Images of Science Linked to Labwork: A Survey of Secondary School and University Students

Resea ch in Science Educa ion 31: 499–523, 2001. © 2001 Kluwe Academic Publishe s. P in ed in he Ne he lands. Images o Science Linked o Labwo k: A Su ey o Seconda y School and Uni e si y S uden s Ma ie-Gene iè e Sé é1, Manuel Fe nandez-Gonzalez2, Jose A. Gallegos2, F ancisco Gonzalez-Ga cia2, Es eban De Manuel2, F. Ja ie Pe ales2 and John Leach3 1Uni e si é Pa is XI (F ance) 2Uni e sidad de G anada (Spain) 3The Uni e si y o Leeds (U.K.) Abs ac This pape p esen s indings abou he images o science d awn upon in labo a o y wo k, by uppe seconda y and uni e si y s uden s, in academic s eams wi h a science ocus. Da a we e collec ed h ough ou w i en ques ions, adminis e ed o a o al o 368 s uden s. The ques ions all equi ed s uden s o commen on labo a o y in es iga ions ca ied ou by esea ch scien is s o by science s uden s. We show ha s uden s’ easoning has an epis emological and an on ological dimension, and ha i o en di e s signi ican ly om accep ed pe spec i es on he na u e o science. The issue o eaching appea s o be showing s uden s wha coun s (and wha does no coun ) as app op ia e easoning in ac ual si ua ions. In o he wo ds, explici eaching abou he a ious ela ionships ha can exis be ween heo y and da a would ans o m labwo k owa ds a mo e c i ical p ocess ha in ol es making and jus i ying decisions. Du ing hei schooling, s uden s acqui e images o science and o scien i ic wo k. In ecen yea s he e ha e been many s udies ocused on cha ac e ising hese images. Se e al s udies ha e shed ligh on he images o science ac ually d awn upon by s u- den s, no ma e how hose images o science we e de eloped. D i e , Leach, Milla and Sco (1996), and Désau els and La ochelle (1998), ha e made comp ehensi e e iews o his li e a u e. The indings o s udies ypically po ay s uden s as naï e ealis s and naï e empi icis s. Fu he mo e, he e is some e idence o sugges ha many eache s also make naï e s a emen s abou he na u e o science (e.g., Led- e man, 1992), and ha eache s’ images o science a e communica ed o s uden s du ing eaching (e.g., B ickhouse, 1990; Hodson, 1993). Howe e , he p ocess o d awing ou implica ions o p ac ice om his li e a- u e is no s aigh o wa d. The e is some e idence o sugges ha s uden s (e.g., Rowell & Dawson, 1983) and p o essional scien is s (e.g., Sama apunga an, 1992) who make naï e e bal s a emen s abou science, demons a e mo e sophis ica ed aci knowledge abou science in ac ion se ings. Fu he mo e, s uden s’ easoning abou science is in ima ely connec ed o con ex (e.g., Mo ime , 1995; Leach, Milla , Ryde , & Sé é, 2000). Qui e o en, he diagnos ic ques ions used in pape and pencil su eys o elici s uden s’ and eache s’ images o science do no e e o any speci ic 500 M.-G. S ´ ER´ EETAL. con ex . We e e o such ques ions as decon ex ualised ques ions. Fo example, Po - lan A iza, Ri e o Ga cia and Ma in del Pozo (1998) used decon ex ualised ques ions o discuss possible ela ionships be ween eache s’ epis emological op ions on one hand, and hei eaching op ions on he o he . Koulaidis and Ogbo n (1989) p esen ed eache s wi h a se ies o gene al s a emen s, exp essing di e en philosophical iews. In bo h s udies, a p ede ined epis emological posi ion is a ibu ed o each indi- idual s uden o eache . This app oach elici s he s uden ’s o eache ’s espoused images o science. Howe e , he ela ionship be ween hese espoused posi ions, and he knowledge d awn upon by s uden s and eache s o in o m ac ions, is open o ques ion. In his esea ch, we we e in e es ed in elici ing he kind o knowledge ha science s uden s migh d aw upon o in o m hei ac ions in he se ing o a pa icula class- oom ac i i y, namely labwo k. Wi hin he Eu opean p ojec ‘Labwo k in Science Educa ion’1(LSE) (Sé é, 1998), we de eloped some hypo heses conce ning he ela ionship be ween science s uden s’ images o science and hei ac ions du ing labwo k a he le el o uppe seconda y school, and he beginning o uni e si y educa ion (Leach, in p ess). In o de o in es iga e hese hypo heses, we designed ques ions which equi e s uden s o make judgemen s and commen s on speci ic labo a o y si ua ions. The design o hese ques ions ecognises ha he images o science ha s uden s de elop and use a e in ima ely connec ed o he scien i ic ac- i i ies engaged in h ough labo a o y wo k: di e en labo a o y ac i i ies in ol e s uden s in making di e en links be ween concep ual knowledge and images o scien i ic ac i i y. The knowledge a s ake du ing labwo k is no only concep ual: knowledge abou aspec s o he si ua ion such as he ela ionship be ween ideas and da a, he p ocess o measu emen and so on is also in ol ed. We we e in e es ed in he knowledge s uden s use in making decisions abou how much da a o collec , which o he da a o use, and wha can be concluded om he da a a ailable. In he physical sciences, i is o en necessa y o s uden s o make decisions abou he alidi y o knowledge claims om expe imen al wo k, gi en he accu acy and p