scieee Open visual document viewer

Longitudinal Associations between Physical Activity and Educational Outcomes

Kari, Jaana T,Pehkonen, Jaakko,Hutri-Kähönen, Nina,Raitakari, Olli T,Tammelin, Tuija H

Full text

Longi udinal Associa ions be ween Physical Ac i i y and Educa ional Ou comes JAANA T. KARI 1,2 , JAAKKO PEHKONEN 2 , NINA HUTRI-KA ¨HO ¨NEN 3 , OLLI T. RAITAKARI 4,5 , and TUIJA H. TAMMELIN 1 1 LIKES Resea ch Cen e o Physical Ac i i y and Heal h, Jy a¨skyla¨, FINLAND; 2 Jy a¨skyla¨ Uni e si y School o Business and Economics, Jy a¨skyla¨, FINLAND; 3 Depa men o Pedia ics, Uni e si y o Tampe e and Tampe e Uni e si y Hospi al, Tampe e, FINLAND; 4 Resea ch Cen e o Applied and P e en i e Ca dio ascula Medicine, Uni e si y o Tu ku, Tu ku, FINLAND; and 5 Depa men o Clinical Physiology and Nuclea Medicine, Tu ku Uni e si y Hospi al, Tu ku, FINLAND ABSTRACT KARI, J. T., J. PEHKONEN, N. HUTRI-KA ¨HO ¨NEN, O. T. RAITAKARI, and T. H. TAMMELIN. Longi udinal Associa ions be ween Physical Ac i i y and Educa ional Ou comes. Med. Sci. Spo s Exe c., Vol. 49, No. 11, pp. 2158–2166, 2017. Pu pose: This longi udinal s udy examined he ole o leisu e- ime physical ac i i y in academic achie emen a he end o compulso y basic educa ion and educa ional a ainmen in adul hood. Me hods: The da a we e d awn om he ongoing longi udinal Ca dio ascula Risk in Young Finns S udy, which was combined wi h egis e -based da a om S a is ics Finland. The s udy consis ed o child en who we e 12 y (n= 1723, 49% boys) and 15 y (n= 2445, 48% boys) o age a he ime when physical ac i i y was measu ed. The child en we e ollowed up un il 2010, when hei mean age was 40 y . Physical ac i i y was sel - epo ed and included se e al measu emen s: o e all leisu e- ime physical ac i i y ou side school hou s, pa icipa ion in spo s club aining sessions, and pa icipa ion in spo s compe i ions. Indi iduals_ educa ional ou comes we e measu ed wi h he sel - epo ed g ade poin a e age a age 15 y and egis e -based in o ma ion on he yea s o comple ed pos compulso y educa ion in adul hood. O dina y leas squa es models and he ins umen al a iable app oach we e used o analyze he ela ionship be ween physical ac i i y and educa ional ou comes. Resul s: Physical ac i i y in adolescence was posi i ely associa ed wi h educa ional ou comes. Bo h he physical ac i i y le el a age 15 y and an inc ease in he physical ac i i y le el be ween he ages o 12 and 15 y we e posi i ely ela ed o he g ade poin a e age a age 15 y and he yea s o pos compulso y educa ion in adul hood. The esul s we e obus o he inclusion o se e al indi idual and amily backg ound ac o s, including heal h endowmen s, amily income, and pa en s_educa ion. Conclusion: The esul s p o ide e idence ha physical ac i i y in adolescence may no only p edic academic success du ing compulso y basic educa ion bu also boos educa ional ou comes la e in li e. Key Wo ds: ACADEMIC ACHIEVEMENT, EDUCATIONAL ATTAINMENT, PHYSICAL ACTIVITY, REGISTER-BASED DATA Pos compulso y educa ion is one o he mos impo an in es men s in human capi al (4). Besides he s ong co ela ion be ween academic achie emen and la e educa ional a ainmen , p e ious s udies ha e shown ha a highe le el o educa ion is posi i ely associa ed wi h heal h ou comes (8,12,28), employabili y (22), ea nings (4,7), and economic g ow h (14). Apa om he well-known heal h bene i s o physical ac- i i y (18,20,25,39,40), p e ious s udies ha e linked physical ac i i y wi h posi i e e u ns in e ms o academic achie e- men (11,16,17,31,33) and educa ional a ainmen (3,19,23). Fo example, S e ens e al. (33) ound ha physical ac i i y had a posi i e associa ion wi h ma hema ics and eading achie emen among boys and gi ls, whe eas Fox e al. (11) epo ed ha physical ac i i y and spo s eam pa icipa ion we e posi i ely associa ed wi h g ade poin a e age (GPA). Acco ding o Long and Caudill (19), a hle es who a ended EPIDEMIOLOGY Add ess o co espondence: Jaana T. Ka i, M.Sc. (Econ), LIKES Re- sea ch Cen e o Physical Ac i i y and Heal h, Rau pohjanka u 8, 40700 Jy a¨skyla¨, Finland; E-mail: jaana.ka[email p o ec ed]. Submi ed o publica ion Feb ua y 2017. Accep ed o publica ion June 2017. Supplemen al digi al con en is a ailable o his a icle. Di ec URL ci a- ions appea in he p in ed ex and a e p o ided in he HTML and PDF e sions o his a icle on he jou nal_s Web si e (www.acsm-msse.o g). 0195-9131/17/4911-2158/0 MEDICINE & SCIENCE IN SPORTS & EXERCISE Ò Copy igh Ó2017 The Au ho (s). Published by Wol e s Kluwe Heal h, Inc. on behal o he Ame ican College o Spo s Medicine. This is an open- access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion- Non Comme cial-No De i a i es License 4.0 (CCBY-NC-ND), whe e i is pe missible o download and sha e he wo k p o ided i is p ope ly ci ed. The wo k canno be changed in any way o used comme cially wi hou pe mission om he jou nal. DOI: 10.1249/MSS.0000000000001351 2158 colleges and uni e si ies had highe g adua ion a es han nona hle es, whe eas Ba on e al. (3) ound ha a hle ic pa - icipa ion was associa ed wi h a highe pos seconda y le el o educa ion. Simila ly, P ei e and Co neliQen (23) showed ha nonp o essional physical ac i i y du ing childhood and you h inc eased he p obabili y o a aining a highe school deg ee. Al hough he ela ionship be ween physical ac i i y and educa ional ou comes has been es ablished, some o he esea ch indings a e mixed and sugges he exis ence o unmeasu ed he e ogenei y, selec ion issues, and he po en ial endogenei y o physical ac i i y (3,10,26,34). Addi ionally, a ecen sys- ema ic e iew by Donnelly e al. (9) indica ed ha much o he p e ious esea ch explo ed he ela ionship be ween physical ac i i y and academic achie emen only h ough c oss- sec ional s udies. The e o e, li le is known abou he longi- udinal consequences o leisu e- ime physical ac i i y du ing adolescence o academic achie emen and subsequen edu- ca ional a ainmen among he same indi iduals. The pu pose o he p esen s udy was o examine whe he physical ac i i y in adolescence is associa ed wi h academic achie emen a he end o compulso y basic educa ion and wi h pos compulso