Nature at home and at work : Naturally good? Links between window views, indoor plants, outdoor activities and employee well-being over one year

Landscape and U ban Planning 160 (2017) 38–47 Con en s lis s a ailable a ScienceDi ec Landscape and U ban Planning jou nalhomepage: www.else ie .com/loca e/landu bplan Na u e a home and a wo k: Na u ally good? Links be ween window iews, indoo plan s, ou doo ac i i ies and employee well-being o e one yea Kale i Ko pela∗, Jessica De Bloom, Ma jaana Sianoja, Ty i Pasanen, Ulla Kinnunen School o Social Sciences and Humani ies/Psychology, 33014 Uni e si y o Tampe e, Tampe e, Finland h i g h l i g h s •We s udied wo-di ec ional associa ions o na u e exposu e wi h employee well-being. •Physical ac i i y in na u e du ing ee ime p edic ed i ali y. •Exposu e o he na u al wo ld a wo k was no linked o he well-being measu es. •Con e sely, a ec i e well-being p edic ed he ype o na u e exposu e. •Physical ac i i y in na u e is a po en ial s a egy o enhancing employee i ali y. a i c l e i n o A icle his o y: Recei ed 1 Feb ua y 2016 Recei ed in e ised o m 22 No embe 2016 Accep ed 13 Decembe 2016 A ailable online 29 Decembe 2016 Keywo ds: Indoo plan s Window iew Physical ac i i y Na u e exposu e Wo kplace Well-being a b s a c We in es iga ed wo-di ec ional ela ions be ween a ious ypes o exposu e o he na u al wo ld, a wo k and a home, and employee well-being. In o al, 841 employees answe ed an elec onic ques ion- nai e wice wi h a one-yea in e al. Pa h analysis indica ed ha equen physical ac i i y in na u al su oundings du ing ee ime p edic ed g ea e i ali y o e a one-yea pe iod a e including con ol a iables. The use o one’s ya d/ga den and happiness we e ma ginally posi i ely associa ed o e ime. None o he a iables in ol ing exposu e o he na u al wo ld a wo k we e linked o he well-being measu es. In he e e se di ec ion, c ea i i y a wo k p edic ed mo e equen and i ali y less equen use o one’s domes ic ga den. Happiness was ma ginally posi i ely ela ed o he equency o physical ac i i y in na u e. The esul s sugges ha ee ime physical ac i i y in na u al su oundings is a po en ial s a egy o enhancing employee i ali y ac oss ime. © 2016 The Au ho s. Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/). 1. In oduc ion Looking a he na u al wo ld om he window can be ega ded as a “mic o- es o a i e” expe ience ha acco ding o he A en- ion Res o a ion Theo y (ART) can ence o us a ion and boos en husiasm abou one´ ıs job hus p omo ing employee well-being (Kaplan, 1993), i.e., op imal psychological unc ioning and expe i- ence (Ryan & Deci, 2001). Being exposed a wo kplaces o window iews o e g eenspace has been ela ed o se e al well-being ou - ∗Co esponding au ho . E-mail add esses: kale i.ko pela@u a.fi (K. Ko pela), jessica.de.bloom@u a.fi (J. De Bloom), ma jaana.sianoja@u a.fi (M. Sianoja), y i.pasanen@u a.fi (T. Pasanen), ulla.kinnunen@u a.fi (U. Kinnunen). comes like lowe physiological a ousal and anxie y (Chang & Chen, 2005) and be e job sa is ac ion (Lo up, S igsdo e , Meilby, & Claudi, 2013) compa ed wi h buil u ban iews. The link be ween na u e exposu e and well-being is e iden also in he psychophysiological s ess eco e y heo y (SRT), ano he majo heo e ical pe spec i e in en i onmen al psychology (Ul ich e al., 1991;B a man, Hamil on, & Daily, 2012). Acco ding o SRT, a isual encoun e wi h na u al scenes p omp s an au oma ic shi owa ds mo e posi i e emo ional s a es, inc eases pa asympa he ic ac i i y inducing elaxa ion (Gladwell e al., 2012), and blocks neg- a i e emo ions and hough s. Complemen ing he idea o he es o a i e e ec s o na u e by ART and SRT, some s udies ha e sugges ed ha exposu e o he na - u al wo ld also has ( e-) i alizing e ec s (Ryan e al., 2010). Vi ali y includes posi i e eelings ha a e mo e ene gized han he eelings h p://dx.doi.o g/10.1016/j.landu bplan.2016.12.005 0169-2046/© 2016 The Au ho s. Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4. 0/). K. Ko pela e al. / Landscape and U ban Planning 160 (2017) 38–47 39 o es o a ion o elaxa ion (Ryan e al., 2010). Subjec i e i ali y is a eeling o ali eness (Nix, Ryan, Manly, & Deci, 1999) and ac i- a ed posi i e a ec (Ryan e al., 2010). Vi ali y inc eases posi i e coping esponses, be e sel -con ol and enhances p oduc i eness (Ryan & Deci, 2008), all ele an o employee well-being. In he li - e a u e, he i alizing e ec s o na u e exposu e ha e ecei ed less esea ch a en ion han he s ess- educing aspec s. Thus, in his s udy we in es iga e whe he he pe cep ions o he use o na u al elemen s a wo k and a home a e ela ed o posi i e and sel - enhancing aspec s o employee well-being, pa icula ly subjec i e i ali y. Mo eo e , a ele an ou come dese ing in es iga ion in his con ex is he eeling o c ea i i y a wo k which may be ela ed o i ali y and esul om a en ion es o a ion and lowe s ess ul a ousal a e na u e exposu e (A chley, S aye , & A chley, 2012). In he ollowing, we e iew s udies on he associa ions o win- dow iews, plan s, and physical ac i i y in na u e wi h well-being and wo k- ela ed ou comes. We main ain ha impo an gaps in esea ch conce n causal di ec ions and he ela i e impo ance o di e en ypes o na u e exposu e. Ou s udy aims o add ess hese gaps by using a longi udinal design including se e al ypes o na u e exposu e. 1.1. Gaps in knowledge: he ela i e impo ance o he ypes o na u e exposu e and causal di ec ions Despi e a la ge body o empi ical esul s, he ela i e impo ance o di e en o ms o exposu e o he na u al wo ld o well-being emains elusi e. Many s udies ha e ocused on na u e exposu e only wi hin he wo kplace (A ies, Vei ch, & Newsham, 2010; La go- Wigh , Chen, Dodd, & Weile , 2011). In his s udy, we compa e he pe cep ions o exposu e o he na u al wo ld a he wo kplace wi h he pe cep ions o exposu e o he na u al wo ld in esiden- ial and ee ime en i onmen s. To he bes o ou knowledge, only one s udy so a has aken in o accoun he la e when s udying he ela ionship be ween he use o g eenspace a he wo kplace and well-being (Gilch is , B own, & Mon a zino, 2015). The c oss- sec ional esul s o ha s udy indica ed ha knowledge wo ke s’ use o g eenspace nea hei wo kplaces and window iews o na u e we e posi i ely associa ed wi h sel - epo ed well-being when con olling o he use o domes ic ga den and ou doo ac i - i y. O he c oss-sec ional s udies ha e demons a ed se e al well- being benefi s bu ha e no inc eased ou unde s anding abou he empo al and causal na u e o he ela ionships. Fo example, exis ing su ey s udies ha e ela ed o fices wi h plan s ( e sus lean o fices) wi h g ea e o e all job sa is ac ion, sel - a ed quali y o li e (D a igne, Waliczek, Linebe ge , & Zajicek, 2008), sel - epo ed c e- a i i y, and wo k sa is ac ion (B ingslima k, Ha ig, & Pa il, 2007). Rega ding physical ac i i y (PA) in na u al su oundings, a link o less need o eco e y om wo k has been epo ed (Ko pela & Kinnunen, 2011). Con e sely, he highe he le el o emo ional demands a he job, he mo e o en he ou doo ec ea ion a eas in he icini y o home a e isi ed on wo kdays (Degenha d , F ick, Buchecke , & Gu sche , 2011). Expe imen al esea ch on na u e exposu e and well-being, in u n, has ocused mainly on sho - e m e ec s (Knigh & Haslam, 2010) and shown mixed esul s. Fo ins ance, one s udy ound ha wo king in “g een o fices” esul ed in be e pe o mance in an a en ion capaci y es (Raanaas, E ensen, Rich, Sjøs øm, & Pa il, 2011) whe eas ano he s udy ound no change in ei he di ec ed a en ion capaci y o sel - epo ed es o a ion (E ensen, Raanaas, Häge häll, Johansson, & Pa il, 2015). Howe e , expe i- men s wi h ollow-ups ha e indica ed ha en iching a p e iously lean o fice wi h plan s was