Students' Perceptions on Engaging Database Domains and Structures

This is a sel -a chi ed e sion o an o iginal a icle. This e sion may di e om he o iginal in pagina ion and ypog aphic de ails. Au ho (s): Ti le: Yea : Ve sion: Copy igh : Righ s: Righ s u l: Please ci e he o iginal e sion: CC BY 4.0 h ps://c ea i ecommons.o g/licenses/by/4.0/ S uden s' Pe cep ions on Engaging Da abase Domains and S uc u es © 2023 he Au ho s Published e sion Miedema, Daphne; Taipalus, Toni; Ai aloglou, E himia Miedema, D., Taipalus, T., & Ai aloglou, E. (2023). S uden s' Pe cep ions on Engaging Da abase Domains and S uc u es. In SIGCSE 2023 : P oceedings o he 54 h ACM Technical Symposium on Compu e Science Educa ion : Volume 1 o 2. ACM. h ps://doi.o g/10.1145/3545945.3569727 2023 S uden s’ Pe cep ions on Engaging Da abase Domains and S uc u es Daphne Miedema [email p o ec ed] Eindho en Uni e si y o Technology Eindho en, he Ne he lands Toni Taipalus [email p o ec ed] Uni e si y o Jy äskylä Jy äskylä, Finland E himia Ai aloglou [email p o ec ed] Open Uni e si y Hee len, he Ne he lands Del Uni e si y o Technology Del , he Ne he lands ABSTRACT Se e al educa ional s udies ha e a gued o he con ex ualiza ion o assignmen s, i.e., o p o iding a con ex o a s o y ins ead o an abs ac o symbolic p oblem s a emen . Such con ex ualiza ion may ha e bene icial e ec s such as highe s uden engagemen and lowe d opou a es. In he domain o da abase educa ion, ex books and educa o s ypically p o ide an example da abase o con ex . These a e hen used o in oduce key concep s ela ed o da abase design, and o illus a e que ying. Howe e , i emains uns udied wha kinds o da abase con ex s a e engaging o no ices. In his pape , we s udy which aspec s o da abase domain and complexi y s uden s ind engaging h ough s uden e lec ions on a da abase c ea ion assignmen . We iden i y six ac o s ega ding engaging domains, and i e ac o s o engaging complexi y. The main ac- o o domain- ela ed engagemen was Pe sonal in e es , he main ac o o complexi y engagemen was Ma ching in o ma ion e- qui emen s. Ou indings can help da abase educa o s and book au ho s o design engaging exe cise da abases a ge ed o no ices. CCS CONCEPTS •Applied compu ing → Educa ion;•In o ma ion sys ems → Rela ional da abase model. KEYWORDS da abase, educa ion, con ex , engagemen , design ACM Re e ence Fo ma : Daphne Miedema, Toni Taipalus, and E himia Ai aloglou. 2023. S uden s’ Pe cep ions on Engaging Da abase Domains and S uc u es. In P oceedings o he 54 h ACM Technical Symposium on Compu e Science Educa ion V. 1 (SIGCSE 2023), Ma ch 15–18, 2023, To on o, ON, Canada. ACM, New Yo k, NY, USA, 7 pages. h ps://doi.o g/10.1145/3545945.3569727 1 INTRODUCTION Da a managemen skills a e becoming mo e and mo e cen al o he asks o a compu e scien is . Da a managemen is o high im- po ance in ins i u ions using la ge amoun s o da a, o o ganizing da a dis ibu ion and decen alized que y execu ion. I is also co e o he a ea o a i icial in elligence, as no model can be ained This wo k is licensed unde a C ea i e Commons A ibu ion In e na ional 4.0 License. SIGCSE ’23, Ma ch 15–18, 2023, To on o, ON, Canada. ©2023 Copy igh held by he owne /au ho (s). ACM ISBN 978-1-4503-9431-4/23/03. h ps://doi.o g/10.1145/3545945.3569727 wi hou da a. As such, da a managemen echniques a e also an essen ial pa o compu e science educa ion. Da a managemen is ypically augh by means o oy examples [ 23 ], wi h domains such as s o e o de ing sys ems, mo ie en als and company employees [ 21 ]. These examples a e accessible: mos s uden s ha e some idea o wha da a could be in ol ed in hese domains, which hey can use as a sca old o emembe he da abase schema. As such, he domain o he da a is he inhe en con ex o he da abase, bu de e mining wha makes a good da abase con ex no been s udied in-dep h be o e. We can deepen s uden s in e es in cou se ma e ial by ensu ing ha he p ojec s hey wo k on a e au hen ic, and by gi ing hem a choice in he wha and how o he p ojec [ 2 ]. Fo example, eaching CS1 in he s uden s majo ’s con ex has led o inc eased s uden success a es [ 22 ]. Inc eased in e es in a p ojec may also lead he s uden s o spend mo e ime on an assignmen [ 6 ]. Fu he mo e, inc eased in e es makes s uden s become mo e ac i e lea ne s, which in u n inc eases mo i a ion and lea ning [ 11 ]. Howe e , i has been shown ha mo e complex da abases ha e he possible downside o making lea ning SQL mo e di icul [21]. In his pape , we examine ac o s ha s uden s hink make da abases engaging, o answe he ques ion: Wha kind o da abases do no ices deem engaging o s udy da abase design, implemen a ion and que ying, and why? We spli his in o wo esea ch ques ions: RQ1 Which ac o s desc ibe an engaging da abase applica ion do- main o s uden s? RQ2 Which ac o s a e key o an engaging da abase in e ms o s uc u al complexi y? To answe hese ques ions, we an a s udy in which s uden s de- signed and implemen ed hei own ela ional da abase and que ies. We asked hem o elabo a e on wha hey hough was engaging abou he domain o hei da abase, and he complexi y o i . We iden i ied six domain ela ed ac o s, and i e complexi y ela ed ac o s. Ou indings can help eache s and ex book w i e s use and c ea e mo e engaging examples, such ha s