ecision o mea- su emen s. In he li e sciences, s uden s o en ha e o make decisions abou design and sampling in o de o collec use ul da a. Da a analysis ypically in ol es dealing wi h la ge da a se s exhibi ing a high deg ee o a iabili y. We we e in e es ed in he ep esen a ions o he unc ioning o science ha unde pinned s uden s’ decisions du ing labwo k. In he philosophy o science, ques ions abou he na u e and s a us o scien i ic knowledge a e con en ionally discussed a ound wo dimensions: 1. an on ological dimension, which add esses he ela ionship be ween scien i ic models and hei empi ical e e en s; and 2. an epis emological dimension, which add esses he wa an ing o knowledge claims as eliable. These dimensions ha e been used in he li e a u e o desc ibe concep ual lea ning (e.g., Chi, 1992; Tyson, Ven ille, Ha ison, & T eagus , 1997; Vosniadou, 1994). In he s udy epo ed in his pape , howe e , we do no add ess concep ual lea ning. Ra he , we d aw upon on ological and epis emological dimensions o desc ibe he IMAGES OF SCIENCE LINKED TO LABWORK 501 knowledge ha s uden s appea o d aw upon when dealing wi h decisions encoun- e ed du ing lea ning h ough labwo k. Ou ini ial assump ion is ha s uden s’ easoning du ing labo a o y wo k has on- ological and epis emological dimensions, because du ing labwo k s uden s ha e o d aw upon implici knowledge abou he na u e o scien i ic esea ch, he analysis o empi ical da a and he alidi y and eliabili y o knowledge claims. The pu pose o his pape is o demons a e, h ough he analysis o a ques ion- nai e, he epis emological and on ological dimensions ha cha ac e ises uppe sec- onda y and beginning uni e si y s uden s’ easoning abou labwo k. We show ha indi idual s uden s’ easoning exhibi s di e en cha ac e is ics in esponse o di e - en ques ions. We go on o conside he implica ions o ou indings abou s uden s’ easoning o he p ac ice o eaching science h ough labwo k. The Su ey and he Samples In o de o cha ac e ise on ological and epis emological aspec s o s uden s’ ea- soning abou labwo k, i was necessa y o design diagnos ic ques ions which p o- ided s uden s wi h he oppo uni y o discuss he p ac ice o labwo k. Following pilo ing, ou ques ions (con aining bo h open and closed- esponse sec ions) we e adminis e ed o a sample o s uden s in uppe seconda y schools, and a he beginning o hei science s udies a uni e si y. The su ey was adminis e ed in F ance and Spain, no ably due o he linguis ic compe ence o he esea ch eam. I was adminis e ed in ei he F ench o Spanish e- spec i ely. The e we e some di e ences be ween he o ms o ques ions in he F ench and Spanish ques ionnai es; hese will be commen ed upon as ele an o he epo - ing o indings. In Spain, he ques ionnai e was gi en o 89 seconda y school s uden s and o 37 uni e si y s uden s (mainly in Physics o Chemis y, less in Biology). In F ance i was gi en o 76 seconda y school s uden s and 166 uni e si y s uden s, 60 o whom s udy Biology o Geology and 106, Physics o Chemis y. This gi es a o al sample o 368 s uden s (40% emale and 60% male). This sample in ol ed andomly selec ed s uden s in academic s eams, in se e al owns, whose s udies ocused on science. We do no claim ha he sample is na ionally ep esen a i e, hough we do no see any speci ic easons why he sample di e s om he na ional posi ion in ei he coun y. Fu he mo e, gi en he simila i y o ou indings o s udies using simila me hods and in ol ing a wide ange o coun ies (including an anglophone coun y) (Leach, Milla , Ryde , & Sé é, 2000), we suspec ha indings om his s udy migh well be o in e es in coun ies o he han F ance and Spain. A eas o Focus wi hin he Ques ions In his sec ion, we explain he ocus o each diagnos ic ques ion. The ull ex o each ques ion, ansla ed in o English, can be ound in he Appendix. 502 M.-G. S ´ ER´ EETAL. The Theo y Ques ion This is a decon ex ualised ques ion, asking whe he o no i is necessa y o ha e a heo y in mind in o de o in e p e da a. This ells us i s uden s belie e ha he possession o su icien da a allows people o a i e a a heo y, o al e na i ely, ha ha ing a heo y in mind is necessa y o o ien he collec ion o da a. To balance he absence o con ex , i asks s uden s o desc ibe a si ua ion ha exempli ies hei answe . This ques ion is ele an o he na u e o empi ical enqui y. The Cu es Ques ion This ques ion in ol es wo scien is s who in e p e he same da a (a se o poin s) in adically di e en ways. One o he scien is s ep esen s he ela ionship be ween da a poin s by means o a s aigh line, while he o he in e p e s he ela ionship as a cu e. The open- esponse ques ions ask s uden s whe he i is legi ima e o ha e wo in e p e a ions, and how scien is s migh jus i y hei espec i e iewpoin s. A sec- onda y objec i e o his ques ion is o disco e he ex en o which s uden s belie e ha compu e s a e capable o sol ing such p oblems. Since his ques ion in ol es modelling, i is in o ma i e abou s uden s’ iews o he na u e o empi ical enqui y as well as he analysis o da a. The Su p ise Ques ion This ques