• ino ino Ads by Toktok9ja
  • Balvinciglobal Balvinciglobal
  • Murya-Magazine-Banner Murya-Magazine-Banner



Who is Scared Of Pub­lic Opin­ion Polls?

.…who is scared of pub­lic opin­ion polls ? _ Dr Bell Ihua. (@bellemsky via twit­ter)

It was Falz the bahd guy who recent­ly called out crit­ics for “shoot­ing down the mes­sen­ger and miss­ing the mes­sage”, as he respond­ed to black­lash on his “This is Nige­ria” video. Per­haps, this is new ter­ri­to­ry for Falz, but not so for social researchers and pub­lic opin­ion poll­sters. We are often crit­i­cized for the results and find­ings of our polls; more so, by peo­ple who have nei­ther read the research report nor under­stand the method­ol­o­gy adopt­ed in con­duct­ing the poll, sur­vey or research study. The rise or fall of any piece of research or pub­lic opin­ion poll lies in its method­ol­o­gy. In oth­er words, what deter­mines whether any­one should give a hoot about the find­ings of a poll or what con­fers cred­i­bil­i­ty on any piece of research are in its abil­i­ty to answer some ques­tions – how was the poll con­duct­ed? how was pri­ma­ry data col­lect­ed? how were the ques­tions word­ed? how rep­re­sen­ta­tive is the data? what’s the sam­ple size of the poll? and what’s the mar­gin of error, amongst oth­er ques­tions.

Pub­lic opin­ion polls are an essen­tial part of pol­i­tics, democ­ra­cy, mar­kets and social life in devel­oped soci­eties. From polit­i­cal pre­dic­tions and approval rat­ings, to titles of Hol­ly­wood movies and song-of-the-week picks, polls are tools used to mea­sure and gauge pub­lic opin­ion on issues affect­ing the soci­ety. Let’s take a step back­ward to define “pub­lic opin­ion”; it sim­ply refers to the mind, thought and expres­sion of the gen­er­al pop­u­la­tion on a par­tic­u­lar issue or sub­ject mat­ter.

Is pub­lic opin­ion polling new?

Cer­tain­ly not, polling isn’t new. It was Jacques Neck­er, French states­man and Finance Min­is­ter under King Louis XVI, who in the 18th cen­tu­ry, first empha­sized the impor­tance of L’opinion publique, as he advo­cat­ed for the pub­lish­ing of gov­ern­ment accounts and bud­gets in order to boost pub­lic con­fi­dence in the years pre­ced­ing the French Rev­o­lu­tion. How­ev­er, pub­lic opin­ion polling was pop­u­lar­ized in Amer­i­ca. It was George Gallup, founder of the renowned Gallup Poll and father of pub­lic opin­ion polling, who in the 1930s con­clud­ed that there was no dif­fer­ence between polling on tooth­paste and pol­i­tics. In recent times too, poll­sters like Stan­ley Green­berg, have been hailed as the father of mod­ern polling tech­niques and described as a man who doesn’t just have a fin­ger on the people’s pulse, but has an IV inject­ed into it. 

How use­ful are polls?

For gov­er­nance to be effec­tive, it must be inclu­sive and par­tic­i­pa­to­ry. Opin­ion polling is sig­nif­i­cant tool that can enhance inclu­sive­ness and par­tic­i­pa­tion in gov­er­nance. In the Unit­ed States for instance, polling has become an entrenched part of democ­ra­cy, as the need to under­stand what the pub­lic think is con­sid­ered to be at the heart of gov­er­nance. This dis­po­si­tion was apt­ly cap­tured by Abra­ham Lin­coln, who was quot­ed to have said “what I want to get done is what the peo­ple desire to have done, and the ques­tion for me is how to find that out exact­ly.” Pres­i­dent Franklin Roo­sevelt relied on poll­sters like Emil Hur­ja and Prince­ton pro­fes­sor Hadley Cantrill to shape strate­gies and pub­lic pol­i­cy; same with Pres­i­dent JF Kennedy. Like­wise, Pres­i­dent Ronald Reagan’s White House relied on polling sup­port from Dr. Richard Wirth­lin; while Pres­i­dents Bill Clin­ton and George W. Bush relied heav­i­ly on the polling gurus like Jere­my Ros­ner, Justin Wallin and Al Quin­lan. Apart from the Unit­ed States, there’s evi­dence to sug­gest that sev­er­al glob­al lead­ers and mem­bers of par­lia­ment (MPs) have in the recent decades relied on opin­ion research to help shape pub­lic pol­i­cy and reforms in their coun­tries – from Prime Min­is­ters Tony Blair and Julia Gillard, to Pres­i­dents Nel­son Man­dela, Mikheil Saakashvili, and Vik­tor Yushchenko to men­tion a few.

