Measuring Public Opinion

Transcription

1 Measuring Public Opinin We all d n end f feeling and we mistake it fr thinking. And ut f it we get an aggregatin which we cnsider a bn. Its name is public pinin. It is held in reverence. It settles everything. Sme think it is the vice f Gd. Mark Twain

2 Public Opinin...aggregate f individual attitudes r beliefs shared by sme prtin f the adult ppulatin

3 Measuring Public Opinin We need t cmbine these individual pinins s that we can then determine what the public as a whle believes. Must cllect data in a scientifically rigrus fashin.

4 Steps in Measuring Public Opinin target ppulatin identificatin randm sample selectin questin cnstructin pll administratin data analysis

5 Measuring Public Opinin: Target Ppulatin Identificatin Wh s pinin are yu interested in measuring? target ppulatin: the ttal grup f individuals frm which the sample might be drawn generalizability: the extent t which we can apply the findings f ur research t the target ppulatin we are interested in

6 Measuring Public Opinin: Randm Sample Selectin A sample is a subset f peple, items r events frm a larger ppulatin that yu cllect and analyze t make inferences abut the larger ppulatin. T represent the ppulatin well, a sample shuld be randmly cllected and adequately large. Randmly cllected means that every persn in the target ppulatin must have an equal and knwn prbability f being included in the sample. size: The sample must be large enugh t adequately represent the ppulatin but nt larger than necessary r yu waste time, effrt and mney.

7 Scientific vs. Unscientific Plls: Randm Sample The key is whether the sample is randm. Remember that every persn in the target ppulatin must have an equal and knwn prbability f being included in the survey. Allws us t calculate the margin f errr and the cnfidence interval.

8 Scientific vs. Unscientific Plls: Randm Sample margin f errr: hw much the sample s pinins differ frm the ttal ppulatin s pinins (Unless yu survey the entire ppulatin, there will always be a margin f errr.) +/- 3.5% t abut +/- 6% 45% with a 4% margin f errr 41% 45% 49% cnfidence interval: A cnfidence interval is a range arund a measurement that cnveys hw precise the measurement is, hw sure we are f the results t % t 95% cnfidence

9 Scientific vs. Unscientific Plls: Randm Sample Need t be able t determine hw much yur sample differs frm the ttal ppulatin and hw sure yu are f the results. If the sample isn t randm, there s n way t determine that.

10 Measuring Public Opinin: Questin Cnstructin The prper wrding and phrasing f questins are vitally imprtant t prducing reliable, bjective data. laded r leading wrds: Subtle wrding differences can prduce great differences in results (The Supreme Curt culd /shuld/ might change the limits n free speech in light f terrrist activities). Strng wrds, such as prhibit may represent cntrl r actin and influence respnses. Wrd questins and answers in a neutral fashin, withut taking sides n an issue, and representing bth sides fairly. questin rder: Avid questins placed ut f rder r ut f cntext. Use a funnel apprach... brad and general questins at the beginning, mre specific questins, fllwed by mre general easy questins (like demgraphics).

11 Measuring Public Opinin: Questin Cnstructin use clsed-ended vs. pen-ended questins apprpriately clsed-ended: limits respndents with a list f answer chices frm which they must chse t answer the questin (yes r n, like r dislike r nt sure, walk r car r bus r ther, etc) D yu think gvernment cares abut yu? yes n I m nt sure. What des / culd gvernment d t shw it cares abut yu? pen-ended: n answer chices given, respndent answers any way desired (Why did yu chse that answer? What is the purpse f gvernment? etc) clear and understandable questins : Be specific in what yu want t knw. D yu like the president? (persnally? plitically? n the issues? respect? vte fr?) D yu cntact the gvernment regularly? (what s regularly? any kind f cntact? anyne cnnected t gvernment? elected? pst ffice?) length: When yu increase the length f questins, yu decrease the chance f receiving a cmpleted respnse.