uden s can lea n mo e e ec i ely. 2 RELATED WORK Whe he o no s uden s will bene i om eaching and assign- men s ha a e con ex ualized has been a poin o discussion in he compu e science educa ion esea ch communi y o some ime now. On one hand, esea che s a gue ha con ex can inc ease mo i a ion and educe s uden d opou by engaging he s uden in a conc e e way by add essing eal-wo ld ques ions [ 5 , 7 , 13 , 19 ]. On 122 SIGCSE ’23, Ma ch 15–18, 2023, To on o, ON, Canada. Daphne Miedema, Toni Taipalus, and E himia Ai aloglou he o he hand, con ex ualiza ion can lead o nega i e side-e ec s such as high cogni i e load [ 4 ] and hinking o newly lea n hings as only applying in he cu en si ua ion [12]. In CS educa ion esea ch, con ex ualiza ion has ecei ed a en- ion mainly ou side he ealm o da abase educa ion. In pe haps he s udy mos ela ed o his one, Yue used a semi- ealis ic da abase (Sakila by MySQL) o inc ease engagemen in hei da abase cou se by mo ing away om he ypical “ oy” example da abases [ 23 ]. The da abase was ully in eg a ed in he cou se and used ac oss mul iple assignmen s. S uden s ound hese assignmen s bo h in e - es ing and use ul, al hough hey also ound i di icul because i is mo e complex han he examples used in ex books. This inc ease in da abase complexi y has also been shown o nega i ely a ec success ul que y o mula ion among no ices [21]. Bou ie e al. an a s udy o see whe he s uden s pe o m be e on con ex ualized p og amming exe cises, e sus plain, unembel- lished ones [ 4 ]. Fo his pu pose, hey made wo e sions o he same exe cise, wi h and wi hou con ex . The con ex ualized one used as onomy as i s domain, and he plain one jus men ioned numbe s. They ound ha he e was no signi ican di e ence in pe o mance be ween he wo e sions o he exe cise [ 4 ]. Simi- la esul s we e also ob ained by Leinonen, Denny and Whalley, who saw ha , e.g., s uden s ound i easie o implemen a linea equa ion sol e when hey had con ex o he p oblem, han wi h- ou . Fu he mo e, con ex led o a longe ime on ask in seconds, bu he con ex ualized p oblems also equi ed ewe a emp s [ 15 ]. C aig, Smi h and Pe e son an a s udy in which hey compa ed pe o mance on con ex ualized p og amming assignmen s wi h pe o mance on non-con ex ualized assignmen s [ 8 ]. In con as o Leinonen e al., hey did no ind a measu able e ec , leading hem o he conclusion ha o he ac o s, besides con ex , impac di icul y and pe o mance mo e. Beyond hese discussions on pe o mance, i is also in e es ing o see wha s uden s hemsel es hink o con ex ualiza ion. In 2004, Rich, Pe y and Guzdial an a s udy in which hey compa ed a con ex ualized CS1 cou se o a s anda d one [ 19 ]. Thei aim was o a ac and e ain emale s uden s in hei CS1 cou se, and hei esul s sugges ha hey we e success ul: emale s uden s exp essed g ea e enjoymen and in e es in he cou se han hose in he s anda d cou se [ 19 ]. In simila ein, Clibu n and Mille ound ha s uden s liked o ha e a choice as o how o demons a e hei skill le els, a he han all ollowing he same assignmen [6]. I is also wo h no ing ha while con ex ualiza ion is no a sim- ple concep , measu ing i s bene i s can also be di icul . Rade e al. examined whe he s uden s would be mo e in e es ed in socially ele an con ex ualized p og amming asks han in o he ypes o asks [ 18 ]. In a su ey, s uden s esponded ha hey like assign- men s ha show how compu e science can be used o bene i socie y. Howe e , when anking p ojec s in o de o mos in e - es ing o leas in e es ing, s uden s showed a highe p e e ence o simple game-based p ojec s. This p e e ence may be pa ially explained by he complexi y and open-endedness o he human- i a ian p oblem desc ip ions [ 18 ]. Addi ionally, in he con ex o compu e secu i y educa ion, DeWi and Cicalese ound ha s u- den s we e mo e engaged wi h eal-wo ld s o ies o e ic i ious scena ios. Howe e , con ex and con en need o be in balance o accommoda e bo h adi ional compu e science s uden s, and hose ha wan o lea n o apply echniques in eal-wo ld con ex [9]. Finally, choosing he igh con ex may be di icul . Ya osh and Guzdial applied a na a i e con ex ualiza ion in hei media com- pu a ion CS2 cou se, in which hey used one unning example h oughou he cou se [ 22 ]. They ound ha while mos o he s uden s ag eed ha con ex made he class mo e in e es ing, some s uden s ound he con ex o a he aking away om lea ning he cou se con en s in g ea e dep h. The e o e, engagemen is no only abou he p esence o absence o con ex . Some o he dimen- sions o engagemen include connec edness [ 10 ], posi i e emo ions [1, 17, 20], passion [1, 17] and mo i a ion [1, 20]. Engaging assignmen s can make ac i e s uden s ou o passi e ecipien s o knowledge, wi h ac i e s uden s eaching highe le - els o mo i a ion and lea ning [ 11 ]. This leads us o he ques ion o which o hese aspec s o engagemen we can u ilize wi h he con ex o ou da abase assignmen s. In his pape , we ocus on s uden pe cep ions on engaging da abase educa ion. Wi h ou esea ch ques ions, we aim o iden i y ac o s ha educa o s can use o guide exe cise da abase design, as i emains unclea which ac o s con ibu e o engagemen in da abase educa ion. 3 RESEARCH SETTING 3.1 Da a Collec ion The da a we e collec ed om h ee coho s (sp ings o 2019-2021) o mainly hi d