organizing output

[Cleland, Patricia (EDU)]

I’m analysing the results of a survey that was completed by about 100 agencies and am looking for help in how to display the output. 

 

Here’s the background: Most of the agencies describe themselves as offering a single program. A small group describe themselves as offering two separate programs, e.g., separate programs for males and females. Another small group describes themselves as offering four programs, e.g., separate programs for males and females by age group (that is, separate programs for adult males, youth males, adult females, youth females).  

 

All agencies answered the same basic questions, but for the two- and four-program agencies each question was presented two or four times respectively. The survey, which was administered via SurveyMonkey, had skip logic which directed the agencies to the questions appropriate for their program type.  One-program agencies answered questions 4-29; two-program agencies answered questions 30-54 and four-program agencies answered questions 55-79. Thus, each question was present in the survey 7 times (1 plus 2 plus 4).

 

I have analyzed the results separately for the program types, that is, created a set of frequency tables for the one-program agencies, a set of tables for the two-program agencies and tables for the four-program agencies. 

 

My client has indicated that they would prefer that I present a single table for each question that includes the data for all the program types side by side, such as:

 

Response

One-program

Two-Program

Four Programs

Males

Females

Adult Males

Youth Males

Adult Females

Youth Females

# of agencies

% of agencies

# of agencies

% of agencies

# of agencies

% of agencies

# of agencies

% of agencies

# of agencies

% of agencies

# of agencies

% of agencies

# of agencies

% of agencies

Option A

Option B

Option C

Option D

Total Number of Agencies

 

 

The only way that I can think to do this is to export all the tables to excel with each program type on a separate tab and then either cut-and-paste the tables so that they’re side-by-side or create formulae in excel to populate the single table. Is there any (automated) way that I can create these tables within SPSS? I have a total of 8 of these tables to prepare.  I’m using v19.

 

Any help would be appreciated.

Pat

[Maguin, Eugene]

Patricia,

My first thought is to restructure your dataset but there is something I
don't understand about your survey. Based on your table, I'm picturing a
fixed set of questions, for example, q1 to q12, that are answered for each
program. One program agencies answer the questions once, two program
agencies answer them twice, once for each program, etc. So a one program
agency has valid data for q1-q12 and missing for q13-q36 and q37-q84. That
structure could be restructured to this

Agency Type program q1-q12
 10     1     1      .....
 12     2     1      .....
 12     2     2      .....
 67     4     1      .....

Then, I think you could use ctables to produce what is really a
crosstab/means report structure.

Your question numbering, however, doesn't seem to correspond to this
structure. What am I misunderstanding?

Gene Maguin


________________________________

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Cleland, Patricia (EDU)
Sent: Tuesday, November 30, 2010 9:38 AM
To: [hidden email]
Subject: organizing output

I'm analysing the results of a survey that was completed by about 100
agencies and am looking for help in how to display the output.

Here's the background: Most of the agencies describe themselves as offering
a single program. A small group describe themselves as offering two separate
programs, e.g., separate programs for males and females. Another small group
describes themselves as offering four programs, e.g., separate programs for
males and females by age group (that is, separate programs for adult males,
youth males, adult females, youth females).

All agencies answered the same basic questions, but for the two- and
four-program agencies each question was presented two or four times
respectively. The survey, which was administered via SurveyMonkey, had skip
logic which directed the agencies to the questions appropriate for their
program type.  One-program agencies answered questions 4-29; two-program
agencies answered questions 30-54 and four-program agencies answered
questions 55-79. Thus, each question was present in the survey 7 times (1
plus 2 plus 4).

I have analyzed the results separately for the program types, that is,
created a set of frequency tables for the one-program agencies, a set of
tables for the two-program agencies and tables for the four-program
agencies.

=====================
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

[John F Hall]

Bit of a hairy one, this.  I need to think about it, but ViAnn, Bruce or Jon may have an answer.  You should be able to create a new variable with 7 values representing the 1,2 and 4 groups and then do something with CTABLES using FREQUENCIES for your Options variable.  I suspect some sort of macro might be needed, but that's outside my competence.

