Fake news: characterization of different individual profiles in relation to different news topics

i FAKE NEWS: CHARACTERIZATION OF DIFFERENT INDIVIDUAL PROFILES IN RELATION TO DIFFERENT NEWS TOPICS Ma ia F ancisca de Sousa e Al im Lima de Ba os Disse a ion p esen ed as pa ial equi emen o ob aining he Mas e ’s deg ee in In o ma ion Managemen ii NOVA In o ma ion Managemen School Ins i u o Supe io de Es a ís ica e Ges ão de In o mação Uni e sidade No a de Lisboa FAKE NEWS: CHARACTERIZATION OF DIFFERENT INDIVIDUAL PROFILES IN RELATION TO DIFFERENT NEWS TOPICS by Ma ia F ancisca de Sousa e Al im Lima de Ba os Disse a ion p esen ed as pa ial equi emen o ob aining he Mas e ’s deg ee in In o ma ion Managemen , wi h a specializa ion in In o ma ion Sys ems and Technologies Managemen Co-Ad iso : P o esso Dou o Flá io Luis Po as Pinhei o Co-Ad iso : P o esso a Dou o a Diana O ghian July 2021 iii DEDICATION To A ó Guidinha and A ó T iz. i ACKNOWLEDGEMENTS This hesis is he las miles one o comple e my mas e ’s deg ee. I challenged me in many ways and was a e y posi i e lea ning expe ience. I wan o use his oppo uni y o acknowledge all who made he conclusion o his disse a ion possible. Thank you o my ad iso s P o esso Diana O ghian and P o esso Flá io Pinhei o. Thank you o he guidance, sha ed knowledge, ideas and o all he help and suppo p o ided du ing his p ocess. Thank you o unde aking his s udy o he dissemina ion o ake news in Po ugal wi h me. I wan o acknowledge my g a i ude o all pa icipan s in he su ey, wi hou whom would no be possible o ga he da a o his s udy. Las ly, hank you my pa en s F ancisca and Luís and my b o he Gui. Fo he lo e, encou agemen and o being my bigges ans. ABSTRACT The exis ence o ake news is an ex emely opical conce n which calls in o ques ion he e aci y o he b oadcas ed in o ma ion. Since nowadays he sea ch and p oduc ion o news is mainly done online, he cos s wi h con en p oduc ion a e low and he con en ’s each and speed o p opaga ion is e y high. These ac o s acili a e he dissemina ion o ake news in social pla o ms ha a e no specialized means o communica ion, namely in online social ne wo ks. The e o e, his s udy aims o cha ac e ize di e en p o iles o Po uguese indi iduals based on hei suscep ibili y o se e al news opics. The a ainmen o he men ioned p o iles is going o be a aluable con ibu ion o in o ma ion managemen and i is going o allow u u e de ini ion o measu es o mi iga e he p opaga ion o ake news in social pla o ms. To achie e his, c i ical li e a u e e iew was done and accompanied by he c ea ion o a su ey o analyze how academic backg ound and opic o he news pieces in luence he accu acy o indi iduals iden i ying alse news. This disse a ion in en s o unde s and i he e is anyone immune o ake news, o i indi iduals can be mo e o less immune depending on he opic. KEYWORDS Fake news; Social Ne wo k; Opinion Dynamics; Social In e en ions; De ec ion i INDEX 1. INTRODUCTION ............................................................................................................................. 10 2. LITERATURE REVIEW ..................................................................................................................... 12 2.1. HISTORY – FALSE NEWS ARE NOT A NOVELTY ................................................................................. 12 2.2. FALSE NEWS DEFINITION ............................................................................................................. 13 2.3. IT IS STILL A CURRENT TOPIC – PARADIGM SHIFT ............................................................................. 13 2.4. SOCIAL MEDIA AND ITS CHARACTERISTICS THAT ENABLE THE SPREAD OF FALSE NEWS ........................... 13 2.5. WHY IS THE SPREAD A PROBLEM? ................................................................................................ 16 2.6. WHY DO INDIVIDUALS BELIEVE IN FALSE NEWS? ............................................................................. 17 2.7. HOW CAN THE SPREAD OF MISINFORMATION ONLINE BE REDUCED? .................................................. 18 3. METHODOLOGY ............................................................................................................................ 23 3.1. STUDY DESIGN .......................................................................................................................... 23 3.1.1. SURVEY PART ONE – INTRODUCTION ........................................................................................ 24 3.1.2. SURVEY PART TWO – ACCURACY AND PREVIOUS KNOWLEDGE ...................................................... 25 3.1.3. SURVEY PART THREE – COMPREHENSION OF NEWS HEADLINES ..................................................... 27 4. STUDY RESULTS AND DISCUSSION ................................................................................................... 29 4.1. DESCRIPTIVE ANALYSIS ......................................................................................................... 29 4.1.1. STUDY PARTICIPANTS ............................................................................................................. 29 4.1.2. RESPONDENTS PERCEPTION OF NEWS HEADLINES ....................................................................... 34 4.1.3. NEWS CATEGORIES ................................................................................................................ 36 4.1.4. DESCRIPTIVE ANALYSIS OF THE NEWS ........................................................................................ 38 4.1.5. COMPARISON BETWEEN STEM AND NON-STEM BACKGROUND ...................................................... 39 4.2. INFERENTIAL STATISTICS.............................................................................................................. 42 4.2.1. T-TEST ................................................................................................................................. 42 4.2.2. ANALYSIS OF VARIANCE - ANOVA AND MANOVA ........................................................................ 42 4.2.3. PEARSON’S CORRELATION ....................................................................................................... 43 4.2.3.1. TRUE NEWS ...................................................................................................................... 44 4.2.3.2. FALSE NEWS ...................................................................................................................... 45 4.3. DISCUSSION .............................................................................................................................. 47 5. CONCLUSIONS ............................................................................................................................... 49 6. LIMITATIONS AND RECOMMENDATIONS FOR FUTURE WORKS ................................................. 50 7. BIBLIOGRAPHY .............................................................................................................................. 51 8. APPENDIX ...................................................................................................................................... 55 ii LIST OF FIGURES Figu e 1 - Dis ibu ion o Responden s by A ea o Residence ………………………………………………..29 Figu e 2 - Pa icipan s’ Educa ion Le el ……………………………………………………………………………….30 Figu e 3 - Pa icipan s’ F equency o Use o Social Ne wo ks………………………………………………..31 Figu e 4 - Responden s’ Use o Social Media…………………………………………………………………………32 Figu e 5 - Pa icipan s’ Numbe o Twi e Followe s ……………………………………………………………33 Figu e 6 - Pa icipan s' Numbe o F iends on Facebook ……………………………………………………….33 iii LIST OF TABLES Table 1 - News P esen ed o Responden s Du ing he Su ey ……………………………………………….27 Table 2 -Desc ip i e S a is ics o S udy Va iables ………………………………………………………………..35 Table 3 – Compa ison Be ween T ue and False News …………………………………………………………..36 Table 4 – Desc ip i e Analysis o he Six Fea u es pe Ca ego y …………………………………………….38 Table 5 - Compa ison Be ween STEM and non-STEM Responden s - Desc ip i e S a is ics ……40 Table 6 - Accu acy o STEM and non-STEM Responden s pe Ca ego y o News ……………………41 Table 7 - Accu acy o STEM and non-STEM Responden s pe Ca ego y o News on T ue and False News ………………………………………………………………………………………………………….