associa ed wi h less disengagemen (e.g., eelings o apa hy o i edness) among employees, which in u n enhanced hei wo kplace sa is ac ion in he sho - e m (2 weeks) and he long- e m (3.5 mon hs) (Nieuwenhuis, Knigh , Pos mes, & Haslam, 2014). S udies compa ing na u e walks wi h u ban walks ha e shown posi i e sho - e m e ec s o na u e walks on a en ion es o a ion and on ne ous sys em a ousal (Aspinall, Ma os, Coyne, & Roe, 2015; Ha ig, E ans, Jamne , Da is, & Gä ling, 2003; Pa k, Tsune sugu, Kase ani, Kagawa, & Miyazaki, 2010). Simi- la ly, a s udy compa ing di e en lunch ime walks in he wo kplace showed imp o emen in sel - epo ed men al heal h only in he na u e walk g oup (B own, Ba on, P e y, & Gladwell, 2014). Expo- su e o na u e in he o m o ga dening has shown sho - e m es o a i e e ec s on mood and co isol when compa ed o indoo eading ( an den Be g & Cus e s, 2011). Some longi udinal in e en ion s udies sugges a p edic i e ole o na u e exposu e in well-being. Fo example, ga dening as well as na u e-based he apies ha e been ela ed o be e o e all heal h and well-being among adul s (G oenewegen, an den Be g, de V ies, & Ve heij, 2006; Wäh bo g, Pe e sson, & G ahn, 2014). Longi udinal s udies wi h a ocus on he gene al popula ion ha e sugges ed ha mo e g eenspace in esiden ial a eas is ela ed o lowe le els o s ess (Wa d Thompson e al., 2012) and ha mo - ing o g eene a eas is ela ed o g ea e subsequen happiness and li e sa is ac ion (Whi e, Alcock, Wheele , & Depledge, 2013). 1.2. Resea ch ques ions and hypo heses The p esen s udy add esses se e al gaps in he esea ch by explo ing ela ionships be ween sel - epo ed exposu e o he na - u al wo ld bo h a wo k and a home and employees’ expe iences o well-being. We s udy he posi i e and sel -enhancing aspec s o well-being: i ali y, happiness, igo , and c ea i i y. These ela ion- ships a e examined o e ime (one yea ) as longi udinal e idence compa ing di e en ypes o na u e exposu e is sca ce. Fu he - mo e, we aim o in es iga e exposu e o he na u al wo ld in a wide con ex han p e iously by con olling o heo e ically ele- an job cha ac e is ics and he s abili y o well-being expe iences o e ime. To con ibu e o he a e compa isons be ween di e en ypes o exposu e o he na u al wo ld, we selec ed pe cei ed na u e exposu e a iables ha eflec di e en en i onmen s − na u al elemen s a wo k ( he numbe o indoo plan s, iews om win- dows, looking ou o he window), a home (looking ou o he window, being in he ga den), and du ing ee ime (physical ac i - i ies in na u al su oundings). The only s udy so a con olling o he p esence o a domes ic ga den and pa icipa ion in ou doo ac i i ies (Gilch is e al., 2015) did no con ol o he equency o looking ou o windows o he use o he domes ic ga den. Mo eo e , ou na u e exposu e a iables ep esen a pe cei ed dimension o inc easing imme sion in na u al su oundings and an inc easing amoun o physical ac i i y ( om si ing and looking a plan s o ou doo ec ea ion). This concu s wi h some indica ions in he li e a u e o a linea dose- esponse ela ionship be ween expo- su e o he na u al wo ld and well-being ou comes (Fjeld, 2000; Jiang, Li, La sen & Sulli an, 2014). As he majo i y o ea lie s udies a e c oss-sec ional causali y has emained la gely an open ques ion. In ou s udy, we aim o gene a e be e condi ions o causal hypo heses by using a wo- wa e panel design. We use a one-yea ime-lag as p e ious s udies p o ide a ange om one yea (As ell-Bu , Mi chell, & Ha ig, 2014; Kinnunen & Feld , 2013; Whi e e al., 2013) o fi e yea s o po en- ial long- e m e ec s. In a 5-yea ollow-up s udy, g een quali ies a ound he esidence in an in e ac ion wi h physical ac i i y p e- dic ed men al heal h (e.g., happiness, abili y o ace p oblems and enjoy e e yday ac i i ies) o women (Anne s ed e al., 2012). To unde s and hese p e ious esul s and