y educa ion la e in li e. The s udy aug- men ed he p e ious li e a u e in h ee ways: Fi s , we in- es iga ed whe he he physical ac i i y le el a age 15 y o a change in he physical ac i i y le el be ween he ages o 12 and 15 y a e associa ed wi h he GPA a age 15 y and he yea s o pos compulso y educa ion in adul hood. Sec- ond, we examined he ole o pa icipa ion in o ganized spo s in academic achie emen and educa ional a ainmen . Thi d, we used an ins umen al a iable (IV) app oach o deal wi h he po en ial endogenei y o physical ac i i y, omi ed a iable bias, and measu emen e o . We hypo he- sized ha a highe physical ac i i y le el and an inc ease in he physical ac i i y le el in adolescence a e associa ed wi h highe educa ional ou comes. METHODS S udy popula ion. The da a we e d awn om he on- going longi udinal Ca dio ascula Risk in Young Finns S udy (YFS), which was launched in 1980, when 3596 indi iduals age 3, 6, 9, 12, 15, and 18 y pa icipa ed in he baseline s udy. The pa icipan s we e andomly chosen boys (51%) and gi ls (49%) om i e Finnish uni e si y ci ies wi h medical schools (Helsinki, Tu ku, Tampe e, Oulu, and Kuopio) and hei u al su oundings. Since 1980, se en ollow-up phases (1983, 1986, 1989, 1992, 2001, 2007, and 2011) ha e been conduc ed. The examina ions ha e included comp e- hensi e da a collec ion using blood es s, physical measu e- men s, and ques ionnai es (24). To ob ain egis e -based in o ma ion on educa ional a ain- men , he YFS da a we e linked wi h he egis e -based Finnish Longi udinal Employe –Employee Da a (FLEED) om S a- is ics Finland. Finnish Longi udinal Employe –Employee Da a is an annual panel ha includes in o ma ion on indi iduals_educa ional a ainmen ob ained di ec ly om he comp ehensi e adminis a i e egis e s main ained by S a is- ics Finland. Regis e -based in o ma ion on amily backg ound ac o s, including amily income and pa en s_educa ion, was ob ained om he Longi udinal Popula ion Census (LPC) o S a is ics Finland om 1980. To a oid p oblems c ea ed by e o s in eco d linkages (27), FLEED and LPC da a we e combined wi h he YFS da a based on he unique pe sonal iden i ie s. The p esen s udy consis ed o indi iduals who we e age 12 o 15 y when he in o ma ion abou physical ac i i y was collec ed; ha is, he sample consis ed o 2445 indi iduals (48% boys) who we e bo n be ween 1965 and 1977. The e- a e , he indi iduals we e ollowed un il 2010, when he pa icipan s_mean age was 40 y . Depending on he model speci ica ion, he sample size a ied om 1126 o 1628. Da a p o ec ion issues we e aken in o accoun as speci ied in cu en Finnish legisla ion. Mo eo e , he esea ch p o ocol o he YFS was app o ed by he e hics commi ees o he i e uni e si ies (Helsinki, Tu ku, Tampe e, Oulu, and Kuopio), and each pa icipan p o ided w i en in o med consen be o e pa icipa ing in he s udy. Sel - epo ed physical ac i i y. Leisu e- ime physi- cal ac i i y a he ages o 12 and 15 y was measu ed wi h a sel - epo ed ques ionnai e. The pa icipan s we e asked he ollowing ques ions: ‘‘How o en do you engage in leisu e- ime physical ac i i y o a leas hal an hou pe session?,’’ ‘‘How much a e you b ea h- aking and swea ing when you engage in physical ac i i y and spo s?,’’ ‘‘How many imes a week do you usually engage in aining sessions o a spo s club?,’’ ‘‘Wha do you usually do in you leisu e ime?,’’ and ‘‘Do you pa icipa e in egional- o na ional-le el spo s compe i ions?.’’ The esponse al e na i es we e coded om 1 o 3—excep pa icipa ion in spo s compe i ions, which was coded as 1 o 2—and he ea e summed o o m a physical ac i i y index (PAI) wi h sco es anging om 5 o 14 (Table, Supplemen al Digi al Con en 1, The o iginal sco ing and ecoding o PAI; 37, h p://links.lww.com/MSS/A969). Educa ional ou comes. Two measu es we e used o illus a e indi iduals_educa ional ou comes: sel - epo ed GPA a age 15 y and yea s o comple ed pos compulso y educa ion. The in o ma ion on GPA was p o ided by he YFS; he a iable e e ed o a nume ical assessmen on a scale o 4 o 10, whe e 4 deno es ailed, and 10 deno es excellen . In o ma ion on pos compulso y educa ional a - ainmen was p o ided by he egis e -based FLEED. Using o icial es ima es om S a is ics Finland o comple ing a speci ic deg ee, he yea s o educa ion we e o mula ed by ans o ming he highes ob ained deg ee in 2010 in o yea s o educa ion. S a is ical analyses. The analysis was based on o di- na y leas squa es (OLS) models s a i ied by sex. The models used he GPA a age 15 y and he yea s o comple ed edu- ca ion as he dependen a iables and he PAI a age 15 y as an explana o y a iable. All models we e adjus ed by bi h coho and bi h mon h, which we e p ede e mined a iables (2). Be o e unning, he OLS models, he co ela ion coe icien s PHYSICAL ACTIVITY AND EDUCATIONAL OUTCOMES Medicine & Science in Spo s & Exe cise d 2159 EPIDEMIOLOGY we e calcula ed o illus a e he uncondi ional connec ions be ween adolescen physical ac i i y and educa ional ou - comes (Table, Supplemen al Digi al Con en 2, Co ela ion coe icien s, h p://links.lww.com/MSS/A970). The baseline models we e ex ended in he ollowing ways. Fi s , o educe he possibili y ha physical ac i i y e lec s omi ed a iables wi hou ha ing an independen e ec on educa ional ou comes, he models we e augmen ed wi h se e al co a ia es: 1) indi iduals_ch onic condi ions and body a in 1980 (32); 2) amily backg ound ac o s in 1980, including egis e -based amily income, egis e -based pa en s_edu- ca ion, and amily size; and 3) indi idual_s p io academic achie emen measu ed wi h he sel - epo ed GPA a he age o 12 y (see Table, Supplemen al Digi al Con en 3, Desc ip i e s a is ics wi h co a ia es, o addi ional de ails, h p://links.lww.com/MSS/A971). Typically, du ing childhood and you h, pa icipa ion in spo s dec eases wi h age (15,21). An open ques ion is whe he his change is ela ed o academic achie emen and educa- ional a ainmen . To ind an answe , he associa ion be ween educa ional ou comes and he change in he physical ac i i y le el be ween ages 12 and 15 y was examined. Addi ionally, wo bina y a iables we e o mula ed, which equals 1, i he physical ac