Variable Search with paneldata.org¶
With paneldata.org it is also possible to search for variables. For example, if you want to find more information about generated variables, a search with paneldata.org is indispensable. For example, the platform offers comprehensive frequency counts, the chronology of the variables searched for, a cross-study variable linkage via concepts, a syntax generator and a topic list for content search in the SOEP.
bbh5508: Wave “bb” (Survey Year 2011); household questionnaire (“h”), question number 55, item 8
Please select the study SOEP-Core. The SOEP-Core overview contains important general information about the study, e.g. data access, survey method, questionnaires, thematic diversity, terms for missing codes, all available data sets of the study and metadata-based questionnaires. To search for a variable, a data set or a publication, simply enter the desired search term in the search field.
In order for the search to be successful, specific information from the user are necessary. The results window displays all results of the search. It can be seen that the variable “bbh5508” originates from the data provided by SOEP-Core and can be found in the data set “bbh” (survey year 2011). If your search is not so specific, you can also search by keywords. We are still interested in the topic “car”.
To better limit the 1091 results, the filter options on the left should be used. We are looking for variables from the ordered SOEP-Core datasets. In the windows “type” and “study” we select “variable” and “soep-core”.
Now all variables are displayed, which contain the term “Car” in the SOEP-Core data. The variable search can be further limited by specifying the data set or the survey year. For more information about the different data sets in SOEP-Core visit the chapter Data Distribution File. To select original data that can be assigned to a question in the questionnaire, select the subtype “org/net”. The specific selection of the analyzing unit allows you to choose whether the variable should provide information on the household level(“h”) or on the individual level (“p”). If you are interested in household-specific variables, select “h” as the “Analysis unit”. If you are explicitly interested in the survey year 2011, the variable search can be limited to five variables.
There are only five results left, which also shows our searched variable. If you click on the variable “bbh5508” you will get additional information about the variable.
First you see the weighted absolute frequencies for the variable. It is possible to remove the missing codes from the analysis and/or to display the relative frequencies. Even without opening the data set, Paneldata.org gives you a good overview of the frequencies of a variable.
In the Related Variables section you will also find the chronology of the variable you are looking for. The sample variable was collected in 2001, 2003, 2005, 2007, 2011, 2013. Below the survey year, the name of the variable in the respective year is displayed and can be clicked to access the respective variable page. At one glance it is possible to see when a variable was measured, how often it was measured and what its name is in the respective survey year.
In addition, by clicking on “Output variables”, you will see a variable forwarding you to the variable in “long” format. For a more detailed understanding of the long format, read the chapter Data Structure in long Format (long).
As soon as you click on the “long” variable, you will get to the variable overview for this variable in long-format. The overview of variables does not differ. It can be seen that our example variable “bbh5508” can also be found in long-format in the data set “hl” with the variable label “hlf0181”.
The field “Label translations” shows the value labels of the variables in German and English. In addition, all missing codes used in SOEP are listed and explained.
The Label table window shows you the absolute frequencies of the variable at different collection times. This makes it possible to identify initial trends in how response behaviour has changed over a period of time. The assigned value code is output for each possible characteristic value and the absolute frequencies are displayed in parentheses.
In our example output we see that for the variable “th5106” 800 respondents in the wave “t” (2003) state “financial reasons” as the reason for the absence of a car in the household. For our example variable “bbh5508” in the survey year 2011 (wave “bb”) there are already 871 respondents.
Paneldata.org is an excellent way to get an first overview of certain variables.
The info box on the right-hand side provides an overview of all relevant information about the variable and the data set. Beside the basic information you will find the information what kind of variable you are looking for under “Conceptual Dataset”. In our example “bbh5508” you can see that variables with a “Conceptual Dataset: questionnaire” describe original variables that are assigned to a questionnaire. Generated variables are “Conceptual Dataset: gen”. To get an overview of the different data set types of SOEP-Core, visit the chapter Data Distribution File.
In addition to searching for keywords or using the various filter settings, you can also find what you are looking for directly in the data set search. Open paneldata.org, click on the study SOEP-Core and select the menu field “data”.
Now you get to an overview which shows you all data sets contained in SOEP Core.
Enter the data set you are looking for (“bbh”) in the search field at the top right and click on the data set. You are forwarded to an overview which shows you all variables from the “bbh” data set.
Now enter the variable you are looking for in the search field at the top right and click on the desired variable. You are then forwarded to the variable overview and receive detailed information about the variable. Paneldata.org offers the user very different search options to suit the individual search behavior of each user.