………42 Table 8 - One-way ANOVA - E ec o News Ve aci y on Accu acy …………………………..………….43 Table 9 - MANOVA - E ec o News' Ca ego y on Accu acy …………………………………………………43 Table 10 – T ue News - Pea son's Co ela ion Be ween Accu acy and O he Fea u es ………….45 Table 11 – False News - Pea son's Co ela ion Be ween Accu acy and O he Fea u es ………….47 ix LIST OF ABBREVIATIONS AND ACRONYMS STEM Science, Technology, Enginee ing and Ma hema ics ANOVA Analysis o Va iance MANOVA Mul i a ia e Analysis o Va iance ACC. Accu acy KNOW. Knowledge COMPR. Comp ehension TEC.TER. Technical Te ms INT. In e es VAG. Vagueness 16 sp ead o ideas can be s udied acco ding o an epidemiological model, he sp ead o alse news can also be analyzed wi h he help o simila me hods. As s a ed abo e, he mains eam sp ead o alse news can be enhanced mainly online by di e se ac o s such as he na u al cha ac e is ics o online media, he indi iduals ha a e pa o one’s ne wo k, he o ma ion o homogeneous and pola ized g oups in social media, as o example echo chambe s, o by use o social bo s. 2.5. WHY IS THE SPREAD A PROBLEM? As e idenced by Vosoughi e al. (2018) when s udying he di usion o ue and alse s o ies dis ibu ed on Twi e , alsehood di uses a he , as e , deepe , and mo e b oadly han he u h ega dless o he ca ego y o in o ma ion. E en hough he in luence is ha d o quan i y o de ine specially, in he medium o long un he exposi ion o misin o ma ion will na u ally impac on indi iduals’ belie s (Laze e . al., 2018). So, his sugges s ha a piece o alse news in an online social ne wo k is going o sp ead as e han any ue piece o in o ma ion and impac people who encoun e i . Al hough he quali y o he in o ma ion is conside ed as a mo e and mo e impo an issue (Bun ain, C., & Golbeck, J., 2017) and ue is cen al o he good unc ioning o li e in socie y, he sp ead o misin o ma ion online is inc easingly common. This wide dissemina ion o alse in o ma ion ep esen s a se ious h ea o gene al us in he mass media. The exposi ion o alse news is p oblema ic o socie y, because indi iduals a e going o base hei iews, alues, and ul ima ely hei decisions in he alsehood hey a e exposed o. Hence, aking alse news as ue ep esen s a p oblem since i leads o inaccu a e belie s and causes pa isan disag eemen o e se e al issues ha can e en be basic ac s (Pennycook e al., 2019). Hence, he sp ead o alse news is a p oblem and should be aken in o conside a ion as a g ea conce n in socie y and i s s udy is cu en ly a c ucial key-issue o in o ma ion managemen . Du ing he global pandemic, misin o ma ion has been a cons an . The isk emains iden ical: indi iduals will ake alse in o ma ion as a basis o decide (Vaidyana han, G. 2020). In a pandemic, i is e en mo e likely ha he ac ions based upon alse in o ma ion endange e e yday li e. 17 2.6. WHY DO INDIVIDUALS BELIEVE IN FALSE NEWS? The ques ion emains: why do people belie e in alse news? Bond & De Paulo (2006) sugges ha humans a e no so good de ec ing lies on ex : he indings suppo ha people a e only 4% be e dis inguishing ue om alse s a emen s han i hey decided i andomly. In con as , Pennycook e al. (2019) disag ees wi h his “pos - u h” na a i e, and ins ead a gues ha people a e no awa e o he ac ha hey a e sha ing misin o ma ion on social media in he momen hey decide o sha e i . The s udy s a es he e is a disconnec be ween sha ing in en ions and accu acy judgmen s. In o he wo ds, he e is a dissocia ion be ween he con en indi iduals conside accu a e and he ac ual con en hey sha e on social media. The indings poin ou ha he p oblem is jus dis ac ion om accu acy when sha ing con en . When asked, he majo i y o pa icipan s in an expe imen hey conduc ed said i is impo an o only sha e accu a e news. Thus, e en hough people can dis inguish ue and alse in o ma ion, hey won’ ake he in o ma ion accu acy in conside a ion when sha ing i . Ins ead, he ocus is se on o he mo i a ions, such as willingness o please ollowe s, engagemen wi h speci ic con en o e en o show o hei ne wo k ha hey belong o a speci ic g oup. The e o e, sha ing alse pieces o news does no indica e pe se ha he indi idual belie es in his in o ma ion. I is possible ha he is only dis ac ed om he accu acy o he con en . The ac ha a piece o news is conside ed ue o alse by an indi idual migh also be a ec ed by social no ms o by he ela ion wi h he indi iduals’ belie s (Del Vicca io e al., 2016). Se e al s udies conduc ed show ha indi iduals p e e in o ma ion ha suppo s hei p e-exis ing a i udes. By doing his, people show a con i ma ion bias, which means ha he in o ma ion cohe en wi h hei own belie s is pe cei ed as mo e con incing han disco dan in o ma ion. Humans also ha e a desi abili y bias, i. e., hey a e d awn o accep in o ma ion ha is mo e socially accep able o desi able and he e o e sa is ies hem (Laze e . al, 2018). The epe i ion o he same in o ma ion – ac oss se e al pla o ms o h oughou di e en connec ions in he ne wo k, o example – mus be conside ed when s udying he belie in alse news. Indi iduals a e mo e likely o accep in o ma ion ha is epea ed as ue (Swi e e al., 2017). People a e also going o be d awn o in o ma ion ha suppo s hei psychological mo i es, as i can happen when conside ing he belie in conspi acy heo ies. They a e explana ions o impo an e en s ha in ol e sec e plo s by powe ul and male olen g oups (Goe zel, 1994). Conspi acy heo ies a e also 18 usually specula i e and con adic o y, conside he public igno an and subjuga ed o unaccoun able powe s. (Douglas e . al., 2017). Indi iduals a e d i en o belie e in his appa en ly unbelie able con en when i p omises o sa is y one’s desi e o unde s anding, accu acy, subjec i e ce ain y o o he social mo i es. Conspi acy heo ies a e conside ed alse news and he mo i es which d aw people o conspi acy belie can be he same ha d aw people o alse news. The epis emic mo i es behind conspi acy use causal explana ions o educe unce ain y when in o ma ion a ailable is con lic ing, y o ind meaning when he e en s a e appa en ly andom and can de end belie s om e usal. This human desi e o unde s anding and accu acy can lead o using conspi acy heo ies o p o ec one’s belie s when acing unce ain y, con adic ions, o e en when he e is no in o ma ion a ailable (Douglas e . al, 2017). The belie in a conspi acy heo y may be s onge when he e en s conside ed a e pa icula ly signi ican and lea e indi iduals dissa is ied wi h small explana ions. (Leman & Cinni ella, 2013). The dissa is ac ion can lead o a belie in alse in o ma ion since i may ul il he abo e e e ed indi idual mo i a ions. I is also possible ha belie s in conspi acy heo ies a e g ounded in social mo i es, namely in he desi e o belong and main ain bo h a posi i e image o onesel and o he g oups’ one belongs o. In he same way people could be d awn owa d conspi acy heo ies ha seem o a i m o alida e hei poli ical iews, indi iduals can be a ac ed o misin o ma ion ha a o s hei poli ical posi ion. The a o emen ioned ac o s d i e conspi acy belie . Bu i is s ill no clea i he belie s in conspi acy scena ios ul il hese mo i a ions. (Douglas e . al., 2017). Social homogenei y is ano he ac o which can ein o ce a belie in alse news. Usually, indi idual belie s lean owa ds an alignmen wi h he alues o he communi y ha he is pa o . (Laze e . al, 2018). When agg ega ed in homogeneous clus e s i is easy ha indi iduals may selec and sha e con en ha ela ed only o a pa icula plo , igno ing wha does no align wi h i . Consequen ly, hese communi ies enhance opinion pola iza ion based mainly in alse na a i es, such as umo s o conspi acy heo ies. (Del Vicca io e al., 2016). 