POL 322 week5 Lec

POL 322 week5 Lec Oct 15, 2012 Measurement: Part 1 operationalization- moving from concept to scientific measurement -abstract to concrete -determine how to measure all variables -difficult to define -needs empirical expression -don’t usually “see” concepts -ex. education -number of years? elementary? secondary? -these are only formal education, what about practical learning? -political culture, education, gender, income, democracy -even more difficult to define -social sciences like political concepts are the most difficult to define -today, even gender is an issue -is there a difference between gender and sex? -democracy (Paxton reading) -how you define it can completely alter a study -2 avenues of operationalization 1. direct: closest to observing directly -traditionally; gender 2. indirect: involves outlining dimensions of concept -try and measure substantially -not always agreed upon -ex. support for mission in Afghanistan -encompassed in number of casualties, commitment, military -(assignment 1 reading) -must take care so as not to weigh one aspect of an issue more than another -ex. military support more than commitment or casualties -assessing measurement POL 322 week5 Lec Oct 15, 2012 -validity and reality -means of assessing the appropriate values -reliability -consistent results= consistent measures -expect using similar methods to produce similar results ex. weight scale -types of reliability techniques 1. inter-rater: consistency among raters 2. test-retest: repeat test of measure 3. split-half: two versions of administered to groups 4. item-total: link b/w item with rest of index 5. alternate-forms: two forms of equivalent measure -ex. Fletcher/ hove? -can you measure the same concept twice, using alternate methods of measurement -ex. one with a picture of a casualty -some with different coffin pictures -Validity -measuring what is supposed to be measured -are you measuring what you think you’re measuring -if not, you’re at the risk of systematic bias -occurs when inaccurate link b/w concept and variable is chronic and consistent problem -ex. sensitive topics-prejudice -can’t ask a person how prejudice they are -CES experiment? (3 women looking for funding, portuguese catholic, muslims etc.) -voting, everyone says they vote -how do you get around this? -”do you think your neighbor voted?” POL 322 week5 Lec Oct 15, 2012 -no test for validity, but 3 ways to help know 1. face- sniff test; does it seem to make sense? -need to explain why you think so -critique question wording, samples etc. 2. construct- compare with known and non-correlate measures -similar to concurrent -check a dependent variable’s relationship with other independent variables -draw from literature 3. predictive- does it predict outcomes? -how good are predictions? do they predict results to a reasonable degree? -ultimately, we want a measure to be both reliable and valid -reliable but not valid keeps giving you the wrong results consistently -impossible to had a valid, unreliable measure except with external results such as time Paxton article -if you measure democracy by male suffrage alone, you get a totally different image of democracy than that which occurred past women’s suffrage -in the past there were reliable theories on this b