Con­duct­ing sci­en­tif­ic polls

Beyond media vox-pop and adhoc straw polls, which involve peo­ple respond­ing to a set of ques­tions with­out ref­er­ence to how respon­dents are select­ed; there are cer­tain indi­ca­tors that can make a poll, sur­vey or research study sci­en­tif­ic. When we say a poll is “sci­en­tif­ic”, it sim­ply means the poll has been con­duct­ed through a set of rig­or­ous process­es and qual­i­ty con­trol mech­a­nism to pro­duce cer­tain results; and that those process­es can be repli­cat­ed to pro­duce about the same result, with some lev­el of con­fi­dence and per­cent­age of replic­a­bil­i­ty. Sev­er­al ter­mi­nolo­gies can be asso­ci­at­ed with polling, such as: Research pop­u­la­tion, Sam­pling tech­nique, Sam­ple size, Field­ing and data col­lec­tion, CAPI and PAPI, Ques­tion­naire design, Ques­tion word­ing, mar­gin of error, ran­dom­iza­tion and strat­i­fi­ca­tion amongst oth­ers. These would nor­mal­ly make up a full mod­ule in a sur­vey research course, or an arti­cle for anoth­er day. My inten­tion isn’t to make read­ers polling experts from this one arti­cle; but to help demys­ti­fy the myth often asso­ci­at­ed with pub­lic opin­ion polls in Nige­ria, and appre­ci­ate what makes good ver­sus bad polls. Con­se­quent­ly, I would like to touch on a few of those ter­mi­nolo­gies that I con­sid­er fun­da­men­tal to pub­lic opin­ion polling.

Research pop­u­la­tion

            All polls are based on sam­ples drawn from larg­er pop­u­la­tions. The pur­pose of every opin­ion poll is to attempt to use a sam­ple to make infer­ences or form con­clu­sions on the larg­er pop­u­la­tion. With a pop­u­la­tion of between 180 and 200 mil­lion Nige­ri­ans, is it pos­si­ble to access the entire pop­u­la­tion for a poll or sur­vey? As far as social research is con­cerned, it is sim­ply imprac­ti­ca­ble to have access to the entire 180 mil­lion Nige­ri­ans, unless you’re on a mis­sion to con­duct a cen­sus. Cen­sus stud­ies don’t come cheap. Nigeria’s last cen­sus was con­duct­ed in 2006, and ought to be repeat­ed every 10 years; how­ev­er, due to finan­cial con­straints the coun­try missed its 2016 cen­sus tar­get and is still shop­ping for funds to con­duct a fresh cen­sus. Notwith­stand­ing, if the total pop­u­la­tion isn’t acces­si­ble, then we can rely on a research pop­u­la­tion. A research pop­u­la­tion is a well-defined col­lec­tion of indi­vid­u­als or objects that have sim­i­lar char­ac­ter­is­tics to the uni­verse (total pop­u­la­tion), from which sam­ples can be drawn for a poll. The next issue is then how the sam­ples are drawn. Sam­ples have to be select­ed ran­dom­ly from the pop­u­la­tion. In oth­er words, sam­ples should be select­ed in such a way that every indi­vid­ual or object with­in the research pop­u­la­tion has an equal oppor­tu­ni­ty of being select­ed. This is called the prin­ci­ple of ran­dom­iza­tion.