yea so wa e enginee ing and in o ma ion sys ems science s uden s a a Finnish uni e si y. The s uden s had o ha e aken a basic cou se on da abases and da a managemen be o e aking his cou se. The s uden s we e shown a da a p i acy s a e- men p io o pa icipa ion, and could hen op -in o he s udy, whe e pa icipa ion ca ied no incen i es o penal ies. Ou o he 68 s uden s, 56 (82%) chose o pa icipa e (15/17 om he i s coho , 8/10 om he second, and 33/41 om he hi d). The pa icipan s we e asked o design and implemen a ela ional da abase ha hey deemed engaging o no ices who a e lea ning da abase design and que ying. The assignmen consis ed o mul iple asks, among which he ollowing wo a e ele an o his s udy: • Choose an applica ion domain o hei da abase ha hey deemed engaging o a no ice o p ac ice que ying, and e lec on why ha pa icula domain is engaging. • Design he da abase s uc u e o be engaging o a no ice o p ac ice que ying, and e lec on why ha pa icula s uc- u e is engaging. The minimum equi emen o he numbe o ables is i e. In his s udy we analyze he s uden s’ e lec ion epo s om he pe spec i e o ou esea ch ques ions. As he language o ins uc ion was Finnish, all ques ions (including he ones abo e), da abases and e lec ions we e w i en in Finnish. All answe s we e ansla ed h ough Google T ansla e and checked by a na i e speake . 3.2 Da a analysis F om he 56 pa icipan s men ioned abo e, 2 p o ided he da a- base bu did no e lec on hei easoning. These s uden s a e no included in he da a, so we we e le wi h 54 pa icipan s in o al. These pa icipan s a e indica ed by P01 h ough P54. 123 S uden s’ Pe cep ions on Engaging Da abase Domains and S uc u es SIGCSE ’23, Ma ch 15–18, 2023, To on o, ON, Canada. Ou analysis was done in ou s eps: mining ac o s, calib a ion o coding, coding emaining da a, and summa izing. We applied di ec ed con en analysis, which is ecognized as a iable op ion when p e ious li e a u e exis s bu is deemed incomple e [14]. Mining ac o s Be o e we s a ed ou analysis, we used exis ing li e a u e o c ea e a lis o ac o s o sea ch o in he e lec ions. We s a ed his app oach om he ITiCSE wo king g oup epo on con ex ualized p og amming educa ion [ 4 ]. F om his epo we e ie ed i e ini ial ac o s: Pe cei ed ele ance, Re en ion, Connec ion o p ac ice, Impac , and Explo ing social and e hical issues. All i e o hese we e used o coding bo h domain and complexi y ac o s. Calib a ion o coding A e se ing up his baseline o ac o s, all au ho s indi idually coded he same subse o he da a (20%). Fo his we used bo h he exis ing ac o s and in oduced new ac o s based on he ex in he e lec ions. A e wa ds, we agg ega ed o see whe e (dis)ag eemen s a ose in he coding. To calib a e ou coding, we discussed all cases oge he , as well as any new ac o s ha may ha e been in oduced by any o he au ho s. New ac o s we e ei he in oduced sepa a ely, me ged wi h ela ed ac o s, o disca ded a e discussion. Coding emaining da a Then, wo au ho s spli he emaining da a in hal and coded indi idually. Fo all cases ha we e no immedia ely clea , mee ings we e held o discuss bes i s. No new ac o s we e in oduced du ing his ime. Summa izing A e he coding was done, we pe o med analysis by examining and summa izing all codes pe ac o . This wo k was calib a ed by wo king acco ding o an example p o ided by one o he au ho s. Wi hin his p ocess o summa izing, we also decided o me ge o d op some ac o s, as hey we e no con ibu ing o ou esea ch ques ion. Fi s , we me ged he wo ypes o e sa il- i y om bo h domain and complexi y ques ions, he gis o bo h was e sa ili y in di icul y and e sa ili y in ex ensibili y. We also d opped a complexi y ac o a ound he concep o enough da a o p ac ice, no emp y esul ables, and a complexi y ac o whe e s uden s decided how many columns o c ea e by in ui ion. Nei he o he la e wo can be used as ins uc ional guidelines, which is why we decided o lea e hem ou . 4 RESULTS 4.1 Engaging domains D1. Familia i y. Familia i y, including ease and unde s andabili y, was men ioned as a ac o by 25 s uden s. They specula ed ha a da abase domain ha is amilia h ough, e.g., pe sonal li e, p e ious s udies, o wo k would be engaging. Fo example, P07 w o e: “The da abase ela ed o lis ening o music is in e es ing o he s uden , because almos e e yone uses some kind o music s eaming se ice e e y day.” Se e al s uden s poin ed ou ha a amilia domain dec eases he ini ial h eshold o s a ing he lea ning p ocess owa ds da abase heo y and que y w i ing. P03 w o e: “[...] simple a ge a eas make i easie o in e nalize new hings and main ain in e es in he subjec mo e e ec i ely.” On he o he hand, a domain no necessa ily amilia o many s uden s, bu in u n easy o unde s and, was conside ed ano he op ion o an engaging da abase domain. In con as , s uden s high- ligh ed ha a domain ha would equi e ex ensi e amilia iza ion would ake away om he in ended p ima y lea ning ou comes o a da abase cou se: “[...] inancial da a eposi o y wi h special e ms and complex modeling can cause g ay hai o beginne s [...] and he e o e I don’ hink is as well sui ed o aining.” - P20 D2. Connec ion o p ac ice. Rela ed o Familia i y was he ac o Connec ion o p ac ice, men ioned by 30 s uden s. They specula ed ha unde s anding he da abase domain and, e.g., lea ning o que y in his pa icula domain would also help hem in hei u u e wo k. P07 w o e ha : “I am in e es ed in he music and en e ainmen indus y in gene al, so combining he wo ells me abou my own in e es s and desi e o ge a job in he indus y.” In con as , a domain ha could be conside ed niche o a e would be ela i ely useless ou side he scope o a da abase cou se. Ano he conside a ion included in his ac o was he iew o inc easingly common applica ion domains such as elec onic spo s and elec- onic comme ce. Fo example, P01 w o e: “[...] use ul o wo king li e as i could be eal and buying om an online s o e is g owing all he ime. This will equi e mo e da abase wo ke s who can handle ela ional da abases in ha opic.” D3. Lea ning oppo uni y o challenge. 