5.1 Estimation of frequencies

The variable \(\texttt{ENV_AFRDWLK_MCQ}\) represents the CLSA survey question: “People would be afraid to walk alone after dark in this area.” If we want to estimate the total number of people in the study population who agree or disagree about the statement, we can use the following sets of codes to create tables of frequencies with standard errors.

R

svytotal(  ~(ENV_AFRDWLK_MCQ1), design= CLSA.design)

SAS

PROC SURVEYFREQ  data= CLSAData ;        
  TABLE  ENV_AFRDWLK_MCQ1;                                           
  strata GEOSTRAT_TRM ; 
  weight WGHTS_INFLATION_TRM;                                      
RUN; 

SPSS

Click “Analyze” \(\rightarrow\) “Complex Samples” \(\rightarrow\) “Frequency…” \(\rightarrow\) Browse and select the \(\texttt{CLSADesign.csaplan}\) created. \(\rightarrow\) Click “\(\texttt{Next}\)” and import a variable (\(\texttt{ENV_AFRDWLK_MCQ}\)) \(\rightarrow\) Click “\(\texttt{Statistics...}\)” \(\rightarrow\) select “Population size” and “Standard error” \(\rightarrow\) Click “Continue” \(\rightarrow\) Click “OK”.

Stata

 svy linearized : tabulate ENV_AFRDWLK_MCQ, count se ci  stubwidth(20) format(%10.0g)

Result comparison

	R		SAS		SPSS		Stata
Population Est.	Total	SE	Total	SE	Total	SE	Total	SE
Strongly Agree	52666	7452	52666	7452	52666	7452	52666	7452
Agree	65819	8502	65819	8502	65819	8502	65819	8502
Disagree	149611	13811	149611	13811	149611	13811	149611	13811
Strongly Disagree	281096	21512	281096	21512	281096	21512	281096	21512

Note: The total of the population estimates from the table above is much smaller that the CLSA study population. It is because the dataset used for illustration is only a subset of the CLSA dataset. The actual dataset should give much larger population totals.