2011-2012 Using DEWI This document covers the basic features of the Data Extraction Web Interface (DEWI) System. DEWI is an easy-to-use, platform independent one-stop-shop for data discovery and extraction. It is a webbased tool that allows users to find and extract variables from selected datasets within the Stanford University Libraries social science numeric data collection. You can locate variables by browsing through the entire list of variables or by searching on keywords within the data, such as variable names, labels, descriptions, or value labels. You can easily extract the selected variables from these lists without having to download the full dataset. Other useful information, such as variable sample frequencies and descriptions, as well as documentation, such as full-text manuals and technical documentation can be viewed within DEWI. The resulting file can be downloaded in the most popular statistical software formats to your computer or accessed directly from your SUNet account. DEWI is available for use by the Stanford community for research and instruction on the Social Science Data and Software web site. The following URL is where you can open DEWI: http://dewi.stanford.edu You will be asked to enter your SUNet ID (Stanford University Network Identifier). See the following URL for information on obtaining a SUNet ID: http://www.stanford.edu/services/sunetid Table of Contents Getting Started with DEWI... 2 Begin a New Session... 3 Viewing a Codebook...4 Browsing, Searching, and Viewing Variables... 4 DEWI Panels...4 Variable Frequenices and Descriptions...7 Refining Variable for Extraction...7 Naming and Formatting your Extract File... 8 For More Information and Assistance... 9 Help within DEWI...9 Feedback within DEWI... 9 Consulting...9 Stanford University Social Science Data and Software
Getting Started with DEWI All the datasets that are available via the DEWI System are arranged in alphabetical order by dataset title on the homepage within the Browse Tab. There are several ways to browse or search for data in the DEWI System: dataset title, Inter-University Consortium for Political and Social Research (ICPSR) or dataset number, and variable labels. Searching by Dataset Title One way to search is to click on the dataset title on the DEWI System s home page or the Browse tab to choose the dataset to start a new variable extraction session. You can also search for keywords within the dataset title such as election or Ghana. Searching by ICPSR and Dataset Number Another way to browse or search for data in the DEWI System is to do a search by study numbers. Different data archives such as ICPSR and the Roper Center for Public Opinion Research have various ways of searching datasets. ICPSR has specific numbers attached to datasets, while Roper does not attach any specific number. ICPSR study numbers can be found at http://www.icpsr.umich.edu/ 2 - Using DEWI
Searching by Variable Labels You can search the DEWI System by variable labels and view all datasets with a specific keyword. By moving the cursor over each listed dataset, you are able to view the variable name and label with your specified keyword. Note that data providers have labeled their data to varying degrees (e.g., one provider might name a variable 'sex', while another might use 'gender', abbreviations or misspellings may be used in variable labels, and value labels may not be assigned to all values of a variable). Once you have located a dataset through dataset title, study number, or variable keywords, click on the title of the dataset to begin a DEWI extraction session. Begin a New Session Once you select a dataset, you will have to begin and name a new session. The session selection page is where you can name and begin a new session by typing a session name and clicking the red Start New Session button. You can also retrieve a previously saved session by clicking on the Session Name link. The information under Codebooks, Etc. and DEWI Processing Notes provide dataset descriptions and processing notes, such as DEWI add date, variable metadata limitations, and dataset source. Using DEWI - 3
Viewing a Codebook The DEWI System offers links to documentation, web sites, and other online resources. New users of a dataset should browse through the dataset s codebook to learn important information about the data collection and variables. You can view links to full-text online versions of the dataset codebooks, manuals, and other documentation on the Codebooks, etc. tab by selecting a dataset. The 'Codebooks, etc.' page is displayed in a new browser window that can be resized or moved for easy referral while continuing to access your DEWI System variable selection and extraction session. Browsing, Searching and Viewing Variables DEWI Panels The main DEWI System page is divided into three main panels -- the 'navigation panel' (top middle), the (browse or search) 'results panel' (lower left), and the 'extraction panel' (lower right). Read below for basic instructions on using these panels. Clicking on the '?' next to a button will display the specific help window for that button. Active buttons will be highlighted. Click on the DEWI button to return to this page. 4 - Using DEWI
Navigation Panel Browse Variables button The DEWI System allows you to browse all variables in a dataset, 100 variables at a time. Click on the Browse Variables button to view the first page of variables in the 'results panel' (lower left). Click on the Next or Back buttons to navigate to the next or previous page of variables. You can also jump to a particular page by entering the page number into the Go to Page box and clicking the red Go button. Variable frequencies or descriptions (if available) can be viewed by clicking on the variable name link in the list of variables. Browsing is the default search mode when you first enter a DEWI session. Search Variables button The DEWI System allows you to find variables in a dataset by searching for keywords in the variable names and labels and the variable value labels. If you enter more than one keyword, you can specify whether the search will find matches with All Words or Any Words that you enter by using the drop down box to the left of the keyword field. Note that data providers have labeled their data differently (e.g., one provider might name a variable 'sex', while another might Using DEWI - 5