 

Jon sent me some advice on how to use the pivot table menu to produce what I wanted (I'd completely missed one of the buttons) and also some syntax to solve a slightly different problem. 

 

CTABLES
  /TABLE BY hlth1  + hlth2+hlth3+hlth4+hlth5
  /CLABELS COLLABELS=OPPOSITE.

I had already one that looked like this.

 

TABLES
  /FORMAT BLANK MISSING('.') /TABLES
  (LABELS)  BY
  ( v16 + v17 + v18 + v19 + v20 )
  /STATISTICS COUNT ((F5.0) 'Count' ) .

I'll get back to you off-list with the problem I had and examples of the output, to see if it's anything like what you need.

 

John Hall
[hidden email]
http://surveyresearch.weebly.com

----- Original Message -----

From: [hidden email]

To: [hidden email]

Sent: Tuesday, November 30, 2010 3:38 PM

Subject: organizing output

I’m analysing the results of a survey that was completed by about 100 agencies and am looking for help in how to display the output. 

 

Here’s the background: Most of the agencies describe themselves as offering a single program. A small group describe themselves as offering two separate programs, e.g., separate programs for males and females. Another small group describes themselves as offering four programs, e.g., separate programs for males and females by age group (that is, separate programs for adult males, youth males, adult females, youth females).  

 

All agencies answered the same basic questions, but for the two- and four-program agencies each question was presented two or four times respectively. The survey, which was administered via SurveyMonkey, had skip logic which directed the agencies to the questions appropriate for their program type.  One-program agencies answered questions 4-29; two-program agencies answered questions 30-54 and four-program agencies answered questions 55-79. Thus, each question was present in the survey 7 times (1 plus 2 plus 4).

 

I have analyzed the results separately for the program types, that is, created a set of frequency tables for the one-program agencies, a set of tables for the two-program agencies and tables for the four-program agencies. 

 

My client has indicated that they would prefer that I present a single table for each question that includes the data for all the program types side by side, such as:

 

Response

One-program

Two-Program

Four Programs

Males

Females

Adult Males

Youth Males

Adult Females

Youth Females

# of agencies

% of agencies

# of agencies

% of agencies

# of agencies

% of agencies

# of agencies

% of agencies

# of agencies

% of agencies

# of agencies

% of agencies

# of agencies

% of agencies

Option A

Option B

Option C

Option D

Total Number of Agencies

 

 

The only way that I can think to do this is to export all the tables to excel with each program type on a separate tab and then either cut-and-paste the tables so that they’re side-by-side or create formulae in excel to populate the single table. Is there any (automated) way that I can create these tables within SPSS? I have a total of 8 of these tables to prepare.  I’m using v19.

 

Any help would be appreciated.

Pat

[Cleland, Patricia (EDU)]

Sorry, I wasn't clear in how the data is organized.  Let me give you an actual question.

"If the program is offered in more than one location, how are clients assigned to the location?"

Here's how the responses are presented to the one-program agencies. The codes in brackets following the options are the variable names. (Of course, they don't appear in the actual survey.)

a. Site nearest client's home (q7_2a)

b. Site that best meets the client's needs (q7_2b)

c. Other (q7_2c)

For two-program agencies, the question is Q33_2 and the response items are presented in a matrix. I've included the variable names in the matrix.

        Males   Females
Site nearest client's home      Q33_2aM Q33_2aF
Site that best meets the client's needs Q33_2bM Q33_2bF
Other   Q33_2cM Q33_2cF

For four-program agencies, the question is q59_2 and the responses are again presented in a matrix. Again the variable names are included in the matrix for your information.