ion desc ibes a con ex in which he esul s o an expe imen in a class- oom si ua ion do no ma ch canonical knowledge. The eache , who is awa e o he heo e ical concep s o be augh , p oposes a labo a o y expe imen ha is expec ed o e i y hem in some way. The example chosen is a es o s a ch in plan lea es, which should come ou posi i e o lea es le in he ligh and nega i e o lea es kep in he da k. Howe e , when he expe imen is ca ied ou , some o he s uden s do no ob ain he desi ed esul . In he F ench ques ionnai e, his ques ion is in an open- esponse o ma and simply asks wha should be done i some hing like his happens in class. In he Spanish ques ionnai e, he ques ion is o mula ed as mul iple-choice. The ques ion elici s in o ma ion abou he links ha s uden s make be ween heo y and expe imen al design. I also conce ns he analysis o expe imen al da a. The Se ies Ques ion To disco e he s uden s’ iews abou handling measu ed da a, we asked ques ions abou se s o sp ead measu emen s. P e ious s udies (Allie, Bu le , Kaunda, Camp- bell, & Lubben, 1998; Jou neaux & Sé é, 1994; Lubben & Milla , 1996) sugges IMAGES OF SCIENCE LINKED TO LABWORK 503 ha he way ha s uden s ea epea measu emen s can be used o cha ac e ise hei implici assump ions abou he na u e o da a. Di e en se s o i e measu emen s (o mass) we e p esen ed, oge he wi h he mean o each se . The se s had he ollowing cha ac e is ics, selec ed on he basis o indings epo ed in he li e a u e ci ed abo e: 1. The mean may o may no appea as one o he alues o he se ies; 2. Ce ain alues appea up o h ee imes in he same se ies; 3. The ange be ween he smalles and la ges alue a ies in size. The ques ion conce ns wo g oups o nu i ionis s who mus measu e he mass o one decili e o oil i e imes. S uden s we e asked abou di e en si ua ions: 1. he same kind o oil is measu ed by each g oup, and each g oup’s se s o mea- su emen s ha e he same mean; 2. he g oups ha e he same oil bu hei se s o measu emen s ha e di e en means; 3. he g oups ha e di e en oils and hei se s o measu emen s ha e di e en means. The ex o he ques ion sugges s ha a di ec measu emen o mass was made. In ac , he alues ob ained depend on he olume, one decili e, o he sample aken. S ic ly speaking, his signi ies ha he measu emen is no di ec , and he design o he expe imen hus in luences he esul . This ques ion p o ides in o ma ion p incipally abou s uden s’ iews o he analy- sis o expe imen al da a, bu also abou hei iews o he na u e o empi ical enqui y. Ca ego ies o Analysis The ca ego ies o analysis2o he esponses we e elabo a ed by d awing upon ou hypo heses, ou di ec obse a ions o s uden s in ol ed in labwo k, and ou knowledge o he exis ing li e a u e on s uden s’ images o science. A: Ca ego ies Re lec ing On ological and Epis emological Posi ions i) Each Measu emen should be conside ed in i sel This idea ha he only alues o a quan i y ha a e conside ed alid a e hose ac ually ob ained as measu ed alues, has been ca ego ised as Values. Simila ly, in ce ain cases when s uden s ha e o gi e a unique esul om wo se ies o measu e- men s, hey keep he alue ha appea s in bo h se ies. The same ype o easoning is ound o se ies o measu emen s and o pai s o alues (poin s o be in e p e ed as a cu e). I places undue p ominence on he ac ual measu ed alue, gi ing no conside a ion o issues o p obabili y. This a i ude, i sys ema ic, con as s wi h he iew ha conside s all measu es as bea ing in o ma ion o which s a is ical analysis can be applied. I has an in luence on he choice o a esul . 504 M.-G. S ´ ER´ EETAL. ii) Accu acy o unce ain y should be conside ed o each measu emen , in o de o judge each o hem This idea has been ca ego ised ei he as Unce ain y,o asDe ia ion: Unce ain y when quan i a i e alues a e in play; o De ia ion which co esponds o he same epis emological-on ological posi ion, bu gene ally exp essed o non-quan i a i e esul s. The la e means, o ins ance, ha he e is always a di e ence be ween wha is eally ob ained om obse a ion, and he e en o alue deduced om he heo y. Typically, he decision ollowing his a i ude, is equen ly o selec which mea- su emen ough o be kep and which should be disca ded. This a i ude con as s wi h a s a is ical iew o he ea men o di e en measu ed alues. iii) The di e ence be ween he bigges and he smalles measu emen s de e mines he quali y o he da a This ca ego y is called Range. i ) I is necessa y o ake se e al measu emen s (o a leas mo e han he numbe gi en in he ques ion) in o de o a i e a alid esul s This idea has been ca ego ised as NMeasu emen s (in case o quan i a i e e- sul s) and S a Again (in he case o quali a i e esul s: i is necessa y o epea an expe imen because one single expe imen can ne e be enough). ) S a is ical analysis is necessa y in he p ocessing o da a This a i ude is ca ego ised as S a is ics. I is he exac opposi e o Values. i) The e exis s one ue alue, which is knowable and mus be known be o e judgemen s can be made This idea is ca ego ised ei he as T u h (belie in he exis ence o an ideal o heo e ical alue) o