27 s uden s indica ed ha a domain ha p o ides a lea ning oppo uni y o challenge was engaging. On he one hand, hese lea ning oppo uni ies and challenges could be ela ed o applica ion domains ha a e complex in e ms o he esul ing da abase s uc u e, such ha s uden s lea n mo e abou da abase concep s: “The s uden will also lea n o design mul i- ela ionship ela- ionships, such as wi h mul iple o de s, p oduc s, and p oduc ca ego ies.” - P08 On he o he hand, he e a e lea ning oppo uni ies o be ound in s udying he domain i sel in o de o c ea e co ec schemas: “Wo king wi h a da abase o publica ion in o ma ion can deepen a s uden ’s knowledge o scien i ic publishing, o example, in e ms o knowledge o di e en ypes o publica ions.” - P06 D4. Ve sa ili y The e a e wo sides o e sa ili y as a ac o . The i s is e sa ili y in he dimension o que y di icul y. This is linked o he concep o g ow h mindse o educa ion, and ha o “low loo , high ceiling”, meaning ha assignmen s a e accessible o all s uden s om no ices up, bu ha i also has oom o challenge ad anced s uden s [ 3 ]. The o he side is e sa ili y in a da abase ha is easy o expand in size and complexi y. Fo e sa ili y in que y di icul y, en s uden s a gued ha his was a ac o ha made hei domain engaging: “I wan ed o make a posi ion ha would be clea enough o new s uden s, as well as ha i would con ain complex da a ha could also be used o challenge olde s uden s.” - P10 “[...] he e a e s ill oppo uni ies o a bi mo e complex que ies, which makes his pa icula ly in e es ing.” - P15 124 SIGCSE ’23, Ma ch 15–18, 2023, To on o, ON, Canada. Daphne Miedema, Toni Taipalus, and E himia Ai aloglou The e we e ele en s uden s who deemed e sa ili y in e ms o a da abase s uc u e ha is easy o expand a ac o o engagemen . “I also el ha he opic I chose was b oad enough o implemen a wide a ie y o da abases, bu s ill easy o summa ize.” (P45) Wi h a lexible s uc u e, a no ice could men ally selec pa s o he da abase, whe eas an expe could use all pa s. D5. Pe sonal in e es . The mos p ominen mo i a ion o a io- nalizing an engaging domain was pe sonal in e es . As many as 33 s uden s a gued ha he applica ion domain should be in e es - ing o a no ice. S uden s who a gued o an in e es ing domain ypically also associa ed in e es ing domains wi h connec ion o p ac ice h ough hobbies, such as P16 and P51: “The a ge a ea eels pleasan o me, because he e will be a lo o gigs and lis ening o music, so I imagine ha i will inspi e o he s in he same way.” - P16 “I chose he simpli ied online gaming s o e [...] Examples o a eal p oduc a e he Val e S eam se ice o he GOG se ice. I’ e been playing ideo games in my spa e ime since I was young, so he opic seemed na u al.” - P51 O he s uden s ocused on business ideas hey o hei iends we e se ing up, by de eloping a da abase o hei domain: “The a ge a ea was he business in o ma ion o a small com- pany called [anonymized]. This company is eally in business, bu he business is e y small. [...] Agains his backg ound, I am building a da abase s uc u e and possibly a web applica ion p og am. ” - P43 No unexpec edly, domains chosen because o his jus i ica ion we e di e se in na u e. D6. Social and e hical issues. Fi e s uden s deemed ha an en- gaging applica ion domain should add ess social o e hical issues. S uden s a gued ha he applica ion domain should be one ha has a seemingly conc e e and big impac on ou li es. “[...] a mo e se ious app oach han handling games, o example. I someone makes a small ca eless mis ake in a game cha ac- e ’s ha colo , no one may e en no ice i . On he o he hand, i a pa ien has access o hospi al d ug s o es o he su geon is p e en ed om en e ing he ope a ing oom, he consequences can be e y se ious. [...] The da abase is no designed o en e - ainmen use, bu o ensu e and p omo e he sa e y o people in he hospi al en i onmen .” - P12 A domain which, due o i s impo an na u e, equi ed ca e ul design, was also conside ed a sou ce o engagemen . D7. O he . The e we e i e s uden s who a gued o some hing ou side o he a o emen ioned ac o s. They discussed hei chosen domain’s engagemen wi h espec o he di e si y in da a (1), s uc- u e o he da a (1), and expanding, complemen ing o i e a ing a so wa e p ojec ini ia ed du ing ano he cou se (3). “I wan ed o imp o e my P og amming 2 cou se assignmen , S o y S a is ics, which cu en ly wo ks wi h ex iles. Tha p o- g am is s ill in use by me oday and con ains a ela i ely la ge amoun o da a, so some e sion ha uses a decen da