use 'gender', abbreviations or misspellings may be used in variable labels. Submit your search by clicking on the red Search button. The results will be displayed in the 'results panel' (lower left). Results Panel The results panel lists up to 100 variables at a time. You can remove variables from the 'results panel' by checking the variables in the list, clicking on Check All or Uncheck All and clicking the red Save for Extraction button. The selected variables will be moved to the extractions panel for eventual extraction. Extractions Panel The variables that you wish to select for eventual extraction may be checked or unchecked in the 'results panel' (lower left) and saved to the 'extraction panel' (lower right) by clicking on the red Save for Extraction button. After clicking this button, the resulting changes will be reflected in the 'extraction panel'. The 'extraction panel' will always display the variables you have selected either in variable list form (default). In addition, the View Custom Codebook is a custom codebook format (variable sample 6 - Using DEWI
frequencies and descriptions, if available) that displays selected variables. Finally, you can quickly select or unselect all the displayed variables by selecting the Check All or Uncheck All button, respectively. You can remove variables from the 'extraction panel' by unchecking the variable in the list and clicking the Save Changes button. Note: If you do not click these buttons after making changes to the variables list, your changes will not be saved. Variable Frequencies and Descriptions Variable frequencies and descriptions (if available) can be viewed by clicking on the variable name link in the list of variables generated by your keyword search. Using DEWI - 7
Refining Variables for Extraction Another way to filter data is by refining saved variables before extraction. You can select values of saved variables you wish to extract from the dataset. You can also extract a subset of cases for a selected variable by value filtering. Click on the variable name link in the list of saved variables in the extractions panel and check the desired values. Naming and Formatting your Extract File When you have finished adding all your variables for extraction, your list can be submitted for extraction by clicking on the Extract Selected Variables button or the red Extract Variables button in the 'extraction panel'. You will then be taken to the 'Extraction' page to name and begin your extraction job. 8 - Using DEWI
You will need to choose a memorable file name and format for your extract file. Because of file naming restrictions of various statistical packages, the file name should be 21 characters or less. It is best to being with your file name with a letter. The file name should be made up of letters, numbers, and underscores only. In addition, some applications have other limits that your extract file may violate (e.g., Excel supports a maximum of 65,536 rows and 256 columns). Submit your extraction request by clicking on the red Begin Extraction button. Once your extraction request has been submitted, it will be placed in the Extraction Queue and processed according to its position in the queue (to view the status of your extraction job, click the Extraction Job Queue button). The 'Extraction Job Queue' page also contains the 'Completed Extractions' Table. When your extraction job has been processed, a message will be sent to your email address. Download your extraction job directly to your computer from the web link provided in the 'Completed Extractions' table on the 'Extraction Job Queue' page. You will need the appropriate statistical package on the computer that you download your extract file in order to do your analysis. You can also access your extract files and previous extractions directly via your Leland Systems account using the pathname that is emailed to you after your extraction job is completed. For More Information and Assistance This section of the guide directs you to useful resources for using DEWI and resolving any questions or problems that could arise as you become familiar with the DEWI System. Using DEWI - 9
Help within DEWI The DEWI System has a Help tab in the DEWI home view and Help button in the DEWI panel view. To access the Help system, click on the Help tab or button. Feedback within DEWI Tell us how we can improve the DEWI Data System and make it more useful to you. Click on the Feedback tab in the DEWI home view to submit your comments, report data and processing errors, and suggest other data resources for DEWI. You can also request a DEWI demonstration for your class. SSDS Software Services at Stanford If you have questions about using the DEWI System, please contact the data consultants at Social Science Data and Software. Data services provide technical support for DEWI users at Stanford. Users can ask questions or make appointments with the consultants via our website. For more information or to contact us, see the web at: http://ssds.stanford.edu/ Copyright 2011, by The Board of Trustees of the Leland Stanford Junior University. Permission granted to copy for noncommercial purposes, provided we receive acknowledgment and a copy of the document in which our material appears. No right is granted to quote from or use any material in this document for purposes of promoting any product or service. Social Science Data and Software Document revised: 91/7/2011 10 - Using DEWI