O icial Value ( he idea o an o icially ecognised alue is used, wi h no e e ence o how ha alue migh be wa an ed). All hese ca ego ies ha e on ological/epis emological dimensions, as illus a ed in Figu e 1. The axes in Figu e 1 ep esen an on ological dimension (iden i y o he eal wo ld wi h i s scien i ic modelling/dis ance o he eal wo ld om i s scien i ic modelling) and an epis emological dimension (p imacy o da a/p imacy o heo y). On he ex emi ies o he axis p imacy o heo y/p imacy o da a, wo ca ego ies o he decon ex ualised closed ques ion appea : heo y i s /da a i s . IMAGES OF SCIENCE LINKED TO LABWORK 505 Figu e 1: A wo-dimensional ep esen a ion o he ca ego ies o analysis o s uden s’ answe s o ques ions conce ning labwo k. B: Ca ego ies Compa ible wi h Any On ological/Epis emological Posi ions i) Exis ence/ine i abili y o human e o This is classi ied in he ca ego y E o . The e ms used a e ha measu emen s a e impe ec , possibly e en ‘ne e pe ec ,’ o ‘always wi h a ma gin o e o .’ ii) The esul s ob ained a e explained in e ms o he expe imen al p o ocol used This is ca ego ised as P o ocol. C: Ca ego y o Rela i ism Some answe s exp ess Rela i ism. All esponses s a ing ha no hing can be de- cided, ha subjec i i y is in play, o ha any conclusion/decision could be equally alid we e ea ed as ela i is . Examples o such esponses we e ound on each ques ion in he su ey. 506 M.-G. S ´ ER´ EETAL. Resul s We now p esen he esul s o he su ey, using he ca ego ies desc ibed in he las sec ion. This means ha he majo i y o esponses o each ques ion can be si ua ed in a gi en quad an o Figu e 1. As will be seen, he answe s o he i s ques ion we analyse can be in e p e ed only in e ms o unde lying epis emology. They a e si ua ed on he epis emological axis o Figu e 1. The Theo y Ques ion (Focus: he Na u e o Resea ch) The i s pa o he ques ion is no con ex ualised. As shown in he appendix, he s uden s a e asked o ag ee o disag ee wi h he ‘ heo y i s ’ s a emen o he ‘da a i s ’ s a emen . They a e also asked o gi e he example hey ha e in mind when making hei choice. In o he wo ds, s uden s a e asked di ec ly o si ua e hemsel es wi hin he plane ep esen ed in Figu e 1, mo e p ecisely on he ho izon al axis. The ques ion o e s s uden s he choice be ween wo answe s ha a i s seem con adic o y. This is due o ou assump ion ha scien is s ca y ou hei wo k in di e en ways, and he di e se app oaches used canno adequa ely be desc ibed by ei he o he wo s a emen s. Quan i a i e esul s o he whole sample a e shown in Table 1: Table 1 F equencies o Choices (%) Be ween Gene al Sen ences Desc ibing he De elopmen o Science. Ag ee wi h Disag ee wi h No su e (%) s a emen (%) s a emen (%) Uppe Uni e si y Uppe Uni e si y Uppe Uni e si y seconda y seconda y seconda y Theo y i s 49 39 30 19 20 19 s a emen Da a i s 22 11 59 72 12 13 s a emen The mos equen esponse is Ag ee wi h he heo y i s s a emen and Disag ee wi h he da a i s s a emen . This mos equen esponse co esponds o he igh - hand side o Figu e 1. A signi ican numbe o s uden s, pa icula ly a he uni e si y le el, made no esponse o he ques ion. IMAGES OF SCIENCE LINKED TO LABWORK 507 By and la ge, he examples o e ed by s uden s pe ain o physics (almos hal o he examples) o biology (app oxima ely a hi d o he examples). They gene ally conce n his o ical disco e ies o a e de i ed om class oom si ua ions. The ollow- ing examples we e commonly used o jus i y he Theo y i s posi ion: Biologis s who s udy gene ics ha e been guided by he heo y o Mendel and he disco e y o DNA; and, a scien is should possess in ui ion, Plu o ( he plane ) was disco e ed heo e ically. Scien is s such as New on, Huygens, Eins ein, Da win, Cha pak, and Michelson a e also ci ed as examples o jus i y his iew. An example ci ed o jus i y he Da a i s s a emen was: “I scien is s only s a ed om heo e ical knowledge, hey would s ill hink ha he wo ld was la .” S uden s ci ed Eins ein, Le Ve ie , Desca es and Be gson o jus i y s a emen 2. Abou 15% o he s uden s ag eed wi h bo h s a emen s. The mos equen expla- na ion gi en is ha bo h a e possible because one s a emen is no su icien in i sel o co e all scien i ic p ac ice, o example: “The disco e y o he bacillus o Koch, as well as he gene ic enginee ing, which sea ches o one gene o ano he , show ha bo h me hods (i.e. Theo y i s and Da a i s ) a e complemen a y.” Boh , Planck, Eins ein, Koch, Galileo, Ma ie Cu ie, and La oisie a e ci ed as ex- amples ha p o e ha bo h s a emen s a e possible, while Planck and La oisie we e also gi en as e idence o he con a y. This sugges s ha a signi ican p opo ion o he esponses we e loca ed on bo h sides o he ho izon al axis o Figu e 1, depending on he si ua ion. In summa y, s uden s’ esponses we e sp ead along he ho izon al axis o Figu e 1, wi h a p edominance o he igh -hand side. F om a me hodological poin o iew, hese esul s mean ha one mus be ca e ul in he in e p e a ion o decon ex ualised ques ions. The answe s a e hea ily depen- den on p e iously lea ned and memo ised examples, and i is impossible o sepa a e he esul s om hei con ex s. The Cu es Ques ion (Focus: The Na u e o Resea ch and he P ocessing o Da a) This ques ion asks s uden s wha scien is s should do o esol e he dilemma when wo di e en in e p e a ions o he same da a a e a ailable. The ques ion is no con ex ualised o a speci ic discipline o si ua ion, and simply o e s a se o da a poin s. App oxima ely 16% o he s uden s (and as many as 31% o he Spanish seconda y school s uden s) s a ed ha he di e ence is basically due o subjec i i y on he pa o scien is s. Those ela i is answe s canno be posi ioned on Figu e 1. The esul s o he emaining esponses a e p esen ed in Table 2: 514 M.