i i y le el inc eased o dec eased be ween he wo ime poin s, and 0 o he wise. Almos hal o he Finnish child en age 7–18 y epo pa icipa ion in spo s club ac i i ies a leas once a week (36). To examine whe he pa icipa ion in spo s club ain- ing sessions o pa icipa ion in spo s compe i ions ha e an independen ole in educa ional ou comes, he o iginal PAI was di ided in o h ee pa s: 1) a a iable ha indica ed pa icipa ion in spo s club aining sessions (1, pa icipa es in spo s club aining sessions; 0, does no pa icipa e in spo s club aining sessions), 2) a a iable ha indica ed pa ici- pa ion in spo s compe i ions (1, pa icipa es in na ional- o egional-le el spo s compe i ions; 0, does no pa icipa e in spo s compe i ions), and 3) he o iginal PAI sub ac ed om he pa icipa ion in spo s club aining sessions and pa ici- pa ion in spo s compe i ions a iables. The o iginal PAI was eplaced wi h hese h ee new a iables in he equa ion, and he same co a ia es we e used as in he baseline OLS models. To alle ia e he possible endogenei y o physical ac i i y, omi ed a iable bias, and measu emen e o , an IV ap- p oach was used. The app oach equi es ins umen s ha a e co ela ed wi h he included endogenous eg esso and unco ela ed wi h he e o e m o he explana o y equa ion (2,35). The e o e, he selec ed ins umen s should a ec edu- ca ional ou comes only h ough hei e ec s on physical ac- i i y (5). Following P ei e and Co neliQen (23), Rees and Sabia (26), and Eide and Ronan (10), we used he indi idual_s body heigh as an ins umen . Following a ecen s udy by Cabane e al. (6) in which pa en al a is ic ac i i ies we e used as an ins umen o examine he ole o playing music o spo s du ing adolescence in educa ion and heal h ou comes, we also included pa en s_physical ac i i y de ails as an ins umen . As sugges ed by Cabane e al. (6), pa en al a is ic ac i i ies a e a s ong p edic o o an adolescen _s pa icipa ion in music and a e unlikely o in luence he ou come di ec ly. Because he IV es ima es can di e sys ema ically om he a ge pa ame e i he sample size is small (2), he IV analysis was conduc ed using a pooled sample o women and men, and he models we e adjus ed by sex, bi h coho , and bi h mon h. RESULTS Baseline esul s. Rega dless o he used measu es o physical ac i i y and spo s pa icipa ion, a he ages o 12 and 15 y , boys had highe alues han gi ls (PG0.001; Table 1). The a e age PAI was app oxima ely one uni highe among boys compa ed wi h gi ls. Rela i e o gi ls, boys mo e TABLE 1. Sample cha ac e is ics. All Women Men nMean (SD) nMean (SD) nMean (SD) P a Physical ac i i y in adolescence PAI a 12 y ( ange, 5–14 y ) b 1723 9.57 (1.80) 886 9.06 (1.63) 837 10.11 (1.82) PG0.001 PAI a 15 y ( ange, 5–14 y ) 2445 8.97 (1.97) 1271 8.61 (1.77) 1174 9.37 (2.10) PG0.001 $PAI a 12–15 y c 1723 j0.52 (1.84) 886 j0.36 (1.80) 837 j0.70 (1.88) PG0.001 +$PAI a 12–15 y d 1723 0.28 (0.45) 886 0.31 (0.46) 837 0.23 (0.42) PG0.001 j$PAI a 12–15 y e 1723 0.50 (0.50) 886 0.46 (0.50) 837 0.53 (0.50) PG0.004 Pa icipa ion in spo s club aining a 15 y 2445 0.28 (0.45) 1271 0.24 (0.43) 1174 0.32 (0.47) PG0.001 Pa icipa ion in spo s compe i ions a 15 y g 1854 0.24 (0.43) 981 0.19 (0.39) 873 0.30 (0.46) PG0.001 Academic achie emen a he age o 15 GPA a 15 y h ( ange 4–10) 1941 7.82 (0.93) 1022 8.08 (0.88) 919 7.53 (0.90) PG0.001 Educa ional a ainmen a he age o 33–45 Yea s o educa ion i 2445 13.53 (2.88) 1271 13.88 (2.82) 1174 13.23 (2.86) PG0.001 a P alues o sex di e ences ( - es ). b PAI (minimum 5 o maximum, 14) a he age o 12 and 15 y is a summa y o i e a iables ha illus a e he equency and in ensi y o leisu e- ime physical ac i i y, equency o pa icipa ion in spo s clubs aining sessions, pa icipa ion in spo s compe i ions, and he mos common ac i i y du ing leisu e ime. c $PAI a 12–15 y illus a es he change in physical ac i i y le el (PAI) be ween he ages o 12 and 15 y . d +$PAI a 12–15 y is a bina y a iable ha ge s alue 1 i he indi idual has inc eased his o he physical ac i i y le el be ween he ages o 12 and 15 y , and 0 o he wise. e j$PAI a 12–15 y is a bina y a iable ha ge s alue 1 i he indi idual has dec eased his o he physical ac i i y le el be ween he ages o 12 and 15 y , and 0 o he wise. Pa icipa ion in spo s club aining sessions a he age o 15 is a bina y a iable ha ge s alue 1 i an indi idual pa icipa es in o ganized spo s clubs aining sessions, and 0 o he wise. g Pa icipa ion in spo s compe i ions a he age o 15 is a bina y a iable ha ge s alue 1 i an indi idual pa icipa es in na ional- o egional-le el spo s compe i ions, and 0 o he wise. The sample size a ies because in he 1992 ollow-up pa icipa ion in spo s compe i ions was no asked. h GPA a he age o 15 y e e s o a sel - epo ed nume ical assessmen on a scale o 4–10, whe e 4 is ailed and 10 is excellen . i Regis e -based in o ma ion on educa ional a ainmen in adul hood based on he yea s o comple ed educa ion. h p://www.acsm-msse.o g2160 O icial Jou nal o he Ame ican College o Spo s Medicine EPIDEMIOLOGY commonly pa icipa ed in spo s club aining sessions and spo s compe i ions a he age o 15 (PG0.001; Table 1). Be ween he ages o 12 and 15 y , howe e , boys_physical ac i i y le els dec eased mo e han hose o gi ls (PG0.001; Table 1). In e ms o educa ional ou comes, he GPA a age 15 y was app oxima ely 0.6 uni s highe among gi ls compa ed wi h boys (PG0.001; Table 1). Simila ly, he le el o pos compulso y educa ion was highe among women han men; ha is, women had app oxima ely 0.7 mo e yea s o educa ion hanmendu ing he ollow-uppe iod(PG 0.001; Table 1). The baseline OLS es ima es (Table 2, column 1) sug- ges ed ha a highe PAI a age 15 y was associa ed wi h a highe GPA a age 15 y and a highe le el o educa ion in la e li e. Among women, a one-uni inc ease in he PAI was ela ed o a 0.11 uni inc ease in GPA and 0.17 addi ional yea s o educa ion (Table 2, column 1). Among men, he coe icien s we e 0.06 o GPA and 0.17 o yea s o edu- ca ion. As a ough es ima e, a he end o compulso y basic educa ion, child en wi h he highes physical ac i i y le el (PAI = 14 a he age o 15 y ) had a GPA ha was app oxi- ma ely 1.0 uni highe among gi ls and 