2.7. HOW CAN THE SPREAD OF MISINFORMATION ONLINE BE REDUCED? The e a e se e al s udies de eloped on how o educe he sp ead o alse news online. Laze e . al (2018) p oposes wo di e en ca ego ies o s a egies o minimizing he sp ead o alse news. Fi s , hey aim o awake one’s c i ical awa eness by helping o iden i y alse news based on an e alua ion o he in o ma ion 19 encoun e ed. This could be achie ed h ough exposi ion o ac checked pieces o news, o ins ance. Howe e , his is conside ed ha d because mos indi iduals will only ques ion he c edibili y o he in o ma ion i iola es hei belie s o p econcei ed ideas (Vo e al., 2020). Second, i is sugges ed ha s uc u al changes on social media can p e en he exposu e o indi iduals o alse news in he i s place (Laze e . al, 2018). Thus, he highligh gi en o in o ma ion wi h c edible sou ces should be maximized so ha an accu a e sou ce would be a equi emen o a ending in o ma ion. Social media si es such as Facebook al eady pa ne wi h ex e nal ac checke s ce i ied h ough In e na ional Fac -Checke Ne wo k o iden i y and e iew alse pieces o news. None heless, i is s ill no enough because se e al kinds o ad e isemen s and published pieces canno be e oed. A di e en pe spec i e a gues ha d awing social media use s’ a en ion o he concep o accu acy can lead o a educ ion o he amoun o misin o ma ion hey sha e. Tha is sugges ed because indi iduals who a e able o de e mine he accu acy o he con en a e dis ac ed om i in social media, ocusing hei a en ion on o he aspec s. In addi ion, social media pla o ms a e designed o quick sc oll and engagemen wi h se e al ypes o con en and may no in i e use s o pa icipa e in c i ical hinking. The e o e, jus by highligh ing accu acy o e en he echnics used o sp ead misin o ma ion, use s can be mo e awa e o i s e aci y when sha ing con en (Mae ens e al., 2020). The e is also li e a u e ha ocus on he ole o use s conside ed hubs in he ne wo k, who consequen ly wo k as in o ma ion b oke s. Hoch ei e e al. (2013) uses p e ious s udies o sugges ha human in e ac ion is o ganized in small-wo ld ne wo ks ha de ine how pe sons a e linked o each o he . A small- wo ld ne wo k has se e al speci ic p op ie ies and is cha ac e ized by he ac ha is possible o connec any wo e ices among he ne wo k jus h ough a ew links (Ama al e al., 2000). In his way, in a small- wo ld ne wo k a ew indi iduals connec mos indi iduals. The ones who es ablish mos connec ions wo k as hubs, being he connec ion’s b oke s. (Hoch ei e e al., 2013) The e o e, in an online social ne wo k i a “hub” indi idual il e s alse in o ma ion and does no pos i , he can make his ne wo k mo e obus o alse news. Likewise, i a use sp eads u h ul and accu a e in o ma ion, he e is a high p obabili y ha i is going o each o he indi iduals' pa o his ne wo k. Social in luence can ha e a signi ican ole in he educ ion o misin o ma ion online. I is desc ibed as he p ocess whe e indi iduals adjus hei opinion and change hei beha iou o belie s as a esul o a se o social in e ac ions wi h o he indi iduals (Moussaïd e al., 2013). I is a c ucial elemen in o ming public 20 opinion. Social in luence sugges s ha in e ac ions be ween indi iduals a e highly esponsible o each indi idual opinion. In an online social ne wo k, he elemen s o a use s’ ne wo k a e going o in luence his iews and he news ha he is exposed o. Consequen ly, a change o iends o ollowe s can make he sys em mo e obus o alse news. Conside ing ha indi iduals do no exis isola ed om socie y and a e pa o di e en ne wo ks o pe sons, social science p esen s h ee ac o s o explain how indi iduals assimila e in o ma ion. Social con agion is one o hese ac o s and a gues ha one abso bs in o ma ion h ough a ne wo k (Tö nbe g e al, 2018). Howe e , no all beha iou s in a social ne wo k sp ead di ec ly om one pe son o ano he . Fo an indi idual o adop a beha iou , i akes ein o cemen o he same beha iou by se e al people (Vaidyana han, G. 2020). In addi ion, one’s wo ld iew is subs an i e when conside ing how indi iduals ag ee wi h in o ma ion. Cul u e, belie s and li e expe ience ha e a g ea deal o in luence on how someone eac s o a gi en piece o news. Indi idual pe spec i e and wo ld iew in luence wha is aken as ue and alse, and his can be one o he easons why ac checking alone is no e icien o igh misin o ma ion. People pe cei e hei iews as some hing hey canno le go in o de no o jeopa dize hei social ci cle s a us. As men ioned be o e, he way o igh misin o ma ion is much mo e complex han p o iding indi iduals wi h ue in o ma ion and hope i is going o be chosen o e alse in o ma ion (Vaidyana han, 2020). This di icul y in igh ing misin o ma ion is p o en by he ac ha e en hough he sp ead o ake news is a widely s udied heme, he conc e e p oblem p e ails, and indi iduals engage daily wi h se e al alse pieces o news. This disse a ion p ojec is designed o s udy he ac o s subjacen o he dissemina ion o alse news. The aim o his hesis is o analyze how indi iduals’ backg ound – academic and p o essional and he opic o news impac he accu acy o indi iduals iden i ying alse news. This disse a ion is done unde he hypo hesis ha no one is immune o ake news. Wha happens is ha indi iduals a e mo e suscep ible o belie e and ake alse news om ce ain opics as ue. The accu acy iden i ying alsehood migh also be a ec ed by ac o s such as ela ion wi h an a ea o expe ise, social no ms o indi iduals’ belie s. Mo eo e , humans ake easily he in o ma ion hey ge om social media, mos imes wi hou wonde ing abou i s accu acy (Pennycook e al., 2021). I is no only social ne wo ks ha lead indi iduals o belie e in alse news, bu o en imes i is he people who a e pa o one’s ne wo k. 21 Thus, he people who connec wi h an indi idual can be esponsible o his exposi ion o misin o ma ion (Balmau e al., 2018). Ensu ing ha a ne wo k is composed o pe sons wi h di e en sensi i i y o di e en news ca ego ies can be a s a egy o make a g oup o indi iduals mo e obus , and he e o e i is mo e di icul o sp ead ake news. The su ey was de eloped o es he ollowing esea ch ques ions: [RQ1: The ac ha a piece o news is u h o alse has an e ec on accu acy iden i ying he news as ue o alse? Does being a ue o alse piece o news e lec s on esponden s' p e ious knowledge o he news and comp ehension? Does i e lec on pe cei ed use o echnical e ms, agueness o esponden s’ in e es in he news?] We ha e he pe cep ion ha he ac ha a piece o news is ue o alse in luences how indi iduals cha ac e ize i accu a ely as ue o alse. As hypo hesis, a piece o news e aci y e lec s has an impac on p e ious knowledge o he news, in e es and ease o comp ehension. Being a ue o alse new s o y also in luences eade s’ pe cei ed use o echnical e ms and agueness. [RQ2: The ca ego y o he news (Spo , Science, Cul u e, Heal h, Economy, Socie y, Poli ics) has an e ec on classi ying news as accu a ely?] Since indi iduals do no ha e he same amoun o knowledge o awa eness ega ding di e en opics, i is likely ha hey eac dis inc ly when con on ed wi h news ega ding di e en hemes. Thus, wi h his esea ch ques ion i is aimed o es i he news opic impac s an accu a e classi ica ion o news headlines as ue o alse. [RQ3: Does an indi idual’s backg ound – academical and p o essional - in luence his sensi i i y o alse news?] Addi ionally, indi iduals’ backg ound – age, educa ion, p o essional expe ience – in luences how in o ma ion is pe cei ed. Consequen ly, his hypo hesis was o mula ed o es i one’s backg ound in luences sensi i i y owa ds alse pieces o news. Wi h his s udy, we expec o ind i :  Indi iduals iden i y mo e accu a ely ue o alse news. 22  Po uguese people accu acy iden i ying i a piece o news is ue o alse is ela ed wi h hei p e ious knowledge o he s o y, in e es , pe cei ed use o echnical e ns o agueness o he s o y.  