jus i y he one-yea lag, we specula e ha he ypes o na u e exposu e in ou s udy (e.g., look- ing ou , being physically ac i e) ep esen ongoing daily, weekly, 40 K. Ko pela e al. / Landscape and U ban Planning 160 (2017) 38–47 o mon hly ecu en loops o beha io s ha , in p inciple, can a ec well-being o e longe ime pe iods, such as one yea . Thus, encoun e ing na u e exposu es day a e day will inc ease he like- lihood o con inuously high well-being. Ou well-being a iables a e consis en wi h esea ch on i al- i y, a en ion and s ess es o a ion bu eflec he posi i e and sel -enhancing a he han s ess- educing aspec s o well-being. We include expe iences o i ali y and happiness, which ha e been di e en ia ed because i ali y is cha ac e ized by high ene gy o ac i a ion, which is no necessa ily ue o happiness (Nix e al., 1999). Vi ali y- ela ed a ec s a e, o example, peppy, ac i e, and en husias ic, whe eas happiness- ela ed a ec s may be con en and sa isfied (Nix e al., 1999). We also conside igo a wo k as a job- ela ed well-being expe ience o i ali y. Vigo is a key dimension o wo k engagemen and e e s o high le els o ene gy and men al esilience while wo king, he willingness o in es in one’s wo k, and pe se e ance in he ace o di ficul ies (Schau eli, Salano a, González-Romá, & Bakke , 2002). Fu he , c ea i i y a wo k, defined as he p oduc ion o no el, use ul ideas o p oblem solu ions (Amabile, Ba sade, Muelle , & S aw, 2005) may esul om a en ion es o a ion and lowe s ess ul a ousal in na u e (A chley e al., 2012). The e ec s o na u e on posi i e mood (e.g., happiness) and he consequen s imula ion o insigh and di e gen associa ions ha e also been p oposed as mechanisms inc easing pe o mance in c ea i e asks (Shiba a & Suzuki, 2002, 2004). Based on p e ious c oss-sec ional findings we an icipa e ha (H1) di e en pe cei ed ypes and deg ees o na u e exposu e a e all posi i ely and independen ly (a e accoun ing o o he a i- ables) ela ed o well-being ou comes o e one yea . An in e es ing ques ion is whe he happiness, i ali y, igo , and c ea i i y a e in e ela ed psychological p ocesses and show consis en associa- ions wi h na u e exposu e o whe he benefi s di e wi h di e en well-being ou comes. In addi ion, we include adequa e con ol a iables ha ep e- sen al e na i e explana ions o he ela ionships be ween na u e exposu e and well-being. In he employee well-being li e a u e, i ali y, ene gy, and sel - egula o y capaci ies − eflec ed in ou ou comes o i ali y, happiness, igo , and c ea i i y − a e assumed o be enhanced by sa is ac ion o needs o ela edness ( eel- ing connec ed), au onomy (sense o choice), and compe ence (e ec i eness) (Ryan & Deci, 2001, 2008). Consequen ly, job cha - ac e is ics including 1) suppo om colleagues and supe io s ( ela edness), 2) au onomy a wo k (au onomy), and 3) wo kload (indica ing a challenge and po en ial o compe ence) ma ch hese needs and a e also he mos impo an job cha ac e is ics a ec ing employee well-being in Ka asek’s well-known demand-con ol- suppo model (Ka asek & Theo ell, 1990). Thus, we con ol o hem. Ano he al e na i e explana ion no con olled o in p e i- ous s udies is he numbe o b eaks om wo k ha a pe son akes du ing he day. Mo e b eaks ha e been ela ed o be e well-being and less a igue (Tucke , 2003), which means mo e igo and pos- i i e a ec i e s a es. Hence, he numbe o b eaks may be di ec ly linked o well-being, and b eaks may also p o ide mo e oppo u- ni ies o engage wi h na u al elemen s. Fu he mo e, i is known ha posi i e a ec s a e mode a ely s able o e mon hs, e en o e yea s (Lucas & Donnellan, 2007; Wa son & Walke , 1996). We accoun o his empo al s abili y which lea es explana o y powe and a iance only o impo an de e minan s and makes ou s udy design a mo e s ingen es o o he explana o y a iables. We also con ol o age and gende which