0.5 uni highe among boys compa ed wi h hose wi h he lowes physical ac i i y le el (PAI = 5 a he age o 15 y ). In e ms o educa ional a ainmen , he mos physically ac i e gi ls and boys (PAI = 14 a he age o 15 y ) had app oxima ely 1.5 mo e yea s o educa ion in adul hood compa ed wi h hei leas physically ac i e coun e pa s (PAI = 5 a he age o 15 y ). Robus ness checks. Among men, he inclusion o a comp ehensi e se o po en ial con ounding ac o s (Table 2, columns 2–4) kep he physical ac i i y es ima e s a is ically signi ican bu dec eased he poin es ima es. In he ex ended model (Table 2, column 4), a one-uni inc ease in he PAI was ela ed o a 0.03 uni inc ease in GPA and 0.09 addi ional yea s o educa ion. Among women, when he models we e adjus ed by indi iduals_p io academic achie emen (Table 2, column 4), in he case o educa ional a ainmen , he associa ion emained posi i e bu became s a is ically insigni ican . The coe icien o GPA, howe e , emained s a is ically signi ican , indica ing a posi i e asso- cia ion be ween physical ac i i y and academic achie emen . The R 2 o he ex ended models (Table 2, column 4) a ied om 0.28 o 0.67. In he case o GPA, app oxima ely 70% o he a iance was accoun ed by he model, whe eas he pe - cen age o educa ional a ainmen was app oxima ely 30%. The esul s om he baseline OLS models we e quali a- i ely and quan i a i ely simila o women and men. The e o e, om Table 3 onwa d, he esul s a e p esen ed o women and men combined and he models we e adjus ed by sex. In line wi h he baseline OLS esul s, change in he physical ac i i y le el in adolescence ($PAI a 12–15 y ) was posi i ely asso- cia ed wi h academic achie emen a age 15 y and educa ional a ainmen in la e li e (Table 3, panel 1). When he models we e adjus ed by indi idual cha ac e is ics, amily back- g ound ac o s, and p io academic achie emen (Table 3, panel 1, columns 2–4), he poin es ima e dec eased by hal while emaining posi i e and s a is ically signi ican . In he ex ended models, he co esponding coe icien s we e ap- p oxima ely 0.05 (academic achie emen ) and 0.09 (edu- ca ional a ainmen ). When ocusing on hose whose physical ac i i y le els inc eased be ween he ages o 12 and 15 y (+$PAI a 12–15 y ), he esul s implied a posi i e associa- ion be ween he a iables (Table 3, panel 2). Fo academic achie emen , he poin es ima es o physical ac i i y we e app oxima ely 0.3 (Table 3, panel 2, column 1) and 0.16 in he ex ended model (Table 3, panel 2, column 4). Fo educa ional a ainmen , he co esponding poin es ima es we e app oxi- ma ely 0.7 (Table 3, panel 2, column 1) and 0.3 (Table 3, TABLE 2. OLS es ima es o he ela ionship be ween physical ac i i y le el a he age o 15 y , academic achie emen a he age o 15 y , and educa ional a ainmen in adul hood. Academic Achie emen a he Age o 15 y , GPA Educa ional A ainmen a he Age o 33–45 y , Yea s o Educa ion 12341234 Panel 1: women PAI a 15 y a 0.108*** 0.108*** 0.083*** 0.048*** 0.174*** 0.174*** 0.103** 0.028 (0.019) (0.019) (0.018) (0.011) (0.050) (0.050) (0.048) (0.043) N661 661 661 661 850 850 850 850 R 2 0.11 0.11 0.19 0.67 0.05 0.06 0.13 0.28 Panel 2: men PAI a 15 y 0.060*** 0.062*** 0.054*** 0.032** 0.166*** 0.162** 0.139** 0.092* (0.018) (0.018) (0.017) (0.010) (0.051) (0.052) (0.049) (0.043) N586 586 586 586 778 778 778 778 R 2 0.05 0.05 0.19 0.66 0.05 0.05 0.16 0.32 Con ol a iables Bi h coho and bi h mon h b xxxxxxxx Indi idual cha ac e is ics c xxx xxx Family backg ound d xx xx GPA a 12 y e xx Robus s anda d e o s a e in pa en heses. The le el o s a is ical signi icance: ***P G0.001, **P G0.01, *P G0.05. a PAI (minimum, 5 o maximum, 14) a he age o 15 y is a summa y o i e a iables ha illus a e he equency and in ensi y o leisu e- ime physical ac i i y, equency o pa icipa ion in spo s clubs aining sessions, pa icipa ion in spo s compe i ions, and he mos common ac i i y du ing leisu e ime. b All models include con ols o coho (1–5) and bi h mon h. The coho dummies illus a e he yea o bi h: coho 1, bo n in 1977; coho 2, bo n in 1974; coho 3, bo n in 1971; coho 4, bo n in 1968; and coho 5, bo n in 1965. c Indi idual cha ac e is ics: summa y o an indi idual_s ch onic condi ions and body a . d Family backg ound ac o s: amily income, pa en s_educa ion, amily size. e GPA a he age o 12 y e e s o a sel - epo ed nume ical assessmen on a scale o 4–10 (4 is ailing and 10 is excellen ). PHYSICAL ACTIVITY AND EDUCATIONAL OUTCOMES Medicine & Science in Spo s & Exe cise d 2161 EPIDEMIOLOGY panel 2, column 4). Fo hose, whose physical ac i i y le els dec eased be ween he ages o 12 and 15 y (j$PAI a 12– 15 y ), he associa ions we e nega i e in all model speci i- ca ions (Table 3, panel 2, columns 1–4). The models ha examined whe he pa icipa ion in o ga- nized spo s a age 15 y is ela ed o educa ional ou comes a e epo ed in Table 4. In he case o academic achie emen (Table 4, le -hand side), he associa ion was posi i e o all measu es o physical ac i i y: o e all physical ac i i y le el, pa icipa ion in spo s club aining sessions, and pa icipa ion in spo s compe i ions. The inclusion o a comp ehensi e se o po en ial con ounding ac o s (Table 4, le -hand side, columns 2–4) kep he es ima es s a is ically signi ican . In he case o educa ional a ainmen (Table 4, igh -hand side), howe e , no such clea ela ion was obse ed. In he baseline model (Table 4, igh -hand side, column 1), he associa ion depended on he measu emen ype o physical ac i i y. Mo e p ecisely, pa icipa ion in spo s club aining sessions was posi i ely ela ed o educa ional a ainmen (Table 4, igh - hand side, panel 1, column 1), whe eas he associa ion be ween pa icipa ion in spo s compe i ions and educa ional a ainmen was s a is ically insigni ican (Table 4, igh -hand side, panel 2, column 1). In he ex ended model (Table 4, igh -hand side, column 4), ega dless o he measu es o physical ac i i y, he poin es ima es became s a is ically insigni ican . The IV esul s a e epo ed in Table 5. Rega dless o he ins umen , he IV es ima es we e posi i e and s a is ically signi ican , sugges ing ha adolescen physical ac i i y is pos- i i ely ela ed o educa ional