The e a e speci ic ac o s o cha ac e is ics - o example age, le el o educa ion o o e all use o social media - ha in luence Po uguese indi iduals’ c i ical sense in ela ion o di e en opics o news.  Academical and p o essional backg ound has an e ec on accu acy classi ying he e aci y o news. 23 3. METHODOLOGY In his sec ion is p esen ed he me hodology employed o ca y ou his s udy. In he i s place, a c i ical e iew o exis ing li e a u e was done. The p e ious li e a u e and esea ch we e used o de ine some o he heo e ical concep s in he basis o his disse a ion. The esea ch a ailable was used in o de o unde s and he scope and ele ance o he p oblem unde s udy and i s ela ionship wi h exis ing knowledge. To comple e he in o ma ion collec ed when de eloping he li e a u e e iew, a ques ionnai e des ined o he da a collec ion was designed and buil . A ques ionnai e is a w i en ool ha p esen s i s esponden s wi h a se ies o ques ions o s a emen s, so ha hey answe by selec ing one o he esponse op ions o by answe ing openly (B own, 2001). The e o e, he ques ionnai e is a s uc u ed in e iew, whe e each esponden answe s o ques ions in a p e iously o ganized o m (B ace, 2008). The se ies o ques ions mus be o ganized iden ically o all esponden s (Chi omo, 2006). Ques ionnai es ca y se e al ad an ages such as low da a collec ion cos s, possibili y o each geog aphically dis an loca ions and a high numbe o indi iduals. The absence o a bias due o he absence o he in e iewe can also be conside ed as an ad an age (Magwa & Magwa, 2015). Th ough he analysis o he in o ma ion collec ed, i is going o be possible o cha ac e ize se e al g oups o people conside ing hei beha io in ela ion o di e en opics o news. The ques ionnai e is going o p esen se e al news i les o esponden s. Ti les a e a highly dis inc i e ac o when dis inguishing alse om ue news pieces. Gene ally, he i le holds he piece o news's main a gumen and allows eade s o skip he ull a icle. Thus, alse news a e a ge ed o indi iduals who will no ead beyond he i le (Ho ne, B., & Adali, S., 2017). The news we e ex ac ed om Po uguese ac -checke websi e Políg a o and subsequen ly alida ed by he esea che in o de be ween 1s o Augus 2019 and 28 h o Feb ua y 2020. 3.1. STUDY DESIGN The s udy was app o ed by NOVA IMS Resea ch E hics Commi ee and complied wi h all e hical aspec s. The ques ionnai e (Appendix A) was de eloped as pa o he s udy o analyze he suscep ibili y o 24 di e en g oups o Po uguese people in ela ion o di e en opics o news. He eby, he s udy is a ge ed o Po uguese ci izens and Po uguese na i e speake s and was deployed in June 2020. The ques ionnai e was designed wi h h ee pa s. In he i s pa , he e was an in oduc ion o he s udy. Responden s consen ed o pa icipa e in he s udy and we e p esen ed wi h a se o ques ions ega ding one’s backg ound and social media usage. The second pa p esen ed 14 di e en pieces o news which esponden s classi ied as ue o alse and indica ed whe he hey knew he news p e iously o no . Be ween he second and hi d pa s o he su ey he e was a ques ion used as an a en ion check ( o mo e see Appendix G). Ul ima ely, i was no u ilized in he analysis since was conside ed oo icky and conse a i e. The hi d pa con ained a comp ehension analysis, done h ough se en di e en pieces o news. The ques ionnai e was designed o a ge Po uguese na ionals and he e o e en i ely w i en in Po uguese. 3.1.1. SURVEY PART ONE – INTRODUCTION In he i s pa o he ques ionnai e, esponden s we e in oduced o i wi h a welcome message (Appendix A) which explained he ques ionnai e objec i es, i s a e age du a ion and he equi emen s o pa icipa e in he s udy. The ques ionnai e welcome message in o med ha o be pa o he su ey i was equi ed o be Po uguese o luen in Po uguese and o e 18 yea s old. To pa icipa e in s udy esponden s had o gi e hei consen . The ea e , esponden s we e asked abou demog aphic backg ound da a: yea o bi h, gende , dis ic o esidency, las le el o educa ion concluded, p o ession and cu en a ea o wo k. In case esponden s s a ed ha hei p o ession was “S uden ”, hey we e asked abou hei a ea o specializa ion a he han hei a ea o wo k. To s udy social media usage and online in luence, he i s pa o he ques ionnai e also included ques ions abou social media u iliza ion. Pa icipan s we e ques ioned how o en hey used social media and which social ne wo ks hey used among he ollowing: Facebook, Ins ag am, Twi e , LinkedIn and Reddi . Fo each social ne wo k selec ed, pa icipan s we e ques ioned abou how many iends, ollowe s o connec ions hey possessed. Las ly, pa icipan s we e asked how o en hey sha ed news on social media. 25 3.1.2. SURVEY PART TWO – ACCURACY AND PREVIOUS KNOWLEDGE To unde s and how Po uguese people pe cei e he accu acy o di e en opics o news i was equi ed o de elop he second pa o he ques ionnai e. I s a ed wi h he ollowing message: Now, I would ask you o classi y as ue o alse se e al news dis ibu ed by Po uguese media in ecen mon hs. Each esponden was p esen ed o a se o 14 di e en pieces o news. Following each piece o news, esponden s we e asked wo ques ions: is his piece o news ue o alse? and did you al eady know his piece o news? I he esponden answe ed he al eady knew he news a icle, he was asked wha was he sou ce ha in o med you abou his piece o news? To answe his ques ion, he gi en op ions we e ia mass media (newspape s, magazines, ele ision, social media), old o ally by someone else, o o he . In his las op ion, esponden s could ype o speci y he sou ce. To c ea e his sec ion o he ques ionnai e, h ee g oups wi h 14 i les o news we e de eloped ( o mo e see Table 1). Each g oup consis ed o wo news om each ca ego y – Science, Hea h, Cul u e, Spo , Socie y, Economics and Poli ics. F om he wo pieces o news in each ca ego y, one was ue and one was alse. The e o e, each esponden encoun e ed se en ue and se en alse news s o ies, displayed in a andomized o de . Table 1 News P esen ed o Responden s Du ing he Su ey ca ego y ype news Science T ue The closes black hole o ou plane , which is a housand ligh yea s away om Ea h, has been disco e ed. Science False A s udy by Ha a d scien is s concludes ha ca ying weapons educes c ime. Cul u e T ue In 2017, 200 housand eu os o public unds we e spen o p omo e he au omaquica cul u e so ha i could be classi ied as Imma e ial Cul u al He i age o Po ugal unde he UNESCO Con en ion. Cul u e False Leo Fende , he c ea o o he well-known Fende gui a s, was a alen ed gui a is . Heal h T ue The Cons i u ion o he Po uguese Republic does no allow manda o y hospi aliza ion in case o con agious diseases. Heal h False Be ween 1950 and 2020, China was a he o igin o nine di e en epidemics - H2N2, H3N2, H5N1, SARS, a ian lu, swine lu, swine e e , Co ona i us, Co id- 19. 32 Apa om being inqui ed abou hei social media usage equency, pa icipan s we e also asked which social ne wo ks hey u ilized. Facebook was he mos used social ne wo k (144 pa icipan s, 74.23%), ollowed closely by Ins ag am (132 pa icipan s, 68.04%) among i s use s. LinkedIn was used by 81 pa icipan s (41.75%), being he e o e he hi d mos widely used social ne wo k. Twi e and Reddi we e he wo less used social ne wo ks, being used by 23 (11.86%) and 7 (3.61%) pa icipan s, espec i ely. Fo mo e, see Figu e 4. Figu e 4 Responden s’ Use o Social Media F om all social ne wo ks, Twi e and Facebook we e he only wo social media si es men ioned in his hesis p e ious analysis o li e a u e. The e o e, we will analyze pa icipan s’ numbe o Twi e ollowe s and Facebook iends wi h mo e de ail. Fo mo e on Twi e ollowe s, see Figu e 5. 