        Males-Adult     Males-Youth     Females-Adult   Females-Youth  
Site nearest client's home      Q59_2aMA        Q59_2aMY        Q59_2aFA        Q59_2aFY       
Site that best meets the client's needs Q59_2bMA        Q59_2bMY        Q59_2bFA        Q59_2bFY       
Other   Q59_2cMA        Q59_2cMY        Q59_2cFA        Q59_2cFY       

The survey was structured this way to minimize the burden on the agency staff that were completing the survey.

I have frequency tables for each of the questions above: one for q7_2, two for q33_2 (separately for males and females) and four for q59_2, one for each of the programs.

Here's the table that my client wants to see:

Response        One-program     Two-Programs    Four Programs  
                Male    Female  Male-Adult      Male-Youth      Female-Adult    Female-Youth   
        # of agencies   % of agencies   # of agencies   % of agencies   # of agencies   % of agencies   # of agencies   % of agencies   # of agencies   % of agencies   # of agencies   % of agencies   # of agencies   % of agencies  
Site nearest the client's home                                                                                                                 
Site that best meets the clients needs                                                                                                                 
Other                                                                                                                  
Total Number of Agencies                                                                                                                       

Pat

-----Original Message-----
From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Gene Maguin
Sent: November 30, 2010 10:28 AM
To: [hidden email]
Subject: Re: [SPSSX-L] organizing output

Patricia,

My first thought is to restructure your dataset but there is something I

don't understand about your survey. Based on your table, I'm picturing a

fixed set of questions, for example, q1 to q12, that are answered for each

program. One program agencies answer the questions once, two program

agencies answer them twice, once for each program, etc. So a one program

agency has valid data for q1-q12 and missing for q13-q36 and q37-q84. That

structure could be restructured to this

Agency Type program q1-q12

 10     1     1      .....

 12     2     1      .....

 12     2     2      .....

 67     4     1      .....

Then, I think you could use ctables to produce what is really a

crosstab/means report structure.

Your question numbering, however, doesn't seem to correspond to this

structure. What am I misunderstanding?

Gene Maguin

________________________________

From: SPSSX(r) Discussion [[hidden email]] On Behalf Of

Cleland, Patricia (EDU)

Sent: Tuesday, November 30, 2010 9:38 AM

To: [hidden email]

Subject: organizing output

I'm analysing the results of a survey that was completed by about 100

agencies and am looking for help in how to display the output.

Here's the background: Most of the agencies describe themselves as offering

a single program. A small group describe themselves as offering two separate

programs, e.g., separate programs for males and females. Another small group

describes themselves as offering four programs, e.g., separate programs for

males and females by age group (that is, separate programs for adult males,

youth males, adult females, youth females).

All agencies answered the same basic questions, but for the two- and

four-program agencies each question was presented two or four times

respectively. The survey, which was administered via SurveyMonkey, had skip

logic which directed the agencies to the questions appropriate for their

program type.  One-program agencies answered questions 4-29; two-program

agencies answered questions 30-54 and four-program agencies answered

questions 55-79. Thus, each question was present in the survey 7 times (1

plus 2 plus 4).

I have analyzed the results separately for the program types, that is,

created a set of frequency tables for the one-program agencies, a set of

tables for the two-program agencies and tables for the four-program

agencies.

=====================

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

[John F Hall]

Pat

 

I'm still a little confused.  What do the data actually look like for (q7_2a)  (q7_2b)  (q7_2c)

 

Are there three variables or only one, but with three values?  Are they coded as numeric integers or as alpha strings?  Ditto for the other variables.

 

Can you send sample output for

 

freq q7_2 q33_2 q59_2 .

 

disp lab .

 

John Hall
[hidden email]
http://surveyresearch.weebly.com

 

----- Original Message -----

From: [hidden email]

To: [hidden email]

Sent: Tuesday, November 30, 2010 6:18 PM

Subject: Re: organizing output

Sorry, I wasn't clear in how the data is organized.  Let me give you an actual question.

"If the program is offered in more than one location, how are clients assigned to the location?"