may be associa ed wi h he dependen a iables (Shiba a & Suzuki, 2004). Ou las aim is o in es iga e bo h di ec ions o he ela ion- ship be ween exposu e o he na u al wo ld and well-being. The e is compelling e idence ha na u e exposu e imp o es well-being, bu as longi udinal s udies a e s ill ew and expe imen al e idence is limi ed, he e e se o ecip ocal di ec ions o he ela ionships emain plausible. The e e se assump ion s a es ha be e well- being inc eases exposu e o na u e (H2), whe eas he ecip ocal assump ion iews his ela ionship as bidi ec ional (H3). In sup- po o he e e sed e ec s, i is well known ha mood a ec s an indi idual’s selec ion o ce ain places, ac i i ies and expe iences while he e, and decisions o lea e (Ke & Tacon, 1999). In ac , he no ion ha people use pa icula places o sel - and emo ion egu- la ion is a common p esupposi ion in esea ch on a o i e places and es o a i e en i onmen s (Ko pela, Ha ig, Kaise , & Fuh e , 2001). The niche-building heo y, based on a b oade pe spec i e o he pe son-en i onmen fi (Robe s & Robins, 2004), posi s ha people end o c ea e, seek ou , o end up in physical and social en i- onmen s ha co ela e wi h hei pe sonal needs, disposi ions, and abili ies (Tesse , 2002). In a el and ou ism s udies, he desi e o elaxa ion and es o a ion, and he p e e ence o na u al en i- onmen s has been ecognized as one basis o he choice o a el des ina ions (Leh o, 2013). F om leisu e s udies, i is known ha leisu e in ol emen ha includes physical ac i i y ollows om se e al in insic mo i es, such as iden i y a fi ma ion, enjoymen , o physical fi ness (Aal onen e al., 2012). Hence, a mul i ude o pe - spec i es and empi ical esul s suppo s he no ion o he ac i e choice o en i onmen s o egula e one´ ıs well-being, po en ially p oducing e e se e ec s. 2. Me hod 2.1. Sample and p ocedu e The pa icipan s we e employees wo king in ele en o ganiza- ions in di e en sec o s, he la ges o which we e educa ion, in o ma ion echnology, and media. By sampling di e en o gani- za ions, we wan ed o include employees om a a ie y o di e en jobs. The o ganiza ions we e mainly ec ui ed om he clien o ga- niza ions o a company supplying occupa ional heal h ca e se ices. As no all ypes o na u al en i onmen s may be es o a i e − o ins ance, dense wooded a eas may also e oke ea and s ess (Ga e sleben & And ews, 2013) − in he p esen s udy, we ec ui ed wo kplaces nea u ban pa ks o woodlands (max. 150 m dis ance) ha we e well main ained and did no include h ea ening ea u es. The ques ionnai e da a we e collec ed in wo wa es as a pa o a la ge p ojec on eco e y a wo k aiming ul ima ely o a 3-wa e design. In he fi s phase, in sp ing 2013 (Time 1) an elec- onic ques ionnai e was sen ei he di ec ly o he employees’ wo k e-mail add esses (in se en o ganiza ions) o he link o he ques ionnai e was deli e ed by he con ac pe sons o he employ- ees (in ou o ganiza ions). O he employees con ac ed (N = 3593), 1347 e u ned he comple ed ques ionnai e a e wo eminde s, yielding a esponse a e o 37.5%. Second, in sp ing 2014 (Time 2) ano he elec onic ques ionnai e was sen o hose employees who esponded in 2013 and who we e s ill employed in he same o ga- niza ions (N = 1192). O hese, a o al o 841 employees e u ned he comple ed ques ionnai e, yielding a esponse a e o 70.6%. In bo h s udy phases, he employees ecei ed in o ma ion abou he goals o he s udy (“a longi udinal s udy abou eco e y om wo k including ques ionnai e wi h he hemes o wo k, wo k en i on- men , ee- ime, well-being, and heal h”) wi h he assu ance ha esponses would be ea ed confiden ially and ha pa icipa ion was olun a y. All he a iables used in his s udy we e measu ed a bo h ime poin s T1 and T2. A Time 1 o his longi udinal sample (N = 841), 58.6% o pa ic- ipan s we e women. The pa icipan s’ a e age age was 47.1 yea s ( ange 21–67, SD = 