Unformatted text preview:

Measurement17.871Mapping from the Abstract to the Measurement Some abstract things we try to measure Alienation Traditional morality Democracy Party identification Fear of defeat Terrorism IdeologyMapping from the abstract to the measurementXYxyexeyTheoryObservationSources of Error in Measurement Random and Systematic Conceptual or design error Survey question wording Transcription, calculation & mechanization errorsWhat to Do About Error Practice safe data Know where your data come from Use multiple measurement techniques!!!!! Collect as much data as possible and disaggregate Show your data and watch for anomalies Run the summarize command before all analysesPractical things to do about measurement Distinction between a measure and an indicator Measure: straightforward quantification of a variable of interest Indicator: quantification of a variable that is believed (or known) to be highly correlated with the “real” variable of interestExamples of Measures Income in $$ Age in years Votes Number of wars Campaign contributions Measurement issues tend to focus on the quality of the data-gathering method, especially sampling – e.g., Iraqi deathsExamples of Indicators Most measures are really indicators Public opinion Party identification Trust in government Ideology Economics Gross domestic (national) product Characterizations of political systems Freedom Transparency Democracy Quality of collegesMultiple-measure indicators “Moral traditionalism” batteryOther Multiple-Variable Indicators Americans for Democratic Action Support Scores Freedom House freedom assessment MIT Teaching Quality Transparency International’s “Corruption Perceptions Index” DNominate ScoresInstructions for data entry The first entry in a column should be a unique variable name that describes the variable, e.g., “Bush_vote” or “male” Variable names and data should have  no spaces and No special characters (e.g., \,$,*,@, #) Data should always be entered numerically (e.g., Republican vote coded to 1, Democratic vote coded to 0), unless data are unique strings, e.g., proper namesInstructions for data entry Each column should code for a single variable If concepts have multiple components, try to break them into separate variables E.g., if you are interested in Democratic incumbents versus Republican incumbents, code the party as one variable and incumbency as another In Excel, format the columns as numbers, not dollars or percentages  Include the source information for variables in the same file as a comment on the cell, on a separate sheet or page, or to the left of the data columns Whenever possible, both print or photocopy the source of the data and save the source in the same folder as the


View Full Document

MIT 17 871 - Measurement

Documents in this Course
Load more
Download Measurement
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Measurement and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Measurement 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?