ou comes. When he indi idual_s body heigh was used as an ins umen (Table 5, columns 2), he Fs a is ics a ied om 4.67 o 6.23, indica ing ha body heigh is only weakly co ela ed wi h physical ac i i y. The e- o e, he IV es ima o may be biased owa d he OLS es ima o , and he esul s should be ea ed wi h cau ion. In he case o pa en s_physical ac i i y (Table 5, columns 3–4), he F alues a ied om 26.85 o 70.19, indica ing ha pa en s_physical ac i i y is highly co ela ed wi h hei child en_s physical ac i - i y. In gene al, he IV es ima es we e g ea e han he co e- sponding OLS es ima es sugges ing an e en s onge associa ion be ween physical ac i i y and educa ional ou comes. DISCUSSION Summa y o he esul s. Using da a om a popula ion- based ollow-up s udy om you h o adul hood combined wi h egis e -based in o ma ion on pos compulso y educa ion, TABLE 3. Changes in physical ac i i y le els be ween ages o 12 and 15 y in associa ion wi h academic achie emen a he age o 15 y and educa ional a ainmen in adul hood. Academic Achie emen a he Age o 15 y , GPA Educa ional A ainmen a he Age o 33–45 y , Yea s o Educa ion 12341234 Panel 1: change in physical ac i i y le el $PAI a 12–15 y a 0.099*** 0.099*** 0.080*** 0.050*** 0.209*** 0.208*** 0.148*** 0.088* (0.015) (0.015) (0.014) (0.009) (0.042) (0.042) (0.040) (0.036) PAI a 12 y b 0.064*** 0.064*** 0.052*** 0.032*** 0.103* 0.101* 0.073 0.022 (0.016) (0.016) (0.016) (0.009) (0.016) (0.047) (0.045) (0.040) N1191 1191 1191 1191 1556 1556 1556 1556 R 2 0.17 0.17 0.26 0.70 0.06 0.06 0.16 0.31 Panel 2: inc ease in physical ac i i y le el +$PAI a 12–15 y c 0.297*** 0.297*** 0.229*** 0.156*** 0.678*** 0.675*** 0.474** 0.326* (0.059) (0.059) (0.055) (0.034) (0.167) (0.167) (0.156) (0.141) PAI a 12 y 0.037* 0.037* 0.030* 0.019* 0.055 0.054 0.039 0.005 (0.016) (0.016) (0.015) (0.009) (0.044) (0.044) (0.042) (0.038) N1191 1191 1191 1191 1556 1556 1556 1556 R 2 0.15 0.17 0.25 0.70 0.06 0.06 0.16 0.31 Panel 3: dec ease in physical ac i i y le el j$PAI a 12–15 y d j0.263*** j0.264*** j0.201*** j0.123*** j0.696*** j0.696*** j0.537*** j0.362** (0.053) (0.053) (0.051) (0.032) (0.149) (0.149) (0.142) (0.130) PAI a 12 y 0.044** 0.044** 0.035* 0.021* 0.073 0.072 0.055 0.015 (0.016) (0.016) (0.015) (0.009) (0.044) (0.045) (0.042) (0.038) N1191 1191 1191 1191 1556 1556 1556 1556 R 2 0.15 0.15 0.25 0.70 0.06 0.07 0.16 0.31 Con ol a iables Bi h coho and bi h mon h e xxxxxxxx Sex xxxxxxxx Indi idual cha ac e is ics xxx xxx Family backg ound g xx xx GPA a 12y h xx Robus s anda d e o s a e in pa en heses. The le el o s a is ical signi icance: ***P G0.001, **P G0.01, *P G0.05. a $PAI a 12–15 y illus a es he change in physical ac i i y le el (PAI) be ween he ages o 12 and 15 y . b PAI (minimum, 5 o maximum, 14) a he age o 12 y is a summa y o i e a iables ha illus a e he equency and in ensi y o leisu e- ime physical ac i i y, equency o pa icipa ion in spo s clubs aining sessions, pa icipa ion in spo s compe i ions, and he mos common ac i i y du ing leisu e ime. c +$PAI a 12–15 y is a bina y a iable, which ge s alue 1 i he physical ac i i y le el inc eases, and 0 o he wise. d j$PAI a 12–15 y is a bina y a iable, which ge s alue 1 i he physical ac i i y le el dec eases, and 0 o he wise. e All models include con ols o coho (1–5) and bi h mon h. The coho dummies illus a e he yea o bi h: coho 1, bo n in 1977; coho 2, bo n in 1974; coho 3, bo n in 1971; coho 4, bo n in 1968; and coho 5, bo n in 1965. Indi idual cha ac e is ics: summa y o an indi idual_s ch onic condi ions and body a . g Family backg ound ac o s: amily income, pa en s_educa ion, amily size. h GPA a he age o 12 e e s o a sel - epo ed nume ical assessmen on a scale o 4–10 (4 is ailing and 10 is excellen ). h p://www.acsm-msse.o g2162 O icial Jou nal o he Ame ican College o Spo s Medicine EPIDEMIOLOGY he aim o his s udy was o examine whe he adolescen physical ac i i y is associa ed wi h academic achie emen a he end o compulso y basic educa ion and wi h educa ional a ainmen la e in li e. The esul s p o ided e idence ha physical ac i i y in adolescence was posi i ely associa ed wi h educa ional ou comes: The physical ac i i y le el a age 15 y and an inc ease in he physical ac i i y le el be ween he ages o 12 and 15 y we e posi i ely ela ed o GPA a age 15 y and yea s o subsequen educa ion in adul hood. The esul s we e obus o he inclusion o addi ional co a ia es, including in- di idual and amily backg ound ac o s. Physical ac i i y and educa ional ou comes. Ou indings a e consis en wi h p e ious li e a u e showing ha physical ac i i y is posi i ely ela ed o academic achie e- men (11,16,17,31,33) and educa ional a ainmen (3,19,23). Howe e , he ques ions conce ning physical ac i i y ha e been s udy speci ic, and he e o e, a compa ison o he esul s would no be s aigh o wa d. Mo eo e , o ou knowledge, he longi- udinal associa ions be ween you h physical ac i i y and edu- ca ional ou comes a e a less s udied. In e ms o ca dio ascula i ness, Swedish s udy demons a ed ha he change in ca dio- ascula i ness be ween he ages o 15 and 18 y p edic s cog- ni i e pe o mance a age 18 y , and ha ca dio ascula i ness a age 18 y p edic s highe subsequen educa ion (1). The possible sel -selec ion in o physical ac i i ies is an issue ha has o be discussed while in e p e ing he esul s. As S e enson (34) sugges s, subs an ial sel -selec ion in o physical ac i i ies may exis , and wi hou adequa e co a ia es, he impac o physical ac i i y on educa ional ou comes can be upwa dly biased. Fu he mo e, indi iduals_academic achie emen and educa ional a ainmen can be highly co e- la ed wi h hei pa en s_educa ion le el and socioeconomic s a us. Thus, he baseline OLS models we e augmen ed wi h indi iduals_ch onic condi ions, body a , p io academic TABLE 5. IV es ima es o he ela ionship be ween PAI a he age o 15 y and educa ional ou comes. Academic Achie emen a he Age o 15 y , GPA Educa ional A ainmen a he Age o 33–45 y , Yea s o Educa ion 123 4 123 4 PAI a 15 y 0.081*** (0.013) 0.966* (0.476) 0.285*** (0.063) 0.277*** (0.061) 0.169*** (0.036) 2.501* (1.086) 0.850*** (0.192) 0.833*** (0.186) Fi s -s age F — 4.67 52.06 26.85 — 6.23 70.19 36.07 N1247 1247 1247 1247 1628 1628 1628 1628 Robus s anda d e o s a e in pa en heses. The le el o s a is ical signi icance: ***P G0.001, *P G0.05. All models include con ols o coho (1–5), bi h mon h and sex. The coho dummies illus a e he yea o bi h: Coho 1 = bo n in 1977, Coho 2 = bo n in 1974, Coho 3 = bo n in 1971, Coho 4 = bo n in 1968, and Coho 5 = bo n in 1965. 