abase could be use ul a some poin , especially i I wan ed o de elop he s uc u e o he da a somehow.” - P32 4.2 Engaging complexi y C1. Simplici y. A commonly a gued cha ac e is ic o an engaging da abase schema, men ioned by 24 pa icipan s, was ha i should be simple and unde s andable. P17 summa izes his as: “The da abase had o be in e es ing o he s uden so a da abase ha looks oo challenging will educe he s uden ’s in e es , I hink.” The main a gumen s in a o o simplici y we e ha i makes i easie o p ac ice w i ing que ies (3 s uden s), o popula e he da a- base (1), o expand i (2), and o a oid e o s (2), which we e also linked o inc eased complexi y due o no maliza ion. Fou s uden s explici ly connec ed schema simplici y o lea ning, as “ he complexi y o he da abase could become a ba ie o lea ning” (P06) and “in e ms o lea ning, a simple da a model is easie o adop ” (P20). Fou s uden s designed a simple da abase o ma ch hei beginne da abase skills, o example: “I wan ed o keep he model as simple as possible, as his is he i s da abase I ha e ac ually s a ed o implemen mysel . The main ocus was on ge ing e e y hing o wo k wi hou a big headache.” - P42 C2. Ma ching in o ma ion equi emen s. 36 s uden s desc ibed designing hei da abase acco ding o he pe cei ed use o in o - ma ion equi emen s. Those equi emen s pe ain o he da abase domain o he so wa e applica ion ha would be based on he da abase. In many cases, s uden s p o ided a jus i ica ion pe able and ela ionship and hey desc ibed he p ocess hey ollowed o come up wi h he inal schema, o example P30 desc ibes: “I s a ed by hinking abou he ea u es I would like o ha e in an imagina y se ice. I se ou o implemen he s uc u e o he da abase h ough an ER diag am and wonde ed wha i ems and wha kind o ela ionships he p ope ies I wan ed equi ed. I d ew an ER diag am and selec ed he app op ia e a ibu es o he objec s and ela ionships. I ans o med he diag am in o a ela ional model acco ding o he ans o ma ion ules.” S uden s commonly desc ibed ha he da abase domain led hem o he de ini ion o speci ic a ibu es o hei da abase, o example: “In choosing he numbe o ables and a ibu es, I s a ed by hinking abou which a ibu es o he games we e essen ial and impo an o me o include in he da abase [.. ] So I didn’ eally delibe a ely choose jus ha numbe o ables and columns, bu he numbe o hem came as dic a ed as needed as he plan p og essed.” - P37 Fou s uden s explici ly men ioned ma ching how a eal wo ld da abase would be c ea ed, o example: “In my opinion, he da abase I ha e c ea ed is use ul o wo king li e, because he da abase is buil o a speci ic eal pu pose, al hough he o ma implemen ed du ing he cou se is no exac ly he same as i is o would be in eal li e.” - P16 Some aspec s ela ed o in o ma ion equi emen s we e men ioned a ely as d i ing he da abase design. These we e he cus ome equi emen s (men ioned by P05), enabling sea ch condi ions (P24), and suppo ing da abase expansion ( wo s uden s), o example: 125 S uden s’ Pe cep ions on Engaging Da abase Domains and S uc u es SIGCSE ’23, Ma ch 15–18, 2023, To on o, ON, Canada. “[...] hough ha i publishe s o de elope s e e wan ed o expand wi h new columns, he changes would be easy when hey al eady ha e sepa a e ables o hem.” - P27 C3. Lea ning oppo uni y o challenge. Nine s uden s a gued ha he da abase schema should suppo lea ning. Mos o hem e- e ed o lea ning abou da abases in gene al o in he con ex o he da abases cou se, bu h ee o hem explici ly e e ed o lea ning as p epa a ion o wo king li e; o example, P06 men ioned: “[...] in my expe ience, a da abase wi h i e ables ha mee s he minimum equi emen s o he assignmen would be oo small o gi e he s uden a co ec unde s anding o he da abases ha will come up in la e wo king li e.” S uden s also ga e us insigh in o how a da abase schema and i s complexi y a ec lea ning. A complex da abase schema enables p ac icing wi h “memo izing concep ual en i ies” (P06), encou ages he use o a da abase schema desc ip ion ( he example o he En i y Rela ionship diag am was men ioned by a s uden ) when o mula - ing que ies, and could “[...] demons a e all he di e en concep s used in da abases (e.g. di e en ypes o ac ables in he Kimball model, di e en ypes o dimensions in he ables, mo e me ada a columns and ables co esponding o he co ec implemen a ion, ables on log da a)” - P20 P21 ag eed wi h he lea ning bene i s o a simple da abase schema: “I hink lea ning he basics is suppo ed by mo ing om simple schemes o mo e complex ones. Wi h a simple schema da abase, i is good o p ac ice making que ies and lea n o unde s and logic ela ed o joins, o example.” A complex schema can also make que y o mula ion mo e challeng- ing, as P55 exp esses: “Su e, he e could ha e been mo e ables and columns o he cus ome ’s add ess, o example, bu I eel i wouldn’ ha e con ibu ed mo e o my lea ning. I hink he e we e enough ables and columns o lea n how o use he en i onmen , bu no oo many o make i oo di icul in sol ing p oblems.” C4. Ve sa ili y. Eigh s uden s men ioned ha a da abase schema should be complica ed enough o ep esen a ious da abase cha - ac e is ics ( h ee s uden s) and o enable p ac icing wi h que y w i ing ( i e s uden s). The da abase cha ac e is ics ha s uden s e e ed o we e da a