-G. S ´ ER´ EETAL. Table 8 Pe cen age o Responses Ag eeing wi h Va ious Conclusions. F ance (means Spain (means di e ing by 0.2 g) di e ing by 1.2 g) Uppe Uni e si y (%) Uppe Uni e si y (%) seconda y seconda y school (%) school (%) I is possible o a i m ha he measu ed oils di e in densi y. 17 18 55 32 I is possible o a i m ha he measu ed oils do no di e in densi y. 49 48 18 19 i is impossible o ell.” 2. Unce ain y (8%–22% acco ding o he samples). 3. As p e iously, he ca ego y P o ocol is a e (0%–8% acco ding o he samples). In summa y, hese esul s show ha : answe s in e p e able in e ms o on ology and epis emology a e sp ead all o e he egions o Figu e 1, wi h a p edominance in he le -hand side; he impo ance o he Values ca ego y, close o he endency o iden i y eali y and model, depends on he con ex ; and, hough di icul o coun p ecisely, endencies o Rela i ism a e exp essed occasionally in jus i ica ions. S uden s a e no quick o c i icise he expe imen al p ocedu e i sel . They p obably a e no ained o use his app oach. Syn hesis o Resul s In he p e ious sec ion, we showed ha s uden s’ esponses o he ques ionnai e could be cha ac e ised in e ms o epis emological and on ological aspec s o hei easoning. We used he ou quad an s o Figu e 1 o ep esen hese epis emo- logical and on ological aspec s. S uden s did no d aw upon a single epis emolog- ical/on ological aspec in esponding o he ques ionnai e. The mos equen ly gi en answe s o di e en ques ions we e o en si ua ed in di e en quad an s. Gene ally speaking, mo e s uden s’ answe s we e loca ed in he quad an s on he igh o he plane, and he quad an s a he op. This means ha he endency p imacy o heo y (on he epis emological axis) and dis ance heo y/model (on he on ological axis) IMAGES OF SCIENCE LINKED TO LABWORK 515 we e mo e commonly chosen by s uden s in connec ion wi h labwo k. Howe e , some esponses we e o be ound in he quad an cha ac e ised by iden i y eali y- model and p imacy o da a ( hough he quan i y was gene ally less han 20%). This leads o posing he ollowing ques ions: I is possible o si ua e answe s in he quad- an s o Figu e 1 – is i also possible o si ua e indi iduals? In o he wo ds, does each indi idual, a he ime he/she answe s he ques ionnai e, d aw upon a gi en posi ion (on ology, epis emology, a endency o ela i ism, o ano he philosophical posi ion)? Does a esponden gi e answe s which all de i e om he same posi ion? The me hod used o add ess hese ques ions was o ca y ou a c oss-ques ion s udy5 o he consis ency and inconsis ency o esponses o di e en pa s o ques ions, and ac oss ques ions. In ac , examples o inconsis ency a e nume ous. The mos clea -cu ones a e as ollows: 1. None o he s uden s answe S a is ics consis en ly in he Cu es ques ion and he h ee pa s o he Se ies ques ion, al hough 49% o e all answe his way on one occasion om he ou possible occasions. 2. Among six oppo uni ies o answe T u h, e y ew s uden s used his kind o esponse on h ee o mo e occasions. 3. No co ela ion can be ound be ween he ca ego ies Unce ain y,De ia ion and Range (each ound on he uppe le quad an o he plane). 4. I is possible o use he Values ca ego y on i e occasions in he ques ionnai e. 36% o he sample used i once, 12% used i wice, e y ew used i h ee imes. We ied o iden i y i some shi om one ca ego y o ano he could be consid- e ed as equen o e en sys ema ic. An example is p o ided by he Se ies ques ion. Among he s uden s who s a wi h Range (“same oil, same mean”), 55% also e- sponded in his way on he “same oil-di e en means” ques ion. Those who changed (45%) ended o selec Values. The o he examples o equen ypes o changes ha we no iced led us o he same conclusion: The mo e complex he si ua ion is, he mo e equen ly Values is chosen. Howe e , ce ain consis encies exis in esponses ac oss ques ions. This can be seen in he h ee pa s o he Se ies ques ion. S uden s emain in he same quad an , he uppe - igh one, by answe ing NMeasu emen s and/o S a is ics. As seen be o e, he consis ency o hese s uden s is limi ed o he Se ies ques ion and ends as soon as hey answe he Cu es ques ion. Simila ly, when he pe cen ages o he ca ego y T u h and O icial Value a e added (lowe - igh quad an ) in he h ee pa s o he Se ies ques ion, he e is conside able consis ency in he esponses o seconda y s u- den s. Again he consis ency ends when answe ing he Su p ise ques ion. I appea s ha consis ency is mainly he consequence o a simila i y in he kind o hing being asked in he ques ion. By con as , he e is no consis ency when s uden s a e dis- cussing di e en kinds o labwo k (such as ha in ol ing quan i a i e and quali a i e da a). The same can be said o he ca ego y P o ocol: Uni e si y s uden s choose i ai ly consis en ly wi hin he h ee pa s o he Se ies ques ion, d awing upon de ails o he si ua ion. The consis ency is limi ed o he h ee pa s o he Se ies ques ion. 516 M.