1, Baseline OLS. Ins umen s used: 2, Heigh . 3, Physical ac i i y, a he . 4, Physical ac i i y, a he and mo he . TABLE 4. Pa icipa ion in spo s club aining sessions and pa icipa ion spo s compe i ions a 15 y in associa ion wi h academic achie emen a he age o 15 y and educa ional a ainmen in adul hood. Academic Achie emen a he Age o 15 y , GPA Educa ional A ainmen a he Age o 33–45 y , Yea s o Educa ion 12341234 Panel 1: pa icipa ion in spo s club aining sessions O e all physical ac i i y a 15 y a 0.078*** (0.024) 0.078*** (0.024) 0.067** (0.023) 0.052*** (0.014) 0.128 (0.075) 0.124 (0.075) 0.095 (0.072) 0.078 (0.065) Pa icipa ion in spo s club a 15 y b 0.210*** (0.061) 0.210** (0.061) 0.155** (0.058) 0.088* (0.037) 0.445* (0.209) 0.452* (0.210) 0.265 (0.202) 0.116 (0.186) N1126 1126 1126 1126 1168 1168 1168 1168 R 2 0.15 0.15 0.25 0.69 0.06 0.06 0.14 0.30 Panel 2: pa icipa ion in spo s compe i ions O e all physical ac i i y a 15 y a 0.086*** (0.023) 0.086*** (0.023) 0.066** (0.022) 0.056*** (0.014) 0.161* (0.072) 0.158* (0.072) 0.101 (0.070) 0.089 (0.063) Pa icipa ion in spo s compe i ions a 15 y c 0.186** (0.060) 0.185** (0.061) 0.186*** (0.057) 0.077* (0.036) 0.274 (0.202) 0.276 (0.203) 0.256 (0.196) 0.049 (0.181) N1126 1126 1126 1126 1168 1168 1168 1168 R 2 0.15 0.15 0.25 0.69 0.06 0.06 0.14 0.30 Con ol a iables Bi h coho and bi h mon h d x x x xxxxx Sex x x x xxxxx Indi idual cha ac e is ics e xx x xxx Family backg ound xx xx GPA 12y g xx Robus s anda d e o s a e in pa en heses. The le el o s a is ical signi icance: ***P G0.001, **P G0.01, *P G0.05. a The o iginal PAI sub ac ed om he pa icipa ion in spo s compe i ions and pa icipa ion in spo s club aining session a iables. b A a iable ha indica ed pa icipa ion in spo s club aining sessions (1, pa icipa es in spo club aining sessions; 0, does no pa icipa e in spo s club aining sessions). c A a iable ha indica ed pa icipa ion in spo s compe i ions (1, pa icipa es in na ional- o egional-le el spo s compe i ions; 0, does no pa icipa e in spo s compe i ions). d All models include con ols o coho (1–5) and bi h mon h. The coho dummies illus a e he yea o bi h: coho 1, bo n in 1977; coho 2, bo n in 1974; coho 3, bo n in 1971; coho 4, bo n in 1968; and coho 5, bo n in 1965. e Indi idual cha ac e is ics: summa y o an indi idual_s ch onic condi ions and body a . Family backg ound ac o s: amily income, pa en s_educa ion, amily size. g GPA a he age o 12 y e e s o a sel - epo ed nume ical assessmen on a scale o 4–10 (4 is ailing and 10 is excellen ). PHYSICAL ACTIVITY AND EDUCATIONAL OUTCOMES Medicine & Science in Spo s & Exe cise d 2163 EPIDEMIOLOGY achie emen , pa en s_educa ion, amily size, and amily in- come. In line wi h he indings o Ba on e al. (3), he esul s suppo ed he iew ha di e ences ac oss indi iduals_abili- ies, especially p io academic achie emen , pa ially explain he co ela ion be ween physical ac i i y and educa ional ou - comes: When he models we e adjus ed by indi iduals_p io academic achie emen , in all model speci ica ions, he s eng h o he associa ion dec eased. This esul implies ha he con- nec ion be ween physical ac i i y and educa ion may be pa ly a ibu ed o he highe GPA o physically ac i e adolescen s. The e o e, i emains unclea whe he he educa ional bene i s o adolescen physical ac i i y a e caused by pa icipa ion in physical ac i i y ( ea men e ec ) o a e me ely ela ed o he ypes o indi iduals who p e e o be physically ac i e du ing hei leisu e- ime ( he selec ion e ec ). Changes in physical ac i i y le el and educa- ional ou comes. Pa icipa ion in spo s du ing childhood and adolescence ypically dec eases wi h age (15,21). Acco ding o Nade e al. (21), a he age o 9 y , almos all child en me he ecommended 60 min o mode a e- o igo ous-in ensi y physical ac i i y daily, whe eas a he age o 15 y , only 30% achie ed he benchma k. A simila decline was seen in ou da a. An open ques ion was whe he his change is ela ed o academic achie emen and educa ional a ainmen . As an answe , a posi i e associa ion be ween he change in physical ac i i y le el and educa ional ou comes was ound. One no- able ea u e in he esul s was ha in he case o academic achie emen , he physical ac i i y le el a age 12 y as well as he change in physical ac i i y be ween he ages o 12 and 15 y we e associa ed wi h highe GPA a age 15 y . Re- ga ding educa ional a ainmen , howe e , only he change, no he le el o physical ac i i y le el, seemed o ma e . Mo e speci ically, an inc ease in he physical ac i i y le el was associa ed wi h highe educa ional e u ns in adul hood, whe eas he associa ion wi h espec o he dec ease in physical ac i i y le el was nega i e. F om policy-pe spec i e, his ind- ing could encou age de eloping p og ams and in e en ions a ge ed o school-age child en aiming o os e child en_s pa icipa ion in physical ac i i y. This could push he young owa d mo e physically ac i e li es yles and imp o e hei educa ional a ainmen in la e li e, p o iding bo h pe sonal and socie al bene i s (4,8). Pa icipa ion in o ganized spo s and educa ional ou comes. The physical ac i i y a iable PAI used in he s udy was a summa y o i e a iables. In addi ion o he equency o weekly physical ac i i y, he PAI included in- o ma ion abou he indi iduals_pa icipa ion in spo s club aining sessions and pa icipa ion in spo s compe i ions (37). As p io analyses ha e shown, high school a hle ic pa ici- pa ion, pa icipa ion in o ganized spo s, and pa icipa ion in spo s compe i ions (23,26) a e ela ed o educa ional ou - comes. Mo eo e , almos hal o Finnish child en and ado- lescen s pa icipa e in o ganized spo s (36). The e o e, as a obus ness check, we examined he ole o pa icipa ion in spo s club aining