10.0). Mos o hem (76.4%) we e li ing wi h a pa ne (ei he ma ied o cohabi ing), and 45.6% had child en K. Ko pela e al. / Landscape and U ban Planning 160 (2017) 38–47 41 (a e age o wo) li ing a home. O he sample, 38.2% held a uni- e si y deg ee (mas e ’s le el o highe ), 26.6% had a poly echnic deg ee, and he es (35.2%) had a oca ional school qualifica ion o less. O he pa icipan s, 8.3% we e blue-colla wo ke s (e.g., cleane s), 30.0% we e lowe -whi e colla wo ke s (e.g., o fice wo k- e s), 57.8% we e highe whi e-colla wo ke s (e.g., eache s), and 3.8% we e highe -le el manage s (e.g., chie execu i e o fice s). The majo i y had a pe manen job (89.0%), wo ked ull- ime (96.8%), and had a egula day shi (89.7%). A e age hou s wo ked weekly we e 39.1 (SD = 5.9). O he pa icipan s, 55.6% wo ked in he public sec o , and he es (39.4%) wo ked in he p i a e sec o . In analyzing sample a i ion we compa ed he esponden s o he longi udinal sample (i.e., hose who eplied a T1 and T2) wi h he non- esponden s (who did no eply a T2). The esponden s did no di e om he non- esponden s in e ms o gende , age, ha ing a pa ne , numbe o child en, o educa ion. The espon- den s we e mo e o en employed as highe whi e-colla wo ke s (58% s. 50%) han he non- esponden s (p < 0.05) and had mo e o en a pe manen employmen con ac (89% s. 79%) han he non- esponden s (p < 0.001). Also, he esponden s wo ked longe hou s pe week (39.1 s. 37.9 h, p < 0.01) and mo e o en had a eg- ula day shi (90% s. 83%, p < 0.01) han he non- esponden s. When compa ing he s udy a iables a T1 (na u e exposu e, job cha ac e is ics, ou comes) wo di e ences we e ound. The non- esponden s a T2 epo ed ha ing somewha less au onomy a wo k (M = 3.1, SD = 0.85) han he long- e m sample (M = 3.2, SD = 0.82; (1226) = 2.2, p = 0.025) and ha ing sligh ly less social suppo (M = 3.9, SD = 0.74) han he long- e m sample (M = 4.0, SD = 0.66; (672.3) = 2.0, p = 0.043). 2.2. Measu es In o de o ele a e he esponse a e, o co e se e al esea ch hemes and no o bu den he employees excessi ely, we had o measu e he a iables as concise as possible in ou ques ion- nai e. In addi ion o pe cei ed na u e exposu e and well-being, he ques ionnai e included se e al measu es abou he wo k si ua ion, eco e y and ee ime, no epo ed he e. 2.2.1. Na u e exposu e a iables a he wo kplace The numbe o indoo plan s was measu ed by one i em asking “How many (a ificial o eal) plan s o flowe s do you ha e in sigh inside you oom/wo k s a ion?”. The esponse was gi en in numbe s. We included a ificial plan s in ligh o he ac ha e en pos e s o plan s can ha e simila al hough weake e ec s on well- being han eal plan s (Beukeboom, Lange eld, & Tanja-Dijks a, 2012). The ype o iew om he window was measu ed by asking “Do you ha e a window, a glass doo o a glass wall in you oom/wo k s a ion?”. The esponse ca ego ies we e 1) No, 2) Yes, i is o he inside o he building, 3) Yes, i is o he ou side o he building wi h mainly an u ban iew ( o example a building o s ee ) 4) Yes, i is o he ou side o he building wi h mainly a na u al iew ( o example a lake, field, o pa k). Fo he s a is ical analyses, we used a dicho omous a iable (1 = indoo o u ban iew and 2 = na u e iew) and excluded hose wi h no iews (5.9% o he sample). F equency o looking ou o he window was measu ed in he con- ex o ene gy managemen beha io s du ing he wo king day (De Bloom, Kinnunen, & Ko pela, 2015). We ecognized he di ficul y o epo ing on looking a plan s o window iew pe se du ing he wo king day bu hough ha people migh mo e easily ecog- nize he momen s when hey el ele a ed a e looking a hei su oundings. Thus, we asked “To wha ex en do you use each o he beha io s o manage you ene gy du ing you wo k day?” Among a lis o 13 beha io s we included “Look ou he window”. The esponse scale was a Like scale om 1 “Ve y seldom o ne e ” o 5 “Ve y o en o always”. 