-G. S ´ ER´ EETAL. F om his summa y o he c oss-ques ion s udy, i is impossible o conclude ha s uden s ha e unique philosophical posi ions which accoun o he way hey p ocess da a, use heo y du ing an expe imen , and conside he espec i e oles o heo y and da a in he de elopmen o science o pe sonal knowledge. I is impossible o in e p e a gi en s uden ’s answe s as inspi ed by he use o a obus men al model. The same pe sons can possibly change hei posi ion on Figu e 1 each ime a di e en ques ion is answe ed, each ime hey ha e o judge da a o conclude in di e en si ua ions, o ins ance in case o passing om quali a i e o quan i a i e esul s. I means ha he shape o he ques ion i sel has an in luence on s uden s’ answe s, and ha si ua ions hemsel es a e add essed by easoning wi h di e en epis emological and on ological unde pinnings. Conclusion and Pe spec i es o Teaching Ou su ey con ibu es o li e a u e add essing s uden s’ images o science. Ou pa icula con ibu ion is ha we ha e add essed he images o science used by s uden s when alking abou he speci ic si ua ion o labwo k. We ha e p oposed a me hod o classi ying s uden s’ esponses in e ms o hei unde lying epis emology and on ology, and conside ed whe he indi iduals’ esponses ac oss a ange o ques- ions can be cha ac e ised by disc e e posi ions abou he espec i e ole o heo y and da a in expe imen a ion. We p esen e idence o he con a y, and we gi e examples o ypical ways o hinking conce ning labwo k, as well as examples o consis encies and inconsis encies. Some o he consequences a e me hodological. As i is necessa y o ake he cha - ac e is ics o si ua ions in o accoun in he in e p e a ions o esponses, i is no alid o cha ac e ise on ological o epis emological aspec s o easoning by he use o gene al, decon ex ualised ques ions. Such w i en ques ions do ha e some uses i si ua ions a e desc ibed. We will sugges below how eache s could ui ully ex end his so o sho ques ions du ing labwo k. The e a e also pedagogical consequences. Though i is no alid o a ibu e unique philosophical posi ions o indi iduals, i canno be denied ha indi iduals do indeed d aw upon philosophical posi ions in di e en si ua ions. The esul s o his su ey and ou obse a ions in labo a o y classes a he same eaching le el, con e ge on his poin . These philosophical posi ions a e no s able. Some a e app op ia e and some a e inapp op ia e. I appea s ha many s uden s do no lea n wha coun s as app o- p ia e scien i ic easoning as a esul o cu en eaching app oaches. I his is seen as an impo an cu icula goal o science educa ion, hen explici eaching abou he a ious ela ionships ha can exis be ween heo y and da a in di e en si ua ions is equi ed. This migh in ol e, o example, showing s uden s how heo e ical issues in luence expe imen al design in se e al si ua ions, o showing how he choice o measu ing ins umen s in luences indings when a quan i y is ob ained by means o an indi ec measu emen . In e ec , i is necessa y o in oduce s uden s o di e en me hods and concep s o da a p ocessing. IMAGES OF SCIENCE LINKED TO LABWORK 517 The ela ionships be ween heo y and da a ce ainly a y om one discipline o ano he , be ween sub-disciplines, and om one si ua ion o ano he ( hough no usu- ally om one scien is o ano he in a gi en si ua ion). We belie e ha he s udy o hese ela ionships in ac ual si ua ions would ha e he po en ial o ans o m lab- wo k in science educa ion away om a p ocess o he ou ine applica ion o algo- i hms, owa ds a mo e c i ical p ocess ha in ol es making and jus i ying labo a o y decisions. The ques ionnai e epo ed in his pape in ol es a limi ed numbe o opics, and he o mula ion o ques ions is always open o imp o emen . Howe e , we belie e ha eache s could elabo a e simila ques ions wi h he pedagogical aim o ge ing s uden s o e lec on he unde lying a ionale o hei labo a o y ac ions. Fu he - mo e, eache s would become mo e sensi i e o he epis emological and on ological commi men s ha hei s uden s we e likely o use in speci ic eaching and lea ning si ua ions. The ques ionnai e discussed in his a icle, as well as he o he s gene a ed in he LSE p ojec (Leach e al., 1998; Sé é, 1998) can be ega ded as a da abank o eaching esou ces. They can be adap ed o di e en con ex s so as o con ibu e o es ablishing he ounda ions o a me hod o science eaching which will in es iga e and p o ide he means o a ain philosophical objec i es in eal labo a o y con ex s. Acknowledgemen s We would like o hank P o esso Rose Pin o o he Au onomous Uni e si y o Ba celona (Spain), as well as he membe s o he DidaScO g oup o he Uni e si y o Pa is 11 (O say, F ance) o hei help in ca ying ou he su ey. We a e indeb ed o he ansla o , P o esso Pamela Fabe , om he depa men o T ansla ion and In e p e a ion o he Uni e si y o G anada (Spain). The da a se d awn upon in his pape was collec ed as pa o he p ojec Labwo k in Science Educa ion, suppo ed by DGXII o he Eu opean Commission. No es 1. The ull i le o he p ojec was Imp o ing Science Educa ion: Issues and e- sea ch on inno a i e empi ical and compu e -based app oaches o labwo k in Eu- ope. I was unded in 1996–1998 by he Eu opean Commission. I in ol ed esea ch eams om F ance (Co-o dina o ), Denma k, Ge many, G ea B i ain, G eece, I aly, Spain. 