sessions and spo s compe i ions in ed- uca ional ou comes. In gene al, he esul s we e in line wi h he baseline OLS esul s, sugges ing a posi i e associa ion be ween physical ac i i y and educa ional ou comes. How- e e , some ea u es equi ed a en ion. Fi s , i espec i e o he di e en measu es o physical ac i i y, each physical ac i i y a iable was posi i ely ela ed o academic achie e- men . Second, in he case o educa ional a ainmen , no such clea associa ion was obse ed. Ins ead, when he models we e adjus ed by indi idual and amily backg ound ac o s, he poin es ima e became s a is ically insigni ican . The e o e, he esul s suppo he iew ha in he case o pos compul- so y educa ion, wha ma e s is he o e all physical ac i i y le el exp essed as he PAI, no he single measu es o o - ganized spo s. S eng hs and limi a ions. The linked da a and lon- gi udinal se ing o he s udy con ibu e o and ex end he p e ious li e a u e in ou impo an ways. Fi s , he longi- udinal se ing allowed acking he same indi iduals_edu- ca ional ou comes s a ing om compulso y basic educa ion un il he end o pos compulso y g adua ion. Second, ins ead o a single measu e o physical ac i i y, he s udy included ou measu es o illus a e physical ac i i y: leisu e- ime physical ac i i y le el a age 15 y , change in he physical ac i i y le el be ween he ages o 12 and 15 y , pa icipa ion in spo s club aining sessions, and pa icipa ion in spo s compe i ions. Howe e , physical ac i i y de ails we e sel - epo ed; hus, some measu emen e o may exis (29,30). Thi d, ins ead o sel - epo ed in o ma ion on pos compulso y educa ion, his s udy used he yea s o comple ed educa ion ob ained om S a is ics Finland. This mi iga ed he possible measu emen e o in he ou come a iable. Fou h, o alle ia e possible endogenei y o physical ac i i y, omi ed a iable bias, and measu emen e o , an IV app oach was used. Following P ei e and Co neliQen (23), Rees and Sabia (26), and Eide and Ronan (10), indi idual_s body heigh was used as an in- s umen . The da a also allowed o ackle pa en s_physical ac i i y de ails as an ins umen o explain a ia ions in hei child en_s spo s pa icipa ion. Fu u e di ec ions. The esul p o ided e idence ha adolescen physical ac i i y is posi i ely ela ed o educa ional ou comes du ing he li e cou se, bu he explana ion o his associa ion is unclea . The e o e, u u e esea ch should ex- plo e he po en ial mechanisms behind he indings o his s udy. To unco e hese mechanisms, u u e s udies need in- o ma i e da a se s co e ing, o example, heal h endow- men s, cogni i e abili y, pe sonali y, and amily backg ound ac o s (e.g., pa en al in ol emen in you h spo s, amily suppo ). Addi ionally, mo e esea ch is needed o cla i y he longi udinal ela ionship be ween physical ac i i y and edu- ca ional ou comes. Fo now, we a e pe suaded o hink ha he indings o his s udy a e gene alizable o o he de eloped coun ies wi h simila physical ac i i y beha io as well as educa ional a ainmen . Mo e speci ically, as in many o he coun ies (13,38,39), he physical ac i i y le els a e low among Finnish child en (36). Rega ding educa ional a - ainmen , he c oss-coun y compa isons, e eal ha , in e ms o educa ional le el, Finns age 25–64 y a e anked h p://www.acsm-msse.o g2164 O icial Jou nal o he Ame ican College o Spo s Medicine EPIDEMIOLOGY se en h among OECD (The O ganiza ion o Economic Co-ope a ion and De elopmen ) coun ies (22). To u he inc ease unde s anding o his longi udinal associa ion be- ween physical ac i i y and educa ional ou comes, i would be aluable o eplica e his s udy wi h da a d awn om o he coun ies han Finland. CONCLUSIONS Educa ional a ainmen has been used as a measu e o he skills a ailable in he popula ion and he labo o ce (4,22). Apa om he well-known heal h e ec s o physical ac i i y (18,20,39), he posi i e associa ion ound in he p esen s udy p o ides e idence ha adolescen physical ac i i y may p edic be e lea ning ou comes s a ing om compulso y basic educa ion and con inuing un il pos compulso y edu- ca ion in adul hood. This s udy was unded by he Finnish Minis y o Educa ion and Cul u e, Juho Vainio Founda ion, Y jo¨ Jahnsson Founda ion (g an s 6648, 6914), he Academy o Finland (g an 273971), and OP G oup Resea ch Founda ion. The YFS has been inancially suppo ed by he Academy o Finland: g an s 134309 (Eye), 126925, 121584, 124282, 129378 (Sal e), 117797 (Gendi), and 41071 (Skidi), he Social Insu - ance Ins i u ion o Finland, Kuopio, Tampe e and Tu ku Uni e si y Hospi al Medical Funds, Juho Vainio Founda ion, Sig id Juselius Founda ion, Paa o Nu mi Founda ion, Finnish Founda ion o Ca dio as- cula Resea ch and Finnish Cul u al Founda ion, Tampe e Tube culosis Founda ion and Emil Aal onen Founda ion. The au ho s decla e no con lic s o in e es . The esul s o he p esen s udy do no cons i u e endo semen by he Ame ican College o Spo s Medicine. The au ho s decla e ha he esul s o he s udy a e p esen ed clea ly, hones ly, and wi hou ab ica ion, alsi ica ion, o inapp o- p ia e da a manipula ion. REFERENCES 1. Åbe g MA, Pede sen NL, To en K, e al. Ca dio ascula i ness is associa ed wi h cogni ion in young adul hood. P oc Na l Acad Sci USA. 2009;106(49):20906–11. 2. Ang is JD, Pischke JS. Mos ly Ha mless Econome ics: An Empi icis _s Companion. P ince on (NJ): P ince on Uni e si y P ess; 2009. pp. 68, 117, 205. 3. Ba on JM, Ewing BT, Waddel GR. The e ec s o high school a hle ic pa icipa ion on educa ion and labo ma ke ou comes. Re Econ S a . 2000;82(3):409–21. 4. Becke GS. Human Capi al: A Theo e ical and Empi ical Analy- sis, wi h Special Re e ence o Educa ion. Chicago (IL): Uni e si y o Chicago P ess; 2009. pp. 17–21, 29–54. 5. Bound J, Jaege DA, Bake RM. P oblems wi h ins umen al a - iables es ima ion when he co ela ion be ween he ins umen s and he endogenous explana o y a iables is weak. J Am S a Assoc. 1995;90(430):443–50. 6. Cabane C, Hille A, Lechne M. Moza o Pele ´? The e ec s o adolescen s_pa icipa ion in music and spo s. Labou Econ. 2016; 41:90–103. 7. Ca d D. The causal e ec o educa ion on ea nings. In: Ashen el e OC,Ca dD,eds.Handbook o Labo Economics; 1999;3(A): 1801–63. 