2.2.2. Na u e exposu e a iables a home and du ing ee ime F equency o looking ou a a na u e iew a home was measu ed wi h a ques ion “Do you ha e a window o balcony iew o na u- al su oundings, e.g., g eenspace, wa e , o a ga den?” wi h ou esponse ca ego ies 0) No, 1) Yes, bu I look a i ai ly seldom, 2) Yes, and I look a i some imes, 3) Yes, and I look a i o en. Use o one’s own back ya d (a home) wi h na u al elemen s was measu ed wi h one i em “Do you ha e a ga den, ya d, balcony o pa io wi h na u al (e.g., plan s, flowe s, ees) o wa e elemen s (e.g., a oun ain, a pond)?”. The esponse ca ego ies we e 0) No, 1) Yes, bu I use i seldom, 2) Yes, and I use i some imes, 3) Yes, and I use i o en. F equency o physical ac i i ies in na u al su oundings du ing ee ime was measu ed wi h one i em “How o en do you spend ee ime on he ollowing ac i i ies? Physical ac i i ies in na u al su - oundings (e.g., swimming, unning, cycling)” was included in he lis o se en ac i i ies. The esponse ca ego ies we e 1) Ha dly e e o a ew imes pe yea , 2) Abou once pe mon h, 3) A ew imes pe mon h, 4) Abou once pe week, 5) A ew imes pe week, 6) Almos e e y day. 2.2.3. Well-being a iables Happiness was measu ed wi h a single i em (“How happy do you eel in gene al?”) using a 10-poin scale anging om 1 ( e y unhappy) o 10 ( e y happy) (Abdel-Khalek, 2006; De Bloom, Geu s, & Kompie , 2013). Vi ali y was measu ed wi h ou i ems (excluding he syn- onymous exp essions o ene gy) om he se en-i em Subjec i e Vi ali y Scale (Ryan & F ede ick, 1997) (e.g., “Du ing he las mon h, I ha e el ali e and i al”, C onbach’s alpha = 0.88T1 and 0.88T2). The i ems we e a ed on a 5-poin scale om 1 ( e y seldom o ne e ) o 5 ( e y o en o always). Vigo a wo k was measu ed wi h h ee i ems (e.g., “A my wo k, I eel bu s ing wi h ene gy”, C onbach’s alpha = 0.89T1 and 0.90T2) om he sho ened U ech Wo k Engagemen Scale (UWES, Schau eli, Bakke , & Salano a, 2006), o which he cons uc alidi y has been ound o be good in Finnish occupa ional samples (Seppälä e al., 2009). The esponse scale anged om 0 (ne e ) o 6 (e e y day). C ea i i y a wo k was measu ed wi h h ee i ems (e.g., “My head is ull o inno a i e ideas ha a e ela ed o my wo k”, C onbach’s alpha = 0.84T1 and 0.86T2) (Geo ge & Zhou, 2001; Jaussi, Randel, & Dionne, 2007) a ed on a scale om 1 ( e y seldom o ne e ) o 5 ( e y o en o always). 2.2.4. Con ol a iables Job au onomy was measu ed wi h fi e i ems (e.g., “I can influence he amoun o wo k assigned o me“, C onbach’s alpha = 0.77T1 and 0.78T2) a ed on a 5-poin scale om 1 ( e y seldom o ne e ) o 5 ( e y o en o always) om he QPS No dic-ADW (Dallne e al., 2000). Social suppo om colleagues was measu ed wi h h ee i ems (e.g., “I needed, I can ge suppo and help wi h my wo k om my co-wo ke s”) and om supe iso s wi h h ee i ems (e.g., “My wo k achie emen s a e app ecia ed by my immedia e supe io ”) ha we e aken om he QPS No dic-ADW (Dallne e al., 2000). C onbach’s alpha o suppo wi h colleagues and supe iso s was 0.80T1 and 0.82T2. The i ems we e a ed on a 5-poin scale om 1 ( e y seldom o ne e ) o 5 ( e y o en o always). Wo kload was measu ed wi h h ee i ems (e.g., “How o en does you job equi e you o wo k unde ime p essu e?”; C onbach’s alpha = 0.88T1 and 0.87T2) om he QWI (Spec o & Jex, 1998). The 42 K. Ko pela e al. / Landscape and U ban Planning 160 (2017) 38–47 i ems we e a ed on a scale om 1 ( e y seldom o ne e ) o 5 ( e y o en o always). Numbe o b eaks las ing o e 10 min du ing a egula wo king day was elici ed wi h an open-ended ques ion. The esponse was gi en in numbe s. We also con olled o gende and age. 2.2.5. S a is ical analyses MANOVA and pai ed sample - es s wi h SPSS 22.0 we e used o examine he change in he mean le el o he a iables om T1 o T2 and ze o-o de co ela ions we e calcula ed o examine he ela ionships be