12 Measuring Public Opinin: Questin Cnstructin cnfusing r unfamiliar wrding: Avid jargn, acrnyms, rarely used wrds, sentences with multiple phrases r mre than ne negative, language inapprpriate fr all educatinal levels. frced private respnses: If asking abut incme, age, race, etc include a prefer-nt-t-answer chice. nn-exhaustive clsed-ended respnses: If yu re nt sure yu ve included all pssible respnses fr a clsed-ended questin, include an ther-please-specify chice. duble-barreled questins: Make certain a questin isn t really tw questins that might have tw different answers (What is the mst respnsive and efficient frm f gvernment?). dichtmus questins: Answers shuld always be independent. D yu think pliticians are hnest r incmpetent? is nt dichtmus. They culd be bth.

13 Measuring Public Opinin: Pll Administratin Cntact thse selected in the randm sample. Tw f the mst cmmn ways that public pinin plls are cnducted include telephne and/r face-t-face interviews. Other methds include mail and self-administered surveys. telephne plls: use a sample f randmly-generated phne numbers that has been purged t eliminate business numbers, dead lines, etc... Because nearly 25% f the US ppulatin (as f 2014) has a cell phne but n landline, true scientific samples shuld include a subsample f cell phne users. Cell phne sampling cmes with its wn unique challenges. The interviewer will randmly select a persn in the husehld t be interviewed, cmmnly by asking fr the adult wh had the mst recent birthday. face-t-face plls: cnducted with the interviewer and the interviewee next t each ther... The interviewer reads material frm the survey and recrds the respnses. The interviewer may hand a card t the respndent fr him/her t select respnses. self-administered pll: respndent is handed the questinnaire t fill ut... Exit plls are examples f self-administered surveys.

14 Measuring Public Opinin: Data Analysis Interpret what the numbers mean. When researchers analyze plls they mean ne f tw things: statistical analysis f the plls r interpretive analysis f the plling results. statistical analysis: disaggregate (r break dwn) the data and run tables that shw hw different grups in the sample respnded t the questins, usually by using a statistical cmputer sftware package... invlves cunting and rganizing the respnses interpretive analysis: prvides an explanatin f what the data mean... Understanding a pll can be a challenging task, requiring critical thinking and cmmn sense t interpret what the respnses mean. Sme f the things yu might lk fr are indicatins f intensity, salience, cnsensus, divisiveness and/r stability.

15 Measuring and Analyzing Variables: Intensity Hw strngly held is an pinin? Plls may indicate wide supprt fr smething, but fail t indicate that the supprt is nt strngly held, fr example. Strngly Agree Agree Disagree Strngly Disagree

16 Measuring and Analyzing Variables: Salience Hw imprtant is the issue t peple? Mst peple dn t have strng pinins abut r even knwledge f issues. They respnd t plls based n what sme grup r the majrity thinks, what sunds right r etc rather than saying they dn t knw. Yes N Dn t Knw

17 Measuring and Analyzing Variables: Salience Americans are nt very knwledgeable abut the specifics f American gvernment because it s nt very imprtant t them. Yung peple knw even less than relatively ill-infrmed lder Americans.

18 Measuring and Analyzing Variables: Cnsensus Is there general agreement amng the citizenry n an issue? Agree Disagree

19 Measuring and Analyzing Variables: Divisiveness Hw deeply is the public divided n an issue? This is f cncern especially when pinins are als intense, salient and stable. Yes N

20 Measuring and Analyzing Variables: Stability Is public pinin stable (the same ver time) r unstable (likely t change ften)? Stability may als indicate intensity and salience. latency: A latent pinin is frmed n the spt, nly when needed (as distinct frm a deeply held pinin, which is stable ver time). Fr mst Americans, all pinins are latent. t1 t2 t3 t4 t5

21 Measuring and Analyzing Variables: Stability Ppular attitudes tward abrtin have been remarkably stable since Re v. Wade (1973).

22 Factrs that Shape Public Opinin family religin educatin incme media current events race gegraphy gender plitical party age career (There are at least as many factrs as there are individual characteristics.)

23 The Imprtance f Public Opinin Public pinin is imprtant IF it... cnnects us t demcracy shapes public plicy infrms plitical leaders cntrls plitical leaders t sme extent

24 THE END