ypes, able ela ionships, and numbe o columns, o example: “[...] he e mus be enough columns in a leas a ew ables o ask meaning ul ques ions. In addi ion, o que ies ha use agg ega e unc ions, g ouping, and so ing a he same ime, i is a good idea o ha e mo e columns in he able han he numbe o unc ions o exp essions used, so ha he unc ion, g ouping, o so ing can be assigned o di e en columns and mo e da a- ich” - P06 The da abase schema should also be e sa ile enough o suppo ex ension o mo e complex que y p ac ice, as summa ized by P21: “In i s cu en o m, he s uc u e o he da abase is, in my opin- ion, sui ably simple, bu a he same ime i makes i possible o ca y ou a wide ange o da abase que ies.” C5. P ac ical cons ain s. 15 s uden s exp essed ha in hei da abase design hey ook in o accoun p ac ical cons ain s, speci - ically assignmen es ic ions (9), ime limi a ions (4), o challenges wi h gene a ing model da a (2). Assignmen es ic ions pe ain mainly o he minimum numbe o i e ables ha was se . The e ec o ime limi a ions is desc ibed by P51: ‘The bigges ac o limi ing he size o he exe cise a he mo- men is ime. The wo k and o he cou ses ake up mos o my ime and I migh ha e liked o build he da abase mo e ho oughly.” Finally, P47 men ioned he e ec o limi a ions o he ool ha was used o popula ing he da abase wi h da a: “Mock da a con ibu ed in pa o he ac ha i la gely de e - mined he ypes o da a I used in he ables [...] I also added some o he columns because o he e ec o he mock da a, because as I looked a i I no iced mo e hings I had p e iously o go en abou in he design.” C6. O he . Th ee s uden s men ioned desi able cha ac e is ics o he da abase schema ha did no i in o any o he a o emen ioned ones. Those we e o limi he schema o he mos signi ican in- o ma ion and “ ocus only on he mos impo an [ ables]” (P08), o ma ch he scope o he cou se and he e o e a oid adding passwo d da a “because wi h passwo ds you should be awa e o secu i y issues ha he cou se doesn’ ocus on anyway” (P27), and o a oid adding calcula ed columns (P33). 5 DISCUSSION In Sec ion 3.2 we in oduced ha ou con en analysis was led by bo h he da a and he wo k by Bou ie e al. on con ex in CS Educa ion [ 4 ]. We’ll now e lec on whe he he ac o s we iden i ied a e in ag eemen wi h hei s. Pe cei ing ele ance. This heme is e lec ed in mul iple iden- i ied ac o s, bu mos p ominen ly in D3: Lea ning oppo uni y o challenge. The s uden s who men ioned his ac o we e looking o lea n abou ei he hei domain o da abases in mo e de ail, show- ing ha hey saw he ele ance o hese hings. We can also a gue ha i links o D2: Connec ion o p ac ice and D6: Social and e hical issues, bu as we adap ed hese di ec ly om al e na i e ac o s by Bou ie e al. [4], we lea e hem ou o he discussion he e. Re en ion. This ac o did no come up in any o ou s uden s’ answe s. Howe e , his is likely o be an a i ac o he assignmen as a sho - e m e lec ion. We’d like o a gue ha his heme is implici ly e lec ed in D3: Lea ning oppo uni y o challenge, as s uden s ocused on lea ning which hope ully leads o e en ion. Connec ion o p ac ice. We adop ed his heme in ou con en analysis and ound enough e idence ha we kep i in as a s and- alone ac o . Impac . Bou ie e al. w i e abou his ac o as impac on he b oade use communi y. This ac o came up egula ly and is mos ly coded unde Pe sonal in e es , whe e s uden s men ion using hei assignmen da abase o a s a -up by hemsel es o o helping hei iends. Explo ing social and e hical issues. This was kep as a sepa- a e ac o : D6, bu i was men ioned only spo adically. In li e a u e, his is desc ibed as a ac o ha may help emale s uden s s ay 126 SIGCSE ’23, Ma ch 15–18, 2023, To on o, ON, Canada. Daphne Miedema, Toni Taipalus, and E himia Ai aloglou engaged in he ma e ial. Un o una ely, as we do no ha e gende iden i y da a o ou s uden s, no conclusions can be d awn on his. A ac o ha a ose om ou indings, and is no men ioned by Bou ie e al. is Ve sa ili y. I is wo h no ing ha he wo no ions o e sa ili y (in business domains (D4), and in da abase complexi y (C4)) had sub le bu dis inc di e ences. By domain e sa ili y, he s uden s a gued o engagemen h ough being able o men ally di ide he da abase in o sel -con ained subse s. This allows s uden s o ake he subse ha i s hei expe ience le el and p ac ice wi h ha . On he o he hand, a da abase wi h s uc u al e sa ili y was seen o acili a e engagemen h ough being easy o expand. On he pe o mance side o he con ex ualiza ion discussion, he e a e some conce ns wi h he cogni i e load associa ed wi h con ex ualiza ion. Howe e , we wan o no e ha da a managemen and da abases canno be augh ully de-con ex ualized: we need da a on some domain o ill ou da abases and in o m ou que ies. As such, i seems be e o adap (some o ) he ac o s ha ou s uden s wish o han o selec some hing andomly o keep doing wha we ha e always done. Fu he mo e, ou indings expand on he no ion o “use ul and in e es ing” exe cise da abases p esen ed by Yue, who ound ha s uden s p e e mo e ealis ic da abases in lea ning da abase ela ed opics, as hese da abases a e pe cei ed o mo e closely e lec he da abases he s uden s a e likely o encoun e in hei u u e wo k [ 23 ]. Howe e , Taipalus showed ha mo e complex da abases migh nega i ely a ec how s