33 Figu e 5 Pa icipan s’ Numbe o Twi e Followe s Mos s udy pa icipan s (78.3%) who use Twi e ha e be ween ze o and 200 ollowe s. Only 21.7% o pa icipan s wi h a Twi e accoun ha e mo e han 200 ollowe s. Conside ing Facebook, he end is e e sed. Mos o pa icipan s in he s udy (60.4%) ha e mo e han 600 ollowe s, while he mino i y o esponden s (39.6%) has less han 600 ollowe s. Figu e 6 Pa icipan s’ Numbe o F iends on Facebook 34 In o de o unde s and esponden s’ po en ial impac in hei ne wo ks, esponden s we e asked abou how o en hey sha ed pieces o news on social media. Mos esponden s (130 esponden s, 67%) sha e news online wi h hei connec ions. Among hese, h ee esponden s (1.5%) sha ed news on social media se e al imes a day, 10 esponden s (5.2%) decla ed ha hey sha ed news daily and 15 esponden s (7.7%) sha ed news se e al imes a week. In addi ion, 25 pa icipan s (12.9%) sha ed news weekly on social media, 51 pa icipan s (26.3%) sha ed news wi h a mon hly equency and 26 pa icipan s (13.4%) sha e news on a yea ly basis. Con e sely, 64 esponden s (33%) decla ed hey ne e sha e news on social ne wo ks. Fo s udy pu poses he pa icipan s we e di ided in wo g oups acco ding o bo h hei academical and p o essional backg ound: STEM (n = 87) and non-STEM ields (n = 95). 4.1.2. RESPONDENTS PERCEPTION OF NEWS HEADLINES A e answe ing ques ions abou backg ound and social media usage, esponden s classi ied pieces o news as ue o alse, iden i ied i hey we e p e iously awa e o he news and analyzed se e al ea u es o he news pieces. In his wo k we ocus on he analysis o he ollowing six ea u es:  Accu acy, gi en by he a io be ween he numbe o esponden s who classi ied accu a ely he piece o news as ue o alse by he o al amoun o esponden s who classi ied he piece o news du ing he su ey.  Knowledge, compu ed om he ac ion o he numbe o esponden s who al eady knew he piece o news p esen ed in he ques ionnai e and he o e all numbe o esponden s o ha ques ion.  Comp ehension, ob ained by he mean o an o e all sco e on a scale om one – e y di icul - o six – e y easy. The sco e was awa ded by each esponden , conside ing he piece o news in ques ion.  Use o echnical e ms, gi en by he mean o an o e all sco e o classi y he use o echnical e ms in a piece o news on a scale om one – none - o six – e y much. Each pa icipan who aced he piece o news was equi ed o assign a sco e. 35  In e es , ob ained by he mean o he sco e a ibu ed by each esponden . This sco e e e ed o esponden s’ in e es in he piece o news, wi h sco es anged om one - none - o six - e y much.  Vagueness, compu ed by he mean o he sco e gi en by each pa icipan o he piece o news. News we e classi ied as conc e e and de ailed o abs ac and ague using a sco e anged om one - e y conc e e - o six - e y ague. O e all, esponden s could classi y accu a ely he pieces o news (N = 42) p esen ed in he ques ionnai e (M = .63, SD = .19). Fu he mo e, mos esponden s didn’ know p e iously he news p esen ed in he s udy (M = .24, SD = .12). The piece o news which mos esponden s we e al eady awa e be o e pa icipa ing in he su ey was ecognized by 64% o esponden s, while he leas known news s o y was ecognized by only 5% o esponden s. Apa om obse ing accu acy classi ying news as ue o alse and ga he ing in o ma ion abou he p e ious knowledge o news, esponden s’ comp ehension o he pieces o news was addi ionally analyzed. In gene al, esponden s conside ed news easily unde s andable (M = 4.80, SD = .43). Pa icipan s conside ed ha pieces o news had an a e age usage use o echnical e ms (M = 2.52, SD = 0.64), e en hough his ea u e was he lowes - a ed comp ehension indica o . The in e es displayed by pa icipan s in he se e al pieces o news was ai ly high (M = 3.40, SD = .76). When classi ying agueness, pa icipan s ended o ind he news sligh ly owa ds a ague and abs ac piece o news (M= 3.03, SD = .51). Fo mo e see Table 2. Table 2 Desc ip i e S a is ics o S udy Va iables n minimum maximum M SD Accu acy 42 0.29 0.98 0.63 0.19 Knowledge 42 0.05 0.64 0.24 0.12 Comp ehension (1 - Ve y Ha d) 42 3.70 5.75 4.80 0.43 Use o Technical Te ms (1 - None) 42 1.44 4.11 2.52 0.64 In e es (1 - None) 40 1.70 4.66 3.40 0.76 Vagueness (1 - Ve y Much Conc e e) 40 1.97 4.00 3.03 0.51 36 The di e ences classi ying ue and alse news s o ies ega ding Accu acy, Knowledge, Comp ehension, Use o Technical Te ms, In e es and Vagueness we e also analyzed. Desc ip i e s a is ics a e displayed below in Table 3. S udy pa icipan s classi ied alse pieces o news mo e accu a ely han ue news. O e all, esponden s sco e a highe accu acy ca ego izing alse news (M = .72, SD = .18) han ue news (M = .54, SD = .17). When asked i hey we a e amilia wi h news p esen ed in he s udy, pa icipan s knew mo e ue news (M = .26, SD = .13) han alse news (M = .21, SD = .11). Al hough alse news we e conside ed easie o unde s and, he di e ence be ween alse (M = 4.82, SD = .48) and ue (M = 4.78, SD = .38) news s o ies comp ehension was no signi ican . Responden s conside ed ha he use o echnical e ms was g ea e in ue (M = 2.61, SD = .54) han in alse news (M = 2.43, SD = .73). Fu he mo e, esponden s showed a sligh ly g ea e in e es ed in ue (M = 3.58, SD = .67) han in alse news pieces (M = 3.23, SD = .83). Las ly, when acing news in he ques ionnai e pa icipan s conside ed alse news (M = 3.14, SD = .58) somewha ague and mo e abs ac han ue news (M = 2.92, SD = .41). Table 3 Compa ison Be ween T ue and False News T ue News False News n M SD M SD Accu acy 21 0.54 0.17 0.72 0.18 Knowledge 21 0.26 0.13 0.21 0.11 Comp ehension (1 - Ve y Ha d) 21 4.78 0.38 4.82 0.48 Use o Technical Te ms (1 - None) 21 2.61 0.54 2.43 0.73 In e es (1 - None) 20 3.58 0.67 3.23 0.83 Vagueness (1 - Ve y Much Conc e e) 20 2.92 0.41 3.14 0.58 4.1.3. NEWS CATEGORIES To s udy esponden s’ suscep ibili y o di e en opics o news, i was analyzed how esponden s pe o m classi ying news as ue o alse in each ca ego y and hei amilia i y wi h news p esen ed in he s udy as displayed on Table 4. The h ee ca ego ies wi h a highe accu acy pe cei ing news a e Science (M = .78, SD = .07), Socie y (M = .74, SD = .20) and Economy (M = .72, SD = .14). Con e sely, he ou ca ego ies in which a lowes accu acy is obse ed a e Poli ics (M = .62, SD = .15), Heal h (M = .54, SD = .21), Spo (M = .50, SD = .16) and Cul u e (M = .50, SD = .21). 37 As s a ed p e iously, news p esen ed o esponden s in he s udy we e no widely known (M = .24, SD = .12). News conce ning ca ego y Heal h (M = .33, SD = .10) we e he mos known among s udy pa icipan s, ollowed by news conce ning Economy (M = .28, SD = .18), Poli ics (M = .25, SD = .14) and Spo (M = .23, SD = .11) The h ee ca ego ies whe e pa icipan s we e mos un amilia wi h he news headlines p esen ed we e Science (M = .22, SD = .11), Socie y (M = .19, SD = .10) and Cul u e (M = .17, SD = .07). A comp ehension analysis was also de eloped wi hin he ques ionnai e. Also in Table 4 is p esen ed how pa icipan s pe cei e he news in he se en di e en ca ego ies conside ing i s o e all comp ehension, use o echnical e ms, agueness and pa icipan s’ in e es in he piece o news. The mos easily unde s ood news belong o he Cul u e (M = 5.03, SD = .43), ollowed by Spo (M = 4.93, SD = .40), Socie y (M = 4.85, SD = .54) and Science (M = 4.78, SD = .23). The ca ego ies conside ed ha de o unde s and we e Heal h (M = 4.72, SD = .36), Economy (M = 4.62, SD = .58) and Poli ics (M = 4.57, SD = .36). Responden s conside ed ha Economy (M = 2.87, SD = .43) was he opic wi h highes usage o echnical e ms, ollowed by Poli ics (M = 2.63, SD = .78), Heal h (M = 2.60, SD = 1.02) and Science (M = 2.56, SD= .51). Acco ding o s udy pa icipan s, Spo (M = 2.49, SD = .31), Cul u e (M = 2.34, SD = .42), and Socie y (M = 2.17, SD = .81) we e he opics o news whe e echnical e ms we e mos widely used. Rega ding in e es shown in each news ca ego ies, esponden s displayed highes in e es in Poli ics (M = 4.00, SD = .38), Economy (M = 3.90, SD = .46) and Heal h (M = 3.78, SD = .56). Science (M = 3.66, SD = 1.00), Socie y (M = 3.14, SD = .72), Cul u e (M = 2.90, SD = .58) and Spo (M = 2.64, SD = .39) we e he subjec s esponden s conside ed hemsel es less in e es ed in. When asked o classi y he pieces om e y much conc e e o e y much ague, pa icipan s a ed Science (M = 3.43, SD = .57), Economy (M = 3.16, SD = .64), Hea h (M = 3.10, SD = .51) and Socie y (M = 2.99, SD = .31) as he ou mo e abs ac ca ego ies. Con e sely, he ca ego ies conside ed mo e conc e e we e Cul u e (M = 2.71, SD = .27), Spo (M = 2.88, SD = .55) and Poli ics (M = 2.91, SD = .50). 