8. Cu le DM, Lle as-Muney A. Educa ion and Heal h: E alua ing Theo ies and E idence. (NBER Wo king Pape No. 12352). Camb idge (MA): Na ional Bu eau o Economic Resea ch; 2006. 9. Donnelly JE, Hillman CH, Cas elli D, e al. Physical ac i i y, i - ness, cogni i e unc ion, and academic achie emen in child en: a sys ema ic e iew. Med Sci Spo s Exe c. 2016;48(6):1197–222. 10. Eide ER, Ronan N. Is pa icipa ion in high school a hle ics an in- es men o a consump ion good? E idence om high school and beyond. Econ Educ Re . 2001;20(5):431–42. 11. Fox CK, Ba -Ande son D, Neuma k-Sz aine D, Wall M. Physical ac i i y and spo s eam pa icipa ion: associa ions wi h academic ou comes in middle school and high school s uden s. J Sch Heal h. 2010;80(1):31–7. 12. G ossman M. On he concep o heal h capi al and he demand o heal h. J Poli Econ. 1997;80:223–55. 13. Hallal PC, Ande sen LB, Bull FC, e al. Global physical ac i i y le els: su eillance p og ess, pi alls, and p ospec s. Lance . 2012; 380(9838):247–57. 14. Hanushek EA, Woessmann L. Do be e schools lead o mo e g ow h? Cogni i e skills, economic ou comes, and causa ion. J Econ G ow h. 2012;17(4):267–321. 15. Inchley J, Cu ie D, Young T, e al. G owing up unequal: gende and socioeconomic di e ences in young people_s heal h and well-being. HBSC in e na ional Repo om 2013/2014 Su ey. [In e ne ] 2016. Copenhagen: WHO Regional O ice o Eu ope; [ci ed 2017 Jan 4]. A ailable om h p://www.eu o.who.in /__da a/ asse s/pd _ ile/0003/303438/HSBC-No.7-G owing-up-unequal- Full-Repo .pd ?ua=1. 16. Kan omaa MT, S ama akis E, Kankaanpa¨a¨ A, e al. Physical ac i i y and obesi y media e he associa ion be ween childhood mo o unc ion and adolescen s_academic achie emen . P oc Na l Acad Sci U S A. 2013;110(5):1917–22. 17. K is jnsson AL, Sig sd i ID, Alleg an e JP, Helgason SR. Adoles- cen heal h beha io , con en men in school, and academic achie e- men . Am J Heal h Beha . 2009;33(1):69–79. 18. Lee IM, Shi oma EJ, Lobelo F, Puska P, Blai SN, Ka zma zyk PT. E ec o physical inac i i y on majo non-communicable diseases wo ldwide: an analysis o bu den o disease and li e expec ancy. Lance . 2012;380(9838):219–29. 19. Long JE, Caudill SB. The impac o pa icipa ion in in e colle- gia e a hle ics on income and g adua ion. Re Econ S a . 1991; 73:525–31. 20. Mu ay C, Lopez A. Global mo ali y, disabili y, and he con i- bu ion o isk ac o s: global bu den o disease s udy. Lance . 1997;349(9063):1436–42. 21. Nade PR, B adley RH, Hou s RM, McRi chie SL, O_B ien M. Mode a e- o- igo ous physical ac i i y om ages 9 o 15 yea s. JAMA. 2008;300(3):295–305. 22. O ganiza ion o Economic Coope a ion and De elopmen . Educa- ion a a Glance In e im Repo : Upda e o Employmen and Edu- ca ional A ainmen Indica o s. [In e ne ]. 2015; [ci ed 2017 Jan 4]. A ailable om: h ps://www.oecd.o g/edu/EAG-In e im- epo .pd . 23. P ei e C, Co neliQen T. The impac o pa icipa ion in spo s on educa ional a ainmen – new e idence om Ge many. Econ Educ Re . 2010;29(1):94–103. 24. Rai aka i OT, Juonala M, Ro¨nnemaa T, e al. Coho p o ile: he ca dio ascula isk in Young Finns S udy. In J Epidemiol. 2008; 37(6):1220–6. 25. Rai aka i OT, Taimela S, Po kka KV, e al. Associa ions be ween physical ac i i y and isk ac o s o co ona y hea disease: he Ca dio ascula Risk in Young Finns S udy. Med Sci Spo s Exe c. 1997;29(8):1055–61. 26. Rees DI, Sabia JJ. Spo s pa icipa ion and academic pe o mance: e idence om he Na ional Longi udinal S udy o Adolescen s Heal h. Econ Educ Re . 2010;29(5):751–9. 27. Ridde G, Mo i R. The econome ics o da a combina ion. In: Heckman JJ, Leame EE, eds. Handbook o Econome ics. Vol. 6, Pa B. Ams e dam: Else ie Science; 2007. pp. 5469–547. PHYSICAL ACTIVITY AND EDUCATIONAL OUTCOMES Medicine & Science in Spo s & Exe cise d 2165 EPIDEMIOLOGY 28. Ross CE, Wu CI. The links be ween educa ion and heal h. Am Sociol Re . 1995;60(5):719–45. 29. Sallis JF, Saelens BE. Assessmen o physical ac i i y by sel - epo : s a us, limi a ions, and u u e di ec ions. Res Q Exe c Spo . 2000;71(2 Suppl):S1–14. 30. Shepha d RJ. Limi s o he measu emen o habi ual physical ac- i i y by ques ionnai es. B J Spo s Med. 2003;37(3):197–206. 31. Singh A, Uij dewilligen L, Twisk JW, an Mechelen W, Chinapaw MJ. Physical ac i i y and pe o mance a school: a sys ema ic e- iew o he li e a u e including a me hodological quali y assess- men . A ch Pedia Adolesc Med. 2012;166(1):49–55. 32. Slaugh e MH, Lohman TG, Boileau RA, e al. Skin old equa ions o es ima ion o body a ness in child en and you h. Hum Biol. 1988;60(5):709–23. 33. S e ens TA, To Y, S e enson SJ, Lochbaum MR. The impo - ance o physical ac i i y and physical educa ion in he p edic ion o academic achie emen . J Spo Beha . 2008;31(4):368–88. 34. S e enson B. Beyond he class oom: using Ti le IX o measu e he e u n o high school spo s. Re Econ S a . 2010;92:284–301. 35. S ock JH, Yogo M. Tes ing o weak ins umen s in linea IV eg ession. In: And ews DW, S ock JH, edi o s. Iden i ica ion and In e ence o Econome ic Models: Essays in Hono by Thomas Ro henbe g. New Yo k: Camb idge Uni e si y P ess; 2010. pp. 80–108. 36. Tammelin TH, Ai a A, Hakama¨ki M, e al. Resul s om Finland_s 2016 Repo Ca d on physical ac i i y o child en and you h. J Phys Ac Heal h. 2016;13(11 Suppl 2):S157–64. 37. Telama R, Leskinen E, Yang X. S abili y o habi ual physical ac- i i y and spo pa icipa ion: a longi udinal acking s udy. Scand J Med Sci Spo s. 1996;6:371–8. 38. T emblay MS, Ba nes JD, Gonza ´lez SA, e al. Global Ma ix 2.0: Repo Ca d G ades on he Physical Ac i i y o Child en and You h Compa ing 38 Coun ies. J Phys Ac Heal h. 2016;13(11 Suppl 2): S343–66. 39. Wo ld Heal h O ganiza ion. Global Recommenda ions on Physical Ac i i y o Heal h. [In e ne ]. 2010; [ci ed 2016 Dec 30]. A ailable om: h p://whqlibdoc.who.in /publica ions/2010/9789241599979_ eng.pd ?ua=1. 40. Yang X, Telama R, Hi ensalo M, Ma sson N, Viika i JS, Rai aka i OT. The longi udinal e ec s o physical ac i i y his- o y on me abolic synd ome. Med Sci Spo s Exe c. 2008;40(8): 1424–31. h p://www.acsm-msse.o g2166 O icial Jou nal o he Ame ican College o Spo s Medicine EPIDEMIOLOGY