uden s lea n o que y [ 21 ]. Howe e , Taipalus also poin ed ou ha mo e complex da abases a e no necessa ily synonymous o mo e ealis ic da abases, and s uden engagemen migh no depend on ealism, bu a he on pe cei ed ealism. I a eache can use a ela i ely simple exe cise da abase ha is also a gued o be a ealis ic one, his migh be he key o suppo ing bo h que y o mula ion lea ning and engagemen . Ou indings suppo Yue’s obse a ions, as se e al s uden s poin ed ou ha lea ning oppo uni ies and challenges sus ain hei engagemen . Howe e , we also obse ed ha he e a e many o he , some imes con adic ing ac o s which all ela e o s uden engage- men . A guably, and al hough i is possible ha a single exe cise da abase may no ul ill all he engagemen conside a ions, each- e s should conside he ac o s p esen ed in Sec ion 4.1 when hey design exe cise da abase domains and complexi ies. Finally, o ou analysis, we chose o wo k wi h a da abase as- signmen and e lec ion, ins ead o he mo e common ques ionnai e app oach. Rade e al. [ 18 ] in hei s udy ound ha wha s uden s say hey wan o do, di e s om wha hey ac ually do. The ad an- age o ou app oach is ha s uden s had o commi o he domain hey chose. They could no a bi a ily selec some hing, bu had o conside he implica ions o hei choice o he de elopmen o he da abase and que ies. 5.1 Limi a ions and h ea s o alidi y We no e se e al limi a ions o ou s udy. Fi s o all, he assignmen equi ed all da abases o con ain a leas i e ables. Al hough we do no d aw any conclusions om calcula ed complexi y me ics, his equi emen was o en echoed in s uden s a gumen a ion on engagemen . Gi en he indings o Rade e al. [ 18 ] ha s uden s p e e simple games, and ac o C1: Simplici y, we belie e ha s u- den s’ beha io migh ha e been di e en wi hou his es ic ion. Second, he e is he issue o s uden s sel -selec ion o pa ic- ipa ion in his s udy. Al hough all s uden s in he cou se had o comple e he exe cises we s udy in his pape , i may be ha pa ic- ipa ing s uden s we e mo e mo i a ed o deli e a good esul . Finally, we e lec on he alidi y o ou s udy based on he de ini ions by Maxwell [ 16 ]. As ou s udy consis s o documen s con aining s uden a gumen a ion, he e is li le isk o misunde - s anding o misin e p e ing s uden s’ accoun s on he side o he au ho s (high desc ip i e alidi y). Howe e , he e is a h ea he e, as we canno gua an ee ha all s uden s unde s ood he ques ions co ec ly. One clea example o his issue is shown in ac o D5: Pe - sonal in e es , whe e some s uden s ou lined engagemen in e ms o hei own pe sonal in e es , ins ead o in e es o all no ices. All au ho s wo ked oge he on he con en analysis, leading o highe likelihood o eliable da a coding. Al hough some in e - p e a ion o s uden answe s is equi ed o ex ac he ac o s, we belie e ha he equi ed le el o in e p e i e alidi y is me , gi en ha ou analysis is quali a i e and uses quo es om he s uden answe s. Finally, al hough ou sample size can be conside ed small o quan i a i e s udies, i is on he la ge size o quali a i e s ud- ies. As such, we belie e ha he ac o s a e ep esen a i e (bu no exhaus i e) o o he s uden popula ions as well. We wan o conclude his discussion sec ion by acknowledging ha , al hough we can gain in o ma ion abou s uden s’ leisu e pu sui s, i is di icul o c ea e app op ia e assignmen s based on his in o ma ion. To co ec ly con ex ualize acco ding o Coope and Cunningham, we should make he chosen con ex a heme h oughou he cou se [ 7 ], which is a di icul unde aking. One ba ie iden i ied by Blumen eld e al. is ha he assignmen s need o be di icul enough o s uden s o lea n abou he subjec in he equi ed dep h [ 2 ]. I may be ha no all con ex ualized assignmen s a e o he same di icul y, as Clibu n and Mille hypo hesize abou he assignmen s in hei wo k [ 6 ]. Rade e al. also concluded ha some o hei con ex ualized assignmen s we e oo complex and open-ended [ 18 ]. The e o e, he e is a ade-o o be na iga ed be ween how open we can lea e assignmen s o ma ch s uden in e es , and how consis en and in-dep h assignmen s a e. 6 CONCLUSIONS AND FUTURE WORK In his pape we in oduced ac o s ha s uden s pe cei e as engag- ing o Da abase Educa ion. Ou indings a e based on an assign- men in which s uden s designed hei own da abase schema, da a and que ies. S uden s we e asked o e lec on why hei da abase is engaging. Fo engaging domains, we ound he ac o s Familia i y, Connec ion o p ac ice,Pe sonal in e es and Social and E hical issues. Fo engaging complexi y, we iden i ied Simplici y,Ma ching in o - ma ion equi emen s and P ac ical cons ain s. Two sha ed ac o s a e Lea ning oppo uni y o challenge and Ve sa ili y. This pape ep esen s an explo a ion o wha aspec s o con ex make da abases engaging. Al hough he cu en indings can al- eady be used o upda e da abase assignmen s, u u e wo k should explo e in mo e de ail wha he e ec o adop ing ou indings is on s uden engagemen , and possibly pe o mance. Fu he mo e, i would be in e es ing o compa e how such new ma e ials com- pa e o commonly used ma e ials and oy examples om ex books. Finally, we should explo e which domains s uden s seem o ind in e es ing, and whe he any ends can be ound in his da a. 127 S uden s’ Pe cep ions on Engaging Da abase Domains and S uc u es SIGCSE ’23, Ma ch 15–18, 2023, To on o, ON, Canada. REFERENCES [1] Elizabe h F Ba kley and Clai e H Majo . 