2. This so o ca ego isa ion p o ides esul s ha a e e y di e en om an analy- sis in e ms o he use o key wo ds by s uden s, since e y di e en posi ions may be exp essed wi h he same wo ds (unce ain y, accu acy, p ecision, e c.). 3. I mus be no ed ha e y ew esponden s (less han en in all) we e so nai e as o s a e ha a ma hema ical exp ession should always e lec a law. This simplis ic iew seems no o be lea n om class oom p ac ice. 518 M.-G. S ´ ER´ EETAL. 4. In F ance, he disciplines o Biology and Geology a e augh as one subjec in uppe seconda y schools: Sciences de la Vie e de la Te e. 5. To his end, a sample o 250 s uden s has been su icien o d aw conclusions. Co espondence: Ma ie-Gene iè e Sé é, Uni e si é Pa is XI, DidaScO g oup, 91405 O say, F ance E-mail: ma ie-gene ie[email p o ec ed]sud. Re e ences Allie, S., Bu le , A., Kaunda, L., Campbell, B., & Lubben, F. (1998). Fi s -yea physics s uden s’ pe cep ions o he quali y o expe imen al measu emen s. In e na ional Jou nal o Science Educa ion,20, 447–459. B ickhouse, N. W. (1990). Teache s’ belie s abou he na u e o science and hei ela ionship o class oom p ac ice. Jou nal o Teache Educa ion,41, 53–62. Chi, M. T. H. (1992). Concep ual change wi hin and ac oss on ological ca ego ies: Examples om lea ning and disco e y in science. In R. Gie e (Ed.), Cogni i e models o science: Minneso a s udies in he philosophy o science (pp. 129–186). Minneapolis. MN: Uni e si y o Minneso a P ess. D i e , R., Leach, J., Milla , R., & Sco , P. (1996). Young people’s images o science. Buckingham: Open Uni e si y P ess. Désau els, J., & La ochelle, M. (1998). The epis emology o s uden s: The ‘ hingi ied’ na u e o scien i ic knowledge. In B. F aze , & K. Tobin (Eds.), In e na ional handbook o science educa ion (pp. 115–126). Do d ech , NL: Kluwe . Hodson, D. (1993). Philosophic s ance o seconda y school science eache s, cu - iculum expe iences, and child en’s unde s anding o science; some p elimina y indings. In e change, 24, 41–52. Jou neaux, R., & Sé é, M.-G. (1994). T ai emen s a is ique des ince i udes en Physique: P oblèmes scien i iques e didac iques (S a is ical ea men o un- ce ain ies in physics: Scien i ic and didac ical p oblems). Eu opean Jou nal o Physics, 15, 286–292. Koulaidis, V., & Ogbo n, J. (1989). Philosophy o Science: An empi ical s udy o a eache s’ iew. In e na ional Jou nal o Science Educa ion,11, 173–184. La che , C., Sé é, M.-G., & Jou neaux, R. (1994). Di icul és dans l’app en issage du mesu age (Di icul ies in lea ning measu emen ). Ensenanza de las Ciencias, 12, 27–42. Leach, J. (in p ess). The use o seconda y da a in eaching abou da a analysis in a i s yea unde g adua e biochemis y cou se. In D. Psillos, & H. Niedde e (Eds.), Teaching and lea ning in he science labo a o y. Do d ech , NL: Kluwe . Leach, J., Milla , R., Ryde , J., & Sé é, M.-G. (2000). Epis emological unde s anding in science lea ning: The consis ency o ep esen a ions ac oss con ex s.Lea ning and Ins uc ion, 10, 497–527. IMAGES OF SCIENCE LINKED TO LABWORK 519 Leach, J., Milla , R., Ryde , J., Sé é, M.-G., Hammele , D., Niedde e , H., & Tsel es, V. (1998). Su ey 2: S uden s’ images o science as hey ela e o lab- wo k lea ning. (Wo king Pape 4, Labwo k in Science Educa ion P ojec ). Leeds: CSSME. Lede man, N. (1992). S uden s’ and eache s’ concep ions o he na u e o science: A e iew o he esea ch. Jou nal o Resea ch in Science Teaching, 29, 331–359. Lubben, F., & Milla , R. (1996). Chid en’s ideas abou he eliabili y o expe imen al da a. In e na ional Jou nal o Science Educa ion,18, 955–968. Mo ime , E. F. (1995). Concep ual change o concep ual p o ile change? Science and Educa ion, 4, 267–285. Po lan A iza, R., Ri e o Ga cia, A., & Ma in del Pozo, R. (1998). Conocimien o p o esional y epis emologia de los p o eso es, II: Es udios empi icos y con- clusiones. Ensenanza de las Ciencias (Teache s’ p o essional knowledge and epis emology: Empi ical s udies and conclusions), 16, 271–288. Rowell, J. A., & Dawson, C. J. (1983). Labo a o y coun e examples and he g ow h o unde s anding in science. In e na ional Jou nal o Science Educa ion, 5, 203– 215. Sama apunga an, A. (1992, Ap il). Scien is s’ concep ions o science: A s udy o epis emic belie s. Pape p esen ed a he Annual Mee ing o he Ame ican Educa ional Resea ch Associa ion, San F ancisco, CA. Sé é, M.