38 Table 4 Desc ip i e Analysis o he Six Fea u es pe Ca ego y Ca ego y Accu acy Knowledge Comp ehension (1 – Ve y Ha d) Use o echnical e ms (1 – None) In e es (1 – None) Vagueness (1 - Ve y Much Conc e e) M SD M SD M SD M SD M SD M SD Science 0.78 0.07 0.22 0.11 4.78 0.23 2.56 0.51 3.66 1.00 3.43 0.57 Cul u e 0.50 0.21 0.17 0.07 5.03 0.43 2.34 0.42 2.90 0.58 2.71 0.27 Heal h 0.54 0.21 0.33 0.10 4.72 0.36 2.60 1.02 3.78 0.56 3.10 0.51 Spo 0.50 0.16 0.23 0.11 4.93 0.40 2.49 0.31 2.64 0.39 2.88 0.55 Economy 0.72 0.14 0.28 0.18 4.62 0.58 2.87 0.43 3.90 0.46 3.16 0.64 Socie y 0.74 0.20 0.19 0.10 4.85 0.54 2.17 0.81 3.14 0.72 2.99 0.31 Poli ics 0.62 0.15 0.25 0.14 4.57 0.36 2.63 0.78 4.00 0.38 2.91 0.50 4.1.4. DESCRIPTIVE ANALYSIS OF THE NEWS A de ailed analysis o each news headline sco e on Accu acy, Knowledge, Use o Technical Te ms, Comp ehension, In e es and Vagueness is on Appendix F. In his sec ion, we will p esen a ew pieces o news highligh ed in s udy esul s. The news headline esponden s’ classi ied mo e accu a ely was alse and belonged o ca ego y Socie y. I was “G e a Thunbe g has publicly lied abou he age” (M = 0.98). Con e sely, s udy pa icipan s sco ed lowes accu acy iden i ying e aci y o a Spo ue news s o y. I ead “Dejan Lo en, a Li e pool playe , a gues ha Co id19 is being p opaga ed o o ce he wo ld's popula ion o accina e hemsel es wi h d ugs ha a e going o implan ” (M = 0.29). The mos well-known news headline was ue and had o do wi h Economy: “In Po ugal, women's wages a e on a e age 14.5% lowe han men” (M = 0.64). O e all, pa icipan s we e less awa e o a ue news headline in ca ego y Cul u e “B uce Sp ings een's "Bo n in he USA" was he i s CD o be p oduced in he Uni ed S a es” (M = 0.05). Conside ing ease o comp ehension, he piece o news wi h a highe a e was he same pa icipan s’ classi ied mo e accu a ely. I was alse and had o do wi h Socie y: “G e a Thunbe g has publicly lied abou he age (M = 5.75). The ha des news headline o unde s and was alse and belonged o Poli ics: “Nelson 39 de Sousa, Minis e o Planning, is co ec in saying ha Po ugal is a leade in he execu ion o Eu opean unds” (M = 3.70). Bo h highes and lowes a ed news headlines when conside ing use o echnical e ms we e alse and conce ned Heal h. The news s o y wi h highes use o echnical e ms ead “Be ween 1950 and 2020, China was a he o igin o nine di e en epidemics - H2N2, H3N2, H5N1, SARS, a ian lu, swine lu, swine e e , Co ona i us, Co id-19” (M = 4.11). Con e sely, “D inking one glass o ed wine a day is equi alen o doing one hou o physical exe cise” (M = 1.44) sco ed he lowes use o echnical e ms. Pa icipan s’ in e es was highes in he mos well-known news headline. I was a ue news s o y ega ding Economy: “In Po ugal, women's wages a e on a e age 14.5% lowe han men” (M = 4.66). The news headline pa icipan s we e less in e es ed in was alse and belonged o Science. I ead “The FBI has eleased a lis o he Zodiac signs mos likely o commi c imes” (M = 1.70). Vagueness was conside ed highe in a alse news headline om Poli ics: “Nelson de Sousa, Minis e o Planning, is co ec in saying ha Po ugal is a leade in he execu ion o Eu opean unds” (M = 4.00). This piece o news conside ed he agues and mos abs ac was he same ha displayed a lowe comp ehension a e. Con e sely, “Ben ica won he 2018/19 Po uguese league wi h a o al o 103 goals. This was he i s ime a eam sco ed mo e han 100 goals du ing he championship” (M = 1.97) was he mos conc e e news headline in he s udy. I was alse and had o do wi h Spo . 4.1.5. COMPARISON BETWEEN STEM AND NON-STEM BACKGROUND When compa ing o e all su ey pe o mance o esponden s wi h STEM (n = 87) and non-STEM (n = 95) backg ounds, he di e ences we e no g ea ly signi ican as exhibi ed on Table 5. Responden s wi h STEM backg ounds (M = .63, SD = .20) classi ied mo e accu a ely he pieces o news p esen ed han esponden s wi h non-STEM backg ounds (M = .62, SD = .21). In he su ey, he p e ious knowledge o news was no signi ican ly di e en among esponden s wi h STEM (M = .24, SD = .12) and non-STEM (M = .24, SD = .16) academical and p o essional expe ience. The comp ehension o he news was sligh ly easie o pa icipan s wi h non-STEM backg ounds (M = 4.79, SD = .51) han o pa icipan s wi h a STEM backg ound (M = 4.74, SD = .45). Con a y o esponden s who possessed a STEM backg ound (M = 2.50, SD = .56), esponden s wi h a non-STEM backg ound (M = 2.59, SD = .75) belie ed ha he news had a g ea e usage o echnical e ms. The e was no a signi ican di e ence be ween he in e es o he wo g oups in he pieces o news, e en hough STEM esponden s (M = 3.43, SD = .81) had a highes o e all 40 in e es han non-STEM esponden s In e es (M = 3.37, SD = .76). News agueness' was classi ied simila ly by STEM (M = 3.08, SD = .51) and non-STEM (M = 3.02, SD = .59) esponden s. Howe e , STEM esponden s belie ed news we e ague and mo e abs ac . Table 5 Compa ison Be ween STEM and non-STEM Responden s - Desc ip i e S a is ics Backg ound STEM Non-STEM n M SD M SD Accu acy 42 0.63 0.20 0.62 0.21 Knowledge 42 0.24 0.12 0.24 0.16 Comp ehension (1 - Ve y Ha d) 42 4.74 0.45 4.79 0.51 Use o echnical e ms (1 – None) 42 2.50 0.56 2.59 0.75 In e es (1 – None) 40 3.43 0.81 3.37 0.76 Vagueness (1 – Ve y Much Conc e e) 40 3.08 0.51 3.02 0.59 In Table 6 below we analyze he accu acy o he pa icipan s wi h di e en academic and p o essional backg ounds conside ing each one o se en news ca ego ies. Al hough he h ee ca ego ies whe e news s o ies we e a ed mos accu a ely we e he same o esponden s wi h STEM and non-STEM backg ounds, he o de o op h ee was di e en . Socie y (M = .78, SD = .19), Science (M = .77, SD = .07) and Economy (M = .71, SD = .18) we e he ca ego ies ca ego ized mos accu a ely by STEM esponden s. Con e sely, non-STEM esponden s we e mo e able o iden i y accu a ely news ega ding Science (M = .81, SD = .12), Economy (M = .74, SD = .14) and Socie y (M = .71, SD = .22). The opics whe e accu acy classi ying news as ue o alse was lowe we e di e en o STEM and non- STEM esponden s. Indi iduals wi h a STEM backg ound classi ied wo se he e aci y o news s o ies ela ed wi h Cul u e (M = .49, SD = .21), while he ones wi h a non-STEM backg ound sco ed a lowe accu acy on Spo (M = .48, SD = .21). 41 Table 6 Accu acy o STEM and non-STEM Responden s pe Ca ego y o News Ca ego y STEM NON-STEM M SD M SD Science 0.77 0.07 0.81 0.12 Cul u e 0.49 0.21 0.50 0.22 Heal h 0.52 0.22 0.54 0.22 Spo 0.49 0.14 0.48 0.21 Economy 0.71 0.18 0.74 0.14 Socie y 0.78 0.19 0.71 0.22 Poli ics 0.66 0.15 0.59 0.15 In Table 7 he di e en ca ego ies o news we e sepa a ed in o ue and alse news, o s udy he beha io o STEM and non-STEM esponden s owa ds e aci y o news o a ious opics. Responden s wi h a STEM backg ound classi ied mo e accu a ely alse news ela ed wi h he ollowing ca ego ies: Socie y (M = .90, SD = .12), Economy (M = .81, SD = .07) and Science (M = .80, SD = .09), espec i ely. When conside ing ue pieces o news, he ca ego ies wi h a highes p opo ion o accu a ely classi ied news we e Science (M = .75, SD = .06), Socie y (M = .66, SD = .17), and Economy (M = .66, SD = .17). Fo mo e see Table 7. E en hough he o de o news opics iden i ied wi h highe accu acy is di e en , esponden s wi h a non- STEM backg ound classi y he same h ee ca ego ies wi h mo e accu acy as hose wi h a STEM backg ound as displayed in Table 7 below. When a non-STEM esponden encoun e ed alse news, he h ee news opics wi h a highe accu acy a e we e Socie y (M = .87, SD = .16), Science (M = .82, SD = .09) and Economy (M = .82, SD = .09). I he non-STEM indi idual aced a ue news s o y, he h ee ca ego ies wi h a highe accu acy a e we e Science (M = .79, SD = .17), Economy (M = .73, SD = .19) and Socie y ((M = .55, SD = .14). 