Here's how the responses are presented to the one-program agencies. The codes in brackets following the options are the variable names. (Of course, they don't appear in the actual survey.)

a. Site nearest client's home (q7_2a)

b. Site that best meets the client's needs (q7_2b)

c. Other (q7_2c)

For two-program agencies, the question is Q33_2 and the response items are presented in a matrix. I've included the variable names in the matrix.

        Males   Females
Site nearest client's home      Q33_2aM Q33_2aF
Site that best meets the client's needs Q33_2bM Q33_2bF
Other   Q33_2cM Q33_2cF

For four-program agencies, the question is q59_2 and the responses are again presented in a matrix. Again the variable names are included in the matrix for your information.

        Males-Adult     Males-Youth     Females-Adult   Females-Youth  
Site nearest client's home      Q59_2aMA        Q59_2aMY        Q59_2aFA        Q59_2aFY       
Site that best meets the client's needs Q59_2bMA        Q59_2bMY        Q59_2bFA        Q59_2bFY       
Other   Q59_2cMA        Q59_2cMY        Q59_2cFA        Q59_2cFY       

The survey was structured this way to minimize the burden on the agency staff that were completing the survey.

I have frequency tables for each of the questions above: one for q7_2, two for q33_2 (separately for males and females) and four for q59_2, one for each of the programs.

Here's the table that my client wants to see:

Response        One-program     Two-Programs    Four Programs  
                Male    Female  Male-Adult      Male-Youth      Female-Adult    Female-Youth   
        # of agencies   % of agencies   # of agencies   % of agencies   # of agencies   % of agencies   # of agencies   % of agencies   # of agencies   % of agencies   # of agencies   % of agencies   # of agencies   % of agencies  
Site nearest the client's home                                                                                                                 
Site that best meets the clients needs                                                                                                                 
Other                                                                                                                  
Total Number of Agencies                                                                                                                       

Pat

-----Original Message-----
From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Gene Maguin
Sent: November 30, 2010 10:28 AM
To: [hidden email]
Subject: Re: [SPSSX-L] organizing output

Patricia,

My first thought is to restructure your dataset but there is something I

don't understand about your survey. Based on your table, I'm picturing a

fixed set of questions, for example, q1 to q12, that are answered for each

program. One program agencies answer the questions once, two program

agencies answer them twice, once for each program, etc. So a one program

agency has valid data for q1-q12 and missing for q13-q36 and q37-q84. That

structure could be restructured to this

Agency Type program q1-q12

 10     1     1      .....

 12     2     1      .....

 12     2     2      .....

 67     4     1      .....

Then, I think you could use ctables to produce what is really a

crosstab/means report structure.

Your question numbering, however, doesn't seem to correspond to this

structure. What am I misunderstanding?

Gene Maguin

________________________________

From: SPSSX(r) Discussion [[hidden email]] On Behalf Of

Cleland, Patricia (EDU)

Sent: Tuesday, November 30, 2010 9:38 AM

To: [hidden email]

Subject: organizing output

I'm analysing the results of a survey that was completed by about 100

agencies and am looking for help in how to display the output.

Here's the background: Most of the agencies describe themselves as offering

a single program. A small group describe themselves as offering two separate

programs, e.g., separate programs for males and females. Another small group

describes themselves as offering four programs, e.g., separate programs for

males and females by age group (that is, separate programs for adult males,

youth males, adult females, youth females).

All agencies answered the same basic questions, but for the two- and

four-program agencies each question was presented two or four times

respectively. The survey, which was administered via SurveyMonkey, had skip

logic which directed the agencies to the questions appropriate for their

program type.  One-program agencies answered questions 4-29; two-program

agencies answered questions 30-54 and four-program agencies answered

questions 55-79. Thus, each question was present in the survey 7 times (1

plus 2 plus 4).

I have analyzed the results separately for the program types, that is,

created a set of frequency tables for the one-program agencies, a set of

tables for the two-program agencies and tables for the four-program

agencies.

=====================

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