The con­cept of sam­pling

As a poll­ster, one ques­tion I typ­i­cal­ly get asked bor­ders on the issue of sam­ple size – how can a mere 1,000 or 5,000 sam­ple size poll tell us what over 180 mil­lion Nige­ri­ans are think­ing on a par­tic­u­lar issue? My con­stant response to those who ask is sim­ple – it isn’t about the size of the sam­ple, but the selec­tion of the sam­ple. As we often quip in polling par­lance, when you fall ill and vis­it the hos­pi­tal, the lab sci­en­tist doesn’t need to draw out your entire blood to check what’s wrong with you. He sim­ply takes a tiny lit­tle blood sam­ple for the test. This is the exact prin­ci­ple applied in pub­lic opin­ion polling. We mustn’t sam­ple 1 Mil­lion Nige­ri­ans to know what Nige­ri­ans think on any mat­ter. This con­cept has long been estab­lished in social sci­ences. In 1936, it was Gallup who demys­ti­fied polling and dealt a major upset on the US pres­i­den­tial elec­tion pre­dic­tion of Lit­er­ary Digest. The mag­a­zine had sent out over 2 mil­lion dum­my bal­lots from vehi­cle reg­is­tra­tion data­base and tele­phone direc­to­ries, at con­sid­er­able time and cost, in order to pre­dict the elec­tion result. Gallup, on the oth­er hand, used a sam­ple of only 5,000 to pre­dict that Roo­sevelt would take the lead in at least 40 states and car­ry the pop­u­lar vote by 56 per­cent to 44 per­cent mar­gin. Need­less to say, it was a humil­i­at­ing upset for Lit­er­ary Digest, as Gallup suc­ceed­ed in estab­lish­ing that as far as sur­vey sam­ple size is con­cerned, more isn’t always bet­ter. There­fore, a sci­en­tif­i­cal­ly select­ed sam­ple of the gen­er­al pop­u­la­tion was not only much cheap­er and eas­i­er to han­dle, it would pro­duce more accu­rate results.

Good vs Bad Polls

With­out sound­ing aca­d­e­m­ic, there are cer­tain fea­tures that ought to be report­ed in poll reports to be ticked as good. Con­verse­ly, where they do not exist, the find­ings of such polls can sim­ply be con­sid­ered as bad polls. Orga­ni­za­tions like Amer­i­can Asso­ci­a­tion for Pub­lic Opin­ion Research (AAPOR) and the Nation­al Coun­cil for Pub­lic Polls (NCPP) have pro­vid­ed cer­tain stan­dards of dis­clo­sure to guide polling firms. While jour­nal­ists and poll con­sumers may not be able to tell or assess the qual­i­ty con­trol cri­te­ria adopt­ed on the poll; it is expect­ed that good polls should pro­vide infor­ma­tion regard­ing: (1) Sam­pling (from what pop­u­la­tion was the sam­ple select­ed? was data col­lect­ed using a ran­dom or non-ran­dom sam­ple? how many inter­views were com­plet­ed?), (2) Data col­lec­tion (how was data col­lect­ed? was data col­lect­ed using face-to-face, tele­phone, web-based or SMS tech­nique? (3) Ques­tion word­ing (were ques­tions word­ed in a clear and neu­tral man­ner?), (4) Field­ing dates (what were the spe­cif­ic dates of data col­lec­tion?), and (5) Response rate (some sense of how many peo­ple were con­tact­ed and how many respond­ed is always use­ful).

I find most poll con­sumers are unaware of these fea­tures, con­se­quent­ly poll results and reports are sim­ply inter­pret­ed based on the reader’s pre­mo­ni­tion and world­view. In Nige­ria for instance, politi­cians are quick to cel­e­brate polls that seem to be in line with their pre­dis­po­si­tion, and shoot down polls that do not con­form. They are part­ly respon­si­ble for the pro­lif­er­a­tion of fake or bad polls, which seek to sup­port their polit­i­cal lean­ings. Inter­est­ing­ly, there’s usu­al­ly a sea­son­al upsurge in opin­ion polling activ­i­ties and report released dur­ing an elec­tion year; and as we build up to 2019 elec­tions there would be many fly-by-night poll­sters on the loose. The ques­tions for dis­cern­ing con­sumers of opin­ion poll reports should be – where have these poll­sters been before now? Should we assume they are new or bud­ding polling firms? And if they are, great! But would they con­tin­ue their polling ser­vices after elec­tions have come and gone?

So, when next you find your­self con­sum­ing the report from a pub­lic opin­ion poll, just before you hail it or shoot it down, depend­ing on your socio-polit­i­cal lean­ing, be sure to spend a few more min­utes to ascer­tain how it was con­duct­ed using some of the tips dis­cussed above. That way, you would be bet­ter armed to judge for your­self if the poll should be giv­en some atten­tion or sim­ply dis­card­ed as a mere aca­d­e­m­ic exer­cise.

Dr. Ihua is Exec­u­tive Direc­tor at Africa Polling Insti­tute (API), an Opin­ion Research think-tank based in Abu­ja. He was for­mer­ly a fac­ul­ty mem­ber at Coven­try University’s Busi­ness School, Unit­ed King­dom.

Leave a Reply

Your email address will not be published. Required fields are marked *