2020. S uden engagemen echniques: A handbook o college acul y. John Wiley & Sons. [2] Phyllis C. Blumen eld, Ellio Soloway, Ronald W. Ma x, Joseph S. K ajcik, Ma k Guzdial, and Annema ie Palincsa . 1991. Mo i a ing P ojec -Based Lea ning: Sus aining he Doing, Suppo ing he Lea ning. Educa ional Psychologis 26, 3-4 (1991), 369–398. h ps://doi.o g/10.1080/00461520.1991.9653139 [3] Jo Boale , Jack A Dieckmann, Tanya LaMa , Mi iam Leshin, Megan Selbach-Allen, and G aciela Pé ez-Núñez. 2021. The T ans o ma i e Impac o a Ma hema ical Mindse Expe ience Taugh a Scale. F on ie s in Educa ion 6 (2021), 1–13. h ps: //doi.o g/10.3389/ educ.2021.784393 [4] Dennis Bou ie , Ellie Lo elle e, John Ma a, Bedou Alshaigy, B e A. Becke , Michelle C aig, Jana Jacko a, Robe McCa ney, Ka e Sande s, and Ma k Za b. 2016. No ice P og amme s and he P oblem Desc ip ion E ec . In P oceedings o he 2016 ITiCSE Wo king G oup Repo s (A equipa, Pe u) (ITiCSE ’16). Associa ion o Compu ing Machine y, New Yo k, NY, USA, 103–118. h ps://doi.o g/10. 1145/3024906.3024912 [5] E elyn B annock, Robe Lu z, and Nanne e Napie . 2013. In eg a ing au hen ic lea ning in o a so wa e de elopmen cou se: An expe ience epo . In P oceedings o he 14 h annual ACM SIGITE con e ence on In o ma ion echnology educa ion. 195–200. [6] Daniel C. Clibu n and Susan Mille . 2008. Games, S o ies, o Some hing Mo e T adi ional: The Types o Assignmen s College S uden s P e e . In P oceedings o he 39 h SIGCSE Technical Symposium on Compu e Science Educa ion (Po land, OR, USA) (SIGCSE ’08). Associa ion o Compu ing Machine y, New Yo k, NY, USA, 138–142. h ps://doi.o g/10.1145/1352135.1352184 [7] S e e Coope and S e e Cunningham. 2010. Teaching compu e science in con ex . ACM In oads 1, 1 (2010), 5–8. [8] Michelle C aig, Jacqueline Smi h, and And ew Pe e sen. 2017. Familia Con ex s and he Di icul y o P og amming P oblems. In P oceedings o he 17 h Koli Calling In e na ional Con e ence on Compu ing Educa ion Resea ch (Koli, Finland) (Koli Calling ’17). Associa ion o Compu ing Machine y, New Yo k, NY, USA, 123–127. h ps://doi.o g/10.1145/3141880.3141898 [9] Janine DeWi and Cyn hia Cicalese. 2006. Con ex ual in eg a ion: A amewo k o p esen ing social, legal, and e hical con en ac oss he compu e secu i y and in o ma ion assu ance cu iculum. In P oceedings o he 3 d annual con e ence on In o ma ion secu i y cu iculum de elopmen . 30–40. [10] A Fle che . 2015. De ining s uden engagemen : A li e a u e e iew. h ps: //soundou .o g/2015/03/29/de ining-s uden -engagemen -a-li e a u e- e iew/ [11] James E G occia. 2018. Wha is s uden engagemen ? New di ec ions o eaching and lea ning 2018, 154 (2018), 11–20. [12] Ma k Guzdial. 2010. Does Con ex ualized Compu ing Educa ion Help? ACM In oads 1, 4 (Dec. 2010), 4–6. h ps://doi.o g/10.1145/1869746.1869747 [13] Ni a He mann and Je ey L Popyack. 1995. C ea ing an au hen ic lea ning expe ience in in oduc o y p og amming cou ses. In P oceedings o he wen y- six h SIGCSE echnical symposium on Compu e Science Educa ion. 199–203. [14] Hsiu-Fang Hsieh and Sa ah E Shannon. 2005. Th ee App oaches o Quali a i e Con en Analysis. Qual. Heal h Res. 15, 9 (2005), 1277–1288. h ps://doi.o g/10. 1177/1049732305276687 [15] Juho Leinonen, Paul Denny, and Jacqueline Whalley. 2021. Explo ing he E ec s o Con ex ualized P oblem Desc ip ions on P oblem Sol ing. In Aus alasian Compu ing Educa ion Con e ence (Vi ual, SA, Aus alia) (ACE ’21). Associa ion o Compu ing Machine y, New Yo k, NY, USA, 30–39. h ps://doi.o g/10.1145/ 3441636.3442302 [16] Joseph Maxwell. 1992. Unde s anding and alidi y in quali a i e esea ch. Ha - a d educa ional e iew 62, 3 (1992), 279–301. [17] G ea Schools Pa ne ship. 2016. S uden Engagemen . In Glossa y o Educa ion Re o m. h ps://www.edglossa y.o g/s uden -engagemen / [18] Cyndi Rade , Doug Hakka inen, Ba ba a M. Moskal, and Kei h Hellman. 2011. Explo ing he Appeal o Socially Rele an Compu ing: A e S uden s In e es ed in Socially Rele an P oblems?. In P oceedings o he 42nd ACM Technical Symposium on Compu e Science Educa ion (Dallas, TX, USA) (SIGCSE ’11). Associa ion o Compu ing Machine y, New Yo k, NY, USA, 423–428. h ps://doi.o g/10.1145/ 1953163.1953288 [19] Lau en Rich, Hea he Pe y, and Ma k Guzdial. 2004. A CS1 cou se designed o add ess in e es s o women. In P oceedings o he 35 h SIGCSE echnical symposium on Compu e Science Educa ion. ACM New Yo k, NY, USA, 190–194. [20] Ellen A Skinne and Michael J Belmon . 1993. Mo i a ion in he class oom: Recip ocal e ec s o eache beha io and s uden engagemen ac oss he school yea . Jou nal o educa ional psychology 85, 4 (1993), 571. [21] Toni Taipalus. 2020. The e ec s o da abase complexi y on SQL que y o mula ion. Jou nal o Sys ems and So wa e 165 (2020), 110576. h ps://doi.o g/10.1016/j.jss. 2020.110576 [22] S e lana Ya osh and Ma k Guzdial. 2008. Na a ing da a s uc u es: The ole o con ex in CS2. Jou nal on Educa ional Resou ces in Compu ing (JERIC) 7, 4 (2008), 1–20. [23] Kwok-Bun Yue. 2013. Using a Semi-Realis ic Da abase o Suppo a Da abase Cou se. Jou nal o In o ma ion Sys ems Educa ion 24, 4 (2013), 327–336. h p: //jise.o g/Volume24/n4/JISE 24n4p327.pd 128