-G. (1998). Imp o ing science educa ion: Issues and esea ch on inno a i e empi ical and compu e -based app oaches o labwo k in Eu ope. Final epo o he p ojec Labwo k in Science Educa ion, Ta ge ed Socio-Economic Resea ch, Science, Resea ch & De elopmen , Eu opean Commission. B ussels, Belgium: The Eu opean Commission. Tyson, L. M., Ven ille, G. J., Ha ison, A. G., & T eagus , D. F. (1997). A mul idi- mensional amewo k o in e p e ing concep ual change e en s in he class oom. Science Educa ion,81, 387–404. Vosnadiou, S. (1994). Cap u ing and modelling he p ocess o concep ual change. Lea ning and Ins uc ion, 4, 45–69. 520 M.-G. S ´ ER´ EETAL. Appendix: Full Tex o he Ques ionnai e The Theo y Ques ion The ollowing pai o s a emen s is abou how scien is s wo k wi h da a. In each case, please say whe he you ag ee wi h he s a emen , whe he you disag ee wi h i , o whe he you a e no su e whe he you ag ee o disag ee wi h i . Ag ee wi h Disag ee No s a emen wi h su e s a emen Good scien is s ough o ha e heo e ical assump ions which in luence hei analysis o da a. Good scien is s ough o be neu al and ob- jec i e: i is no accep able o scien is s’ heo e ical assump ions o in luence hei analysis o da a. Please explain you easoning: o ins ance gi e he example in biology, chemis y, physics e c... ha you ha e in mind when answe ing. The Cu es Ques ion Two esea che s, A and B, ha e ob ained a se ies o poin s ep esen ed below in he i s pic u e. They gi e wo in e p e a ions shown in he second and hi d pic u e. Measu emen esul s: IMAGES OF SCIENCE LINKED TO LABWORK 521 In e p e a ion o A: In e p e a ion o B: In you opinion, wha could be he easons o he wo di e en cu es? I he wo esea che s would mee and discuss, wha kind o a gumen s you would expec hem o use? Which help can be expec ed om a compu e o help hem decide? Would he compu e be able o make a decision be ween he wo? The Su p ise Ques ion A class is ca ying ou a p ac ical ask in he labo a o y. They a e es ing samples o lea es om plan s which ha e been kep o a ew days in a da k cupboa d, and 522 M.-G. S ´ ER´ EETAL. o he s which ha e been in he ligh . The es s a e o he p esence o s a ch in he lea es. The eache knows ha he e should be s a ch in he lea es om plan s kep in he ligh , bu none in he lea es o plan s kep in he da k. As his is a s anda d esul in all he ex books, he s uden s will be expec ed o know i in hei examina ions. Bu he s uden s do no ye know wha he esul s should be. Some g oups o s uden s in he class ge he esul s he eache expec s. Howe e , some g oups o s uden s do no ge a posi i e s a ch es o any o hei lea es. And some g oups ge a posi i e s a ch es o some o he lea es om plan s kep in he ligh (bu no o all o hem) and also o some o he lea es om plan s kep in he da k (bu no o all o hem). Wha should he eache do in his si ua ion? Explain why you hink his is he bes hing o do: The Se ies Ques ion When he esul is di e en om one measu emen o ano he , how o conclude? Two g oups o nu i ionis s ha e been asked o measu e he mass o 1decili e o oli e oil. Each g oup ecei es one decili e and makes i e measu emen s. These a e hei esul s, a e ha ing classi ied hem in inc easing o de : G oup Measu emen s (g) Mean (g) A94.9 96.6 97.1 98.1 98.9 97.12 B91.9 96.5 97.5 97.5 102.2 97.12 Wi h which o he ollowing s a emen s do you mos closely ag ee? AG oup A’s esul s a e be e , because he ange be ween he la ges and he smalles measu emen is less. BG oup B’s esul s a e be e , because he measu emen s co e a wide ange o alues. CBo h se s o measu emen s gi e he same esul DF om his in o ma ion i is no possible o conclude abou he quali y o measu emen s. Please explain you easoning: Ano he wo g oups o nu i ionis s ha e been asked o measu e he mass o 1decil- i e o nu oil. Each ecei es one decili e and makes i e measu emen s. These a e hei esul s: IMAGES OF SCIENCE LINKED TO LABWORK 523 G oup Measu emen s (g) Mean (g) A92.4 93.3 93.4 94.0 94.4 93.5 B91.9 93.3 94.9 95.0 96.9 94.4 Wi h which o he ollowing s a emen s do you mos closely ag ee? AG oup A’s esul s a e be e , because he ange be ween he la ges and he smalles measu emen is less. BG oup B’s esul s a e be e , because he measu emen s co e a wide ange o alues. CF om he wo se s a alue can be chosen DF om his in o ma ion, i is no possible o conclude and choose a alue Please explain you easoning: I you chose D, please say wha addi ional in o ma ion you would need in o de o be able o say which g oup’s esul s a e be e : A g oup o nu i ionis s wan s o compa e he masses o wo di e en ege able oils. They make i e measu emen s, o each decili e o one so o oil. These a e hei esul s: Measu emen s (g) Mean (g) Oil X 92.2 92.6 93.2 94.1 94.4 93.3 Oil Y (F ance) 92.6 93.2 93.3 94.0 94.4 93.5 Oil Y (Spain) 93.6 94.1 94.3 95.0 95.4 94.5 F om hese esul s, he g oup concludes ha “speci ic mass is g ea e o oli e oil han o nu oil.” Do you ag ee wi h his conclusion? Ag ee Disag ee No Su e Please explain you easons o ag eeing o disag eeing: