnaire) adjacent response choices, such as “3. Agree” and “4. Strongly Agree”. What code should be used 3?, 4?, 3.5? or a missing value code? Does anyone know a good reference for this situation?

Any help will be greatly appreciated.

Thank you,

Stephen Salbod, Pace University, NYC

 

[Art Kendall]

If you are averaging the items within the person, I would just use 3.5 etc. as the value.
If you are going to share the data add a value label so that the metadata is complete.




On 1/30/2013 1:12 PM, Salbod, Mr. Stephen wrote:

Thanks Art,

They’re being averaged.

They completed the 11 other items with single responses.

The items weren’t balanced.

When I did coding in a Market Research house I remember alternating between the numbers. I might have flipped a coin J --Steve

 

From: Art Kendall [[hidden email]]
Sent: Wednesday, January 30, 2013 12:53 PM
To: Salbod, Mr. Stephen
Cc: [hidden email]
Subject: Re: [SPSSX-L] Multiple responses -single response was expected

 

The response of these 4 people underlines the discussion on this list this morning about jittering when graphing variables with a few ordered or interval categories.  Respondents often think of the measure as continuous.

What I would suggest doing would depend.

Are you summing or averaging the items into a summative scale?

Did these 4 people give any other responses to some items?

Were the item stems balanced for direction?


I don't know of a reference, but have been working with Likert scale since 1971.

Art Kendall

Social Research Consultants

On 1/30/2013 12:20 PM, Salbod, Mr. Stephen wrote:

Good Afternoon,

A few respondents (n = 4 out of 82) circled (on  12 item questionnaire) adjacent response choices, such as “3. Agree” and “4. Strongly Agree”. What code should be used 3?, 4?, 3.5? or a missing value code? Does anyone know a good reference for this situation?

Any help will be greatly appreciated.

Thank you,

Stephen Salbod, Pace University, NYC

 

 

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD