Complex sampling plan - problems

[Sheryl Keller]

My sample is a 30 cluster methodology with stratification. It is self weighting. Hiwever the Complex Analysis plan wizard reuires a weight, so I created a weight variabkle with value of 1 in all records, was that correct way to handle this?

And -- main question -- when I run descriptives using this plan I am unavble to get design effects even though I specifiy them in the command. Comes out blank.

What am I doing worng?

This is how the plan reads:

* Analysis Preparation Wizard.
CSPLAN ANALYSIS
 /PLAN FILE='C:\cluster.csaplan'
 /PLANVARS ANALYSISWEIGHT=weight
 /SRSESTIMATOR TYPE=WOR
 /PRINT PLAN
 /DESIGN  STRATA= commune  CLUSTER= villagecl
   /ESTIMATOR TYPE=WR


Here's the my syntax:

* Complex Samples Descriptives.
CSDESCRIPTIVES
 /PLAN FILE = 'C:\cluster.csaplan'
 /SUMMARY VARIABLES =FICyesno
 /MEAN
 /STATISTICS SE DEFF CIN (95)
 /MISSING SCOPE = ANALYSIS CLASSMISSING = EXCLUDE.

And here is what I get back. No deisgn effect shown:

Univariate Statistics
       Estimate Standard Error 95% Confidence Interval Design Effect
                                                                 Lower Upper Lower
Mean FICyesno .5177           .02511        .4656         .5698 .

Same result no matter what variabkle I choose.

Advice????? Thanks!

[Jarrod Teo-2]

Hi,
 
There are these issues that are bothering me about normality. As we know, means comparison of means require normailty.
 
Suppose if I am going to use an ANOVA testing, and my scale variable is not normal even after I do a LN or sqrt, what should I do?
 
It will be nice also if there is a way for me to know which normalisation method to use as in sqrt, LN, or something else based on some Statistics.
 
Thank you for any advice.
 
Dorraj Oet
 

 

[Bruce Weaver]

DorraJ Oet wrote

Hi,

There are these issues that are bothering me about normality. As we know, means comparison of means require normailty.

Suppose if I am going to use an ANOVA testing, and my scale variable is not normal even after I do a LN or sqrt, what should I do?

It will be nice also if there is a way for me to know which normalisation method to use as in sqrt, LN, or something else based on some Statistics.

Thank you for any advice.

Dorraj Oet

What do the distributions look like?  Note that you need to plot the two groups separately.  And what are your sample sizes?


Here are some comments that might give some insight into why I asked those questions.

1. Remember that the normality assumption applies to the errors, which means it is normality within groups, not normality for all of the groups combined.

2. Independence of observations is the far more important assumption.  

3. ANOVA and the t-test are quite robust to violations of normality, especially as the sample size increases.  

All t-tests have the same basic format:

   t = (statistic - parameter|H0) / SE of the statistic

If the statistic in the numerator has a sampling distribution that is approximately normal, the test will be pretty good.  

For the independent groups t-test, the statistic is the difference between the two sample means.  The central limit theorem tells us that:

1. If the populations are normal, the sampling distribution will be normal, regardless of sample size.

2. If the populations are not normal, the sampling distribution will converge on the normal distribution as sample size increases.  (E.g., see the sampling distribution demo here:  http://onlinestatbook.com/stat_sim/index.html.)

So if your population distributions are similar in shape and your sample sizes are large enough, the t-test is probably better than you think.

HTH.

[Jarrod Teo-2]

Hi,
 
Currently I am having an issue with 2 of my rating variables.
 
I am having 2 rating variables of value 1-6 however, these 2 questions are supposed to be 1 question due to wrong questionnaire design by my colleague. Is there a way to combine these 2 rating variables because ultimately, I will like to find the mean. The following is the mockup of my data file.
 
var1 var2
1      1
2      4
3      5
1      6
5      6
5      1
 
These 2 variables have to be combined into 1 so that I can find the mean of this final variable.
 
My personal idea was to sum these 2 variables up, divided by 2, and calculate the mean based on this newly created variable. Is this a wrong approach?
 
Thanks for any advice on this matter.
 
Regards
Dorraj Oet

[John F Hall]

Probably, but will this do?

 

compute (newvar + var1 + var2) /2) .

freq var1 var2 newvar / sta mea .

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

From: [hidden email]

To: [hidden email]

Sent: Wednesday, September 22, 2010 11:50 AM

Subject: Proper way to combining 2 rating variables into 1

Hi,
 
Currently I am having an issue with 2 of my rating variables.
 
I am having 2 rating variables of value 1-6 however, these 2 questions are supposed to be 1 question due to wrong questionnaire design by my colleague. Is there a way to combine these 2 rating variables because ultimately, I will like to find the mean. The following is the mockup of my data file.
 
var1 var2
1      1
2      4
3      5
1      6
5      6
5      1
 
These 2 variables have to be combined into 1 so that I can find the mean of this final variable.
 
My personal idea was to sum these 2 variables up, divided by 2, and calculate the mean based on this newly created variable. Is this a wrong approach?
 
Thanks for any advice on this matter.
 
Regards
Dorraj Oet

[ViAnn Beadle]

This syntax will not run because the compute is wrong.

 

But I’m not sure what you mean by 2 questions are supposed to be 1. Can you be more specific as to what the two variables, var1 and var2, represent. Various ways of combining could be adding, taking the max, or taking the min. Also, how do you want missing values for either variable to be handled in the computation.

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of John F Hall
Sent: Wednesday, September 22, 2010 7:04 AM
To: [hidden email]
Subject: Re: Proper way to combining 2 rating variables into 1

 

Probably, but will this do?

 

compute (newvar + var1 + var2) /2) .

freq var1 var2 newvar / sta mea .

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

From: [hidden email]

To: [hidden email]

Sent: Wednesday, September 22, 2010 11:50 AM

Subject: Proper way to combining 2 rating variables into 1

 

Hi,
 
Currently I am having an issue with 2 of my rating variables.
 
I am having 2 rating variables of value 1-6 however, these 2 questions are supposed to be 1 question due to wrong questionnaire design by my colleague. Is there a way to combine these 2 rating variables because ultimately, I will like to find the mean. The following is the mockup of my data file.
 
var1 var2
1      1
2      4
3      5
1      6
5      6
5      1
 
These 2 variables have to be combined into 1 so that I can find the mean of this final variable.
 
My personal idea was to sum these 2 variables up, divided by 2, and calculate the mean based on this newly created variable. Is this a wrong approach?
 
Thanks for any advice on this matter.
 
Regards
Dorraj Oet

[John F Hall]

My mistake.

 

compute newvar = (( var1 + var2) /2) .

freq var1 var2 newvar / sta mea .

 

. . . but why wasn't the data sorted out before anyone started to analyse it?

 

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

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

From: [hidden email]

To: [hidden email]

Sent: Wednesday, September 22, 2010 4:37 PM

Subject: Re: Proper way to combining 2 rating variables into 1

This syntax will not run because the compute is wrong.

 

But I’m not sure what you mean by 2 questions are supposed to be 1. Can you be more specific as to what the two variables, var1 and var2, represent. Various ways of combining could be adding, taking the max, or taking the min. Also, how do you want missing values for either variable to be handled in the computation.

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of John F Hall
Sent: Wednesday, September 22, 2010 7:04 AM
To: [hidden email]
Subject: Re: Proper way to combining 2 rating variables into 1

 

Probably, but will this do?

 

compute (newvar + var1 + var2) /2) .

freq var1 var2 newvar / sta mea .

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

From: [hidden email]

To: [hidden email]

Sent: Wednesday, September 22, 2010 11:50 AM

Subject: Proper way to combining 2 rating variables into 1

 

Hi,
 
Currently I am having an issue with 2 of my rating variables.
 
I am having 2 rating variables of value 1-6 however, these 2 questions are supposed to be 1 question due to wrong questionnaire design by my colleague. Is there a way to combine these 2 rating variables because ultimately, I will like to find the mean. The following is the mockup of my data file.
 
var1 var2
1      1
2      4
3      5
1      6
5      6
5      1
 
These 2 variables have to be combined into 1 so that I can find the mean of this final variable.
 
My personal idea was to sum these 2 variables up, divided by 2, and calculate the mean based on this newly created variable. Is this a wrong approach?
 
Thanks for any advice on this matter.
 
Regards
Dorraj Oet

[Gosse, Michelle]

Hi there,

 

When looking to combine two questions (and take the mean, for example), I use a Cronbach’s alpha test to see whether the responses on the two items are correlated. If the result is good enough, I use this as the final rationale for combining the two (after looking at the theoretical reason to combine, as in both questions measure different parts of the same factor we are interested in).

 

Here is an example of the syntax I have used. In this example, which is based on a survey, we want to combine the responses from questions 1A and 1B of the survey (these were 2 Likert items, scale 1 to 7):

RELIABILITY

  /VARIABLES=Q1A Q1B

  /SCALE('ALL VARIABLES') ALL

  /MODEL=ALPHA

  /STATISTICS=DESCRIPTIVE CORR COV

  /SUMMARY=MEANS VARIANCE COV CORR

  /ICC=MODEL(MIXED) TYPE(CONSISTENCY) CIN=95 TESTVAL=0.

 

HTH

 

Cheers

Michelle

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of John F Hall
Sent: Thursday, 23 September 2010 1:04 a.m.
To: [hidden email]
Subject: Re: Proper way to combining 2 rating variables into 1

 

Probably, but will this do?

 

compute (newvar + var1 + var2) /2) .

freq var1 var2 newvar / sta mea .

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

From: [hidden email]

To: [hidden email]

Sent: Wednesday, September 22, 2010 11:50 AM

Subject: Proper way to combining 2 rating variables into 1

 

Hi,
 
Currently I am having an issue with 2 of my rating variables.
 
I am having 2 rating variables of value 1-6 however, these 2 questions are supposed to be 1 question due to wrong questionnaire design by my colleague. Is there a way to combine these 2 rating variables because ultimately, I will like to find the mean. The following is the mockup of my data file.
 
var1 var2
1      1
2      4
3      5
1      6
5      6
5      1
 
These 2 variables have to be combined into 1 so that I can find the mean of this final variable.
 
My personal idea was to sum these 2 variables up, divided by 2, and calculate the mean based on this newly created variable. Is this a wrong approach?
 
Thanks for any advice on this matter.
 
Regards
Dorraj Oet

**********************************************************************

This email and any files transmitted with it are confidential and

intended solely for the use of the individual or entity to whom they

are addressed. If you have received this email in error please notify

the system manager.

This footnote also confirms that this email message has been swept by

MIMEsweeper for the presence of computer viruses.

www.clearswift.com

**********************************************************************

[Sheryl Keller]

Any chance of getting a reply to my original post re problems using SPSS CSAMPLE?

"

Sheryl Keller wrote

My sample is a 30 cluster methodology with stratification. It is self weighting. Hiwever the Complex Analysis plan wizard reuires a weight, so I created a weight variabkle with value of 1 in all records, was that correct way to handle this?

And -- main question -- when I run descriptives using this plan I am unavble to get design effects even though I specifiy them in the command. Comes out blank.

What am I doing worng?

This is how the plan reads:

* Analysis Preparation Wizard.
CSPLAN ANALYSIS
 /PLAN FILE='C:\cluster.csaplan'
 /PLANVARS ANALYSISWEIGHT=weight
 /SRSESTIMATOR TYPE=WOR
 /PRINT PLAN
 /DESIGN  STRATA= commune  CLUSTER= villagecl
   /ESTIMATOR TYPE=WR


Here's the my syntax:

* Complex Samples Descriptives.
CSDESCRIPTIVES
 /PLAN FILE = 'C:\cluster.csaplan'
 /SUMMARY VARIABLES =FICyesno
 /MEAN
 /STATISTICS SE DEFF CIN (95)
 /MISSING SCOPE = ANALYSIS CLASSMISSING = EXCLUDE.

And here is what I get back. No deisgn effect shown:

Univariate Statistics
       Estimate Standard Error 95% Confidence Interval Design Effect
                                                                 Lower Upper Lower
Mean FICyesno .5177           .02511        .4656         .5698 .

Same result no matter what variabkle I choose.

Advice????? Thanks!