48 Resul s o he one-way ANOVA, F(1,40) = 10.400, p = .003), e idence ha he ac ha a piece o news is ue o alse eally impac s esponden s’ accu acy iden i ying e aci y o he con en . The T- es suppo s his e idence, sugges ing ha when he news p esen ed o esponden s we e alse, hey ca ego ized i mo e co ec ly; (40) = 3.24, p < .003. The s udied popula ion o Po uguese na ionals sco ed a highe accu acy ca ego izing alse news (M = .72, SD = .18) han ue news (M = .54, SD = .17). The news s o y opic was also conside ed signi ican in accu a ely classi ying alsehood o news. On he one hand, he esul s o he MANOVA conduc ed sugges ha he opic o news has an e ec on how news a e classi ied accu a ely as ue o alse, F(6, 35) = 2.957, p = .019. Howe e , his analysis o MANOVA can be conside ed weak since Science and Cul u e a e he only wo ca ego ies in a o al o se en whe e he ela ionship is s ong. The e o e, i is no clea i his signi icance is caused by he e ec o he co a ia es. On he o he hand, a desc ip i e analysis o s udy esul s e eal ha esponden s beha ed dis inc ly classi ying alsehood o di e en ca ego ies. In he Po uguese na ionals sample unde s udy, accu acy was ela ed wi h how indi iduals pe cei e pieces o news hey encoun e . This ela ionship was di e en in ue and alse pieces o news. Acco ding o Pea son’s co ela ion, an inc ease in esponden s’ accu acy classi ying ue news s o ies indica ed an inc ease in hei p io knowledge o he news and an inc ease in pa icipan s’ in e es on he news. In alse news, an inc ease in accu acy classi ying news s o ies sugges ed a dec ease in he pe cei ed use o echnical e ms. Acco ding o Pea son’s co ela ion, highe p e ious knowledge o alse news indica ed g ea e in e es and a lowe pe cep ion o news pieces as ague and abs ac . An inc ease in comp ehension a e o alse news indica ed bo h a dec ease in pe cei ed use o echnical e ms and in in e es . Addi ionally, a highe use o echnical e ms in alse news indica ed g ea e in e es on news headlines. Mo eo e , bo h in ue and alse news a highe p e ious knowledge o news pieces led o highe in e es on i . In o ma ion ga he ed in he su ey abou indi iduals’ backg ound allowed us o s udy i academical and p o essional a eas o expe ise impac ed accu acy classi ying news as ue o alse. A desc ip i e analysis e idenced ha STEM and non-STEM esponden s beha e dis inc ly: esponden s wi h STEM backg ounds (M = .63, SD = .20) classi ied mo e accu a ely pieces o news p esen ed han esponden s wi h non-STEM backg ounds (M = .62, SD = .21), e en hough o e all di e ence was no signi ican . 49 5. CONCLUSIONS This hesis aimed o iden i y whe he indi iduals wi h di e en backg ounds e eal a di e en sensi i i y o di e en news opics. The main indings show ha no one is immune o belie e in alse news ega dless o age, le el o educa ion o a ea o expe ise. The i s esea ch ques ion aimed o s udy i he ac ha a piece o news is ue o alse in luences indi iduals iden i ying accu a ely i s e aci y. This s udy shows ha o e all indi iduals classi y mo e accu a ely alse news ha ue news. The second esea ch ques ion aimed o s udy i a ca ego y o a piece o news has an impac on how indi iduals classi y i wi h accu acy as ue o alse. I was e idenced ha people eac wi h mo e o less sensi i i y o alsehood depending on he news opic. The ca ego ies ha s udy pa icipan s classi ied mo e accu a ely we e Science, Socie y and Economy. The hi d esea ch ques ion aimed o s udy whe he an indi idual’s backg ound in luences his sensi i i y o alse news. I was possible o conclude ha pa icipan s wi h a STEM backg ound we e able o classi y news mo e accu a ely han esponden s wi h a non-STEM backg ound. While he sample limi s he gene alizabili y o esul s, his app oach p o ides new insigh s in o how Po uguese people pe cei e alse news. This hesis complemen s a c i ical li e a u e e iew wi h a su ey, which we expec will aise awa eness on how Po uguese people eac o alse news and p o ide addi ional in o ma ion o he gene al popula ion and o anyone in e es ed on he s udy o ake news. 50 6. LIMITATIONS AND RECOMMENDATIONS FOR FUTURE WORKS The indings in his s udy ha e o be seen in he ligh o some limi a ions. Fi s , he possibili y o a sample bias o which he esul s could be subjec o. The di e si y o he sample o pa icipan s ga he ed ep esen s only a ac ion o he Po uguese popula ion, mos ly om he Lisbon a ea. This esul s in p o iles o Po uguese indi iduals mos ly om Lisbon and he e o e could unde es ima e he impac o alse news in he Po uguese popula ion. Addi ionally, since he s udy had o be done using an online ques ionnai e ins ead o being applied in-pe son o a g oup o STEM and non-STEM esponden s, i can be di icul o in e he e aci y o some answe s. Tha makes he s udy expensi e, ime consuming and he e o e di icul o eplica e. The second limi a ion conce ns he size o he sample (N = 194). The la ge he sample, he mo e p ecise a e s udy esul s. A simila s udy wi h a bigge sample may ha e a g ea e accu acy. The hi d limi a ion conce ns he a en ion check (Appendix G). I was conside ed oo conse a i e and could mislead esponden s who would ead he i s pa - Whe e do you s and in e ms o poli ical o ien a ion? Poli ical o ien a ion is a sys em ha cha ac e izes di e en poli ical posi ions in ela ion o each o he - and belie ed o ha e unde s ood he ques ion, answe ing immedia ely. Thus, he second pa – So ha we know you a e paying a en ion and a e eading his in o ma ion, lea e his ques ion blank - would go unno iced and ick esponden s. The a en ion check was le blank by 113 esponden s and answe ed by 81. Due o easons men ioned ea lie and in o de o ha e a la ge numbe o esponses, he a en ion check was no u ilized o clean he da a. In u u e esea ch o his na u e, we sugges a clea e a en ion check o esponden s, and one ha is less conse a i e. 51 7. BIBLIOGRAPHY Allco , H., & Gen zkow, M. (2017). Social media and ake news in he 2016 elec ion. Jou nal o economic pe spec i es, 31(2), 211-36. Ama al, L. A. N., Scala, A., Ba helemy, M., & S anley, H. E. (2000). Classes o small-wo ld ne wo ks. P oceedings o he na ional academy o sciences, 97(21), 11149-11152. Balmau, O., Gue aoui, R., Ke ma ec, A. M., Mau e , A., Pa lo ic, M., & Zwaenepoel, W. (2018). Limi ing he Sp ead o Fake News on Social Media Pla o ms by E alua ing Use s' T us wo hiness. a Xi p ep in a Xi :1808.09922. Be encou , L. M., Cin ón-A ias, A., Kaise , D. I., & Cas illo-Chá ez, C. (2006). The powe o a good idea: Quan i a i e modeling o he sp ead o ideas om epidemiological models. Physica A: S a is ical Mechanics and i s Applica ions, 364, 513-536. Bond J , C. F., & DePaulo, B. M. (2006). Accu acy o decep ion judgmen s. Pe sonali y and social psychology Re iew, 10(3), 214-234. B ace, I. (2008). Ques ionnai e Design: How o Plan, S uc u e and W i e Su ey Ma e ial o E ec i e Ma ke Resea ch. Kogan Page Publishe s. B own, J. (2001). Using Su eys in language P og ams. Camb igde: Camb igde Uni e si y P ess. Bun ain, C., & Golbeck, J. (2017, No embe ). Au oma ically iden i ying ake news in popula wi e h eads. In 2017 IEEE In e na ional Con e ence on Sma Cloud (Sma Cloud) (pp. 208-215). IEEE. Cha, M., Haddadi, H., Bene enu o, F., & Gummadi, K. P. (2010, May). Measu ing use in luence in wi e : The million ollowe allacy. In ou h in e na ional AAAI con e ence on weblogs and social media. Chi omo, A. (2006). Resea ch Me hods and Sa is ics in Educa ion: A s uden 's Guide. Gwe u: Be a P in . Chisholm, B. R., Mulle , P. A., Ho n, A. J., & Ellis, Z. S. (2019). A compe ing in ec ion model o he sp ead o di e en iewpoin s o a di isi e idea. The Jou nal o Ma hema ical Sociology, 43(3), 147-163. Con oy, N. J., Rubin, V. L., & Chen, Y. (2015). Au oma ic decep ion de ec ion: Me hods o inding ake news. P oceedings o he Associa ion o In o ma ion Science and Technology, 52(1), 1-4. Del Vica io, M., Bessi, A., Zollo, F., Pe oni, F., Scala, A., Calda elli, G., ... & Qua ociocchi, W. (2016). The sp eading o misin o ma ion online. P oceedings o he Na ional Academy o Sciences, 113(3), 554-559. 52 Del Vica io, M., Vi aldo, G., Bessi, A., Zollo, F., Scala, A., Calda elli, G., & Qua ociocchi, W. (2016). Echo chambe s: Emo ional con agion and g oup pola iza ion on acebook. Scien i ic epo s, 6, 37825. Douglas, K. M., Su on, R. M., & Cichocka, A. (2017). The psychology o conspi acy heo ies. Cu en di ec ions in psychological science, 26(6), 538-542. Eps ein, Z., Pennycook, G., & Rand, D. (2019). Le ing he c owd s ee he algo i hm: Laypeople can e ec i ely iden i y misin o ma ion sou ces. Gen zkow, M., & Yu, N. C. (2018). T ends in he Di usion o Misin o ma ion on Social Media. Goe zel, T. (1994). Belie in conspi acy heo ies. Poli ical Psychology, 731-742. G eene, C. M., Nash, R. A., & Mu phy, G. (2021). Mis emembe ing B exi : Pa isan bias and indi idual p edic o s o alse memo ies o ake news s o ies among B exi o e s. Memo y, 1-18. G inbe g, N., Joseph, K., F iedland, L., Swi e-Thompson, B., & Laze , D. (2019). Fake news on Twi e du ing he 2016 US p esiden ial elec ion. Science, 363(6425), 374-378. Hamelee s, M., B osius, A., & de V eese, C. H. (2021). Whe e’s he ake news a ? Eu opean news consume s’ pe cep ions o misin o ma ion ac oss in o ma ion sou ces and opics. Ha a d Kennedy School (HKS) Misin o ma ion Re iew. h ps://doi.o g/10.37016/m -2020-70 Hoch ei e , R., & Waldhause , C. (2014). Zombie poli ics: E olu iona y algo i hms o coun e ac he sp ead o nega i e opinions. a Xi p ep in a Xi :1401.6420. Jin, F., Wang, W., Zhao, L., Doughe y, E. R., Cao, Y., Lu, C. T., & Ramak ishnan, N. (2014). Misin o ma ion p opaga ion in he age o wi e . IEEE Compu e , 47(12), 90-94. Kucha ski, A. (2016). S udy epidemiology o ake news. Na u e, 540(7634), 525-525. Laze , D. M., Baum, M. A., Benkle , Y., Be insky, A. J., G eenhill, K. M., Mencze , F., ... & Schudson, M. (2018). The science o ake news. Science, 359(6380), 1094-1096. Leman, P. J., & Cinni ella, M. (2013). Belie s in conspi acy heo ies and he need o cogni i e closu e. F on ie s in psychology, 4, 378. Mae ens, R., Roozenbeek, J., Basol, M., & an de Linden, S. (2020). Long- e m e ec i eness o inocula ion agains misin o ma ion: Th ee longi udinal expe imen s. Jou nal o Expe imen al Psychology: Applied. Magwa, S. & Magwa, W. (2015). A guide o conduc ing esea ch: a s uden handbook. S a egic Book Publishing Righ s Agency. 53 Moussaïd, M., Kämme , J. E., Analy is, P. P., & Ne h, H. (2013). Social in luence and he collec i e dynamics o opinion o ma ion. PloS one, 8(11), e78433. Nielsen, R. K., Fle che , R., Newman, N., B ennen, J. S., & Howa d, P. N. (2020). Na iga ing he ‘in odemic’: How people in six coun ies access and a e news and in o ma ion abou co ona i us. Reu e s Ins i u e. Pennycook, G., Eps ein, Z., Mosleh, M., A echa , A. A., Eckles, D., & Rand, D. (2019). Unde s anding and educing he sp ead o misin o ma ion online. Pennycook, G., & Rand, D. G. (2021). The psychology o ake news. T ends in cogni i e sciences. Pe in, A. (2015). Social Media Usage: 2005-2015: 65% o Adul s Now Use Social Ne wo king Si es--A Nea ly Ten old Jump in he Pas Decade. Pew Resea ch T us . Pose i, J., & Ma hews, A. (2018). A sho guide o he his o y o ’ ake news’ and disin o ma ion. In e na ional Cen e o Jou nalis s, 7, 1-19. Rapp, D. N., & Salo ich, N. A. (2018). Can’ we jus dis ega d ake news? The consequences o exposu e o inaccu a e in o ma ion. Policy Insigh s om he Beha io al and B ain Sciences, 5(2), 232-239. Saad, M., Ahmad, A., & Mohaisen, A. (2019, June). Figh ing ake news p opaga ion wi h blockchains. In 2019 IEEE Con e ence on Communica ions and Ne wo k Secu i y (CNS) (pp. 1-4). IEEE. Shao, C., Ciampaglia, G. L., Va ol, O., Flammini, A., & Mencze , F. (2017). The sp ead o ake news by social bo s. a Xi p ep in a Xi :1707.07592, 96, 104. Shu, K., Sli a, A., Wang, S., Tang, J., & Liu, H. (2017). Fake news de ec ion on social media: A da a mining pe spec i e. ACM SIGKDD Explo a ions Newsle e , 19(1), 22-36. Swi e, B., Ecke , U. K., & Lewandowsky, S. (2017). The ole o amilia i y in co ec ing inaccu a e in o ma ion. Jou nal o expe imen al psychology: lea ning, memo y, and cogni ion, 43(12), 1948. Tö nbe g, P. (2018). Echo chambe s and i al misin o ma ion: Modeling ake news as complex con agion. PLoS one, 13(9), e0203958. Tu co e, J., Yo k, C., I ing, J., Scholl, R. M., & Ping ee, R. J. (2015). News ecommenda ions om social media opinion leade s: E ec s on media us and in o ma ion seeking. Jou nal o Compu e - Media ed Communica ion, 20(5), 520-535. Vaidyana han, G. (2020). News Fea u e: Finding a accine o misin o ma ion. P oceedings o he Na ional Academy o Sciences, 117(32), 18902-18905 Vo, N., & Lee, K. (2020). Whe e a e he ac s? sea ching o ac -checked in o ma ion o alle ia e he sp ead o ake news. a Xi p ep in a Xi :2010.03159. 54 Vosoughi, S., Roy, D., & A al, S. (2018). The sp ead o ue and alse news online. Science, 359(6380), 1146-1151. Wagne , C., Mi e , S., Kö ne , C., & S ohmaie , M. (2012, Ap il). When Social Bo s A ack: Modeling Suscep ibili y o Use s in Online Social Ne wo ks. In # MSM (pp. 41-48). Wang, W. Y. (2017). "lia , lia pan s on i e": A new benchma k da ase o ake news de ec ion. a Xi p ep in a Xi :1705.00648. Whi e, D. W. (2014). Wha is STEM educa ion and why is i impo an . Flo ida Associa ion o Teache Educa o s Jou nal, 1(14), 1-9. ou B own, R., B own, J., Rea don, K., & Me ill, C. (2011). Unde s anding STEM: cu en pe cep ions. Technology and Enginee ing Teache , 70(6), 5. Zhang, X., & Gho bani, A. A. (2019). An o e iew o online ake news: Cha ac e iza ion, de ec ion, and discussion. In o ma ion P ocessing & Managemen . Zhao, J., Cheng, J., & Gao, H. (2014, July). Public Opinion P opaga ion Model on Social Ne wo ks. In 2014 Se en h In e na ional Join Con e ence on Compu a ional Sciences and Op imiza ion (pp. 325- 328). IEEE. 55 8. APPENDIX Appendix A Su ey Welcome Message Appendix B Responden s’ P o ession n % S uden 65 33.5 Enginee 12 6.2 Teache 12 6.2 Re i ed 6 3.1 Manage 5 2.6 Consul an 4 2.1 None 4 2.1 Accoun Manage 3 1.5 56 A chi ec 3 1.5 Ci il Enginee 3 1.5 Ele ical Enginee 3 1.5 Jou nalis 3 1.5 Ju is 3 1.5 Ma ke ing 3 1.5 P ojec Manage 3 1.5 Unemployed 3 1.5 Adminis a i e Technician 2 1.0 Diploma 2 1.0 Educa ion Assis an 2 1.0 Finance In e n 2 1.0 Insu ance Agen 2 1.0 IT Technician 2 1.0 Lawye 2 1.0 Resea che 2 1.0 Tou ism Technician 2 1.0 Ac ess 1 0.5 Analys 1 0.5 Banke 1 0.5 Ba man 1 0.5 Biomedical Enginee 1 0.5 Chemical Enginee 1 0.5 Communica ions and Ma ke ing 1 0.5 Communi y De elopmen Technician 1 0.5 Da a Scien is 1 0.5 Designe 1 0.5 Doc o 1 0.5 E-Ma i ime Sys ems Enginee 1 0.5 En ep eneu 1 0.5 Financial Analys 1 0.5 Foo ball Coach 1 0.5 Geological Enginee 1 0.5 Highe Technician 1 0.5 Ho elie 1 0.5 Human Resou ces Manage 1 0.5 IT Analys 1 0.5 Ki chen Assis an 1 0.5 Legal Consul an 1 0.5 Managemen Con olle 1 0.5 Managing pa ne 1 0.5 57 Ma hema ician 1 0.5 Mo he 1 0.5 Os eopa h 1 0.5 P o esso 1 0.5 Psychologis 1 0.5 Publicis 1 0.5 Religionis 1 0.5 RPA Consul an 1 0.5 Sales Di ec o 1 0.5 So wa e Enginee 1 0.5 SQA Enginee 1 0.5 Sys em Adminis a o 1 0.5 Technical Di ec o 1 0.5 Telecomunica ions Enginee 1 0.5 Tou is Guide 1 0.5 T ansla o 1 0.5 T easu e 1 0.5 Appendix C S uden Responden s' A ea o Specializa ion n % Business 14 21.5 Heal h 10 15.4 Enginee ing 9 13.8 Law 5 7.7 Ma ke ing 4 6.2 None 4 6.2 A chi ec u e 3 4.6 A s 2 3.1 Communica ion 2 3.1 Da a Science 2 3.1 In o ma ion Managemen 2 3.1 Philosophy and Teology 2 3.1 Psicology 2 3.1 Biology and Geology 1 1.5 In e na ional Rela ions 1 1.5 Psychomo o Rehabili a ion 1 1.5 Social Se ices 1 1.5 64