ween he s udy a iables. We used a wo-wa e au o eg essi e c oss-lagged panel model, he “ adi ional longi u- dinal model” (Selig & Li le, 2012) o examine he hypo hesized ela ionships ac oss ime. All a iables we e measu ed a T1 and T2. In he model, au o eg essions e eal he s abili y o change o indi iduals’ ela i e s andings on he cons uc om one occasion o he nex (Selig & Li le, 2012). C oss-lagged associa ions a e ela- ionships be ween “independen and dependen ” a iables, i.e., he e ec o a cons uc on ano he measu ed a a la e occasion. C oss- lagged e ec s a e es ima ed con olling o au o eg essions, ha is, he p io le el o he cons uc being p edic ed. The inclusion o au o eg essions ules ou he possibili y ha c oss-lagged associa- ions a e due o he a iables being co ela ed a T1 (Selig & Li le, 2012). We an his “no mal causa ion” model in Mplus 7 including au o eg essions o all a iables and he c oss-lagged eg essions om na u e exposu e and con ol a iables a T1 o ou well-being measu es a T2 (Fig. 1). We also an he “ e e se causa ion” model wi h happiness, i ali y, igo , and he sense o c ea i i y a T1 p e- dic ing na u e exposu e and con ol a iables a T2 o assess he di ec ion o he hypo hesized associa ions. Las , we an he “ ecip- ocal” model which assumes ha hese pa hs be ween a iables we e bidi ec ional. Accep able model fi s we e assessed wi h he ollowing c i e ia: he ␹2 es (Bollen, 1989), Compa a i e Fi Index (CFI) and Tucke Lewis Index (TLI) wi h alues g ea e han 0.90 o .95, Roo Mean Squa e E o o App oxima ion (RMSEA) wi h alues smalle han 0.06 − 0.08, and S anda dized Roo Mean Squa e Residual (SRMR) wi h alues smalle han 0.08 (Sch eibe , No a, S age, Ba low, & King, 2006). Fo he compa isons o non-nes ed models (no mal s. e e sed), Akaike in o ma ion c i e ion (AIC), Bayes in o ma- ion c i e ion (BIC) and sample-size adjus ed BIC we e used wi h he c i e ion “ he smalle , he be e ” (Sch eibe e al., 2006). Fo he compa ison o nes ed models ( ecip ocal s. no mal o e e sed), we used he Sa o a-Ben le scaled ␹2di e ence es . The models we e es ima ed wi h he MLMV es ima o , which is a maximum likelihood −based es ima o ha is obus o non-no mali y o obse ed a iables (Mu hén & Mu hén, 1998-2012). 3. Resul s 3.1. Desc ip i e esul s The co ela ions be ween he na u e exposu e a iables we e low (excep o na u e iew a home and use o ga den, balcony, o ya d a home), indica ing low mul icollinea i y (Table 1). The co e- la ions be ween he na u e exposu e a iables and con ol a iables we e likewise low. As an icipa ed, all na u e exposu e a iables had some significan posi i e co ela ions wi h he well-being a iables bo h c oss-sec ionally and o e ime. The co ela ions be ween job- ela ed con ol a iables (au on- omy, suppo , wo kload, b eaks) and well-being a iables we e o mode a e s eng h and mos ly significan c oss-sec ionally and o e ime. The s onges c oss-sec ional co ela ions we e be ween suppo and i ali y, igo and happiness. Wo kload co ela ed wi h c ea i i y bu no wi h igo , happiness, o i ali y. Numbe o b eaks co ela ed wi h happiness and i ali y bu no wi h igo and c ea i i y. Vigo , happiness, and i ali y we e in e ela ed qui e s ongly and significan ly, c ea i i y ha ing lowe bu s ill signifi- can ela ions o o he well-being a iables. MANOVA e ealed s a is ically significan (Pillai F(13,671) = 8.7, p < 0.001) inc eases om T1 o T2 in he mean le el (Table 1) o equency o looking ou o he window a wo k and a home, and numbe o b eaks. Vigo , c ea i i y, and wo kload dec eased. As high co ela ions be ween dependen a iables in MANOVA may educe powe , we confi med hese esul s wi h sepa a e pai ed samples - es s. 3.2. Au o eg essi e c oss-lagged pa h models Ou hypo hesized model o “no mal causa ion” showed a good fi o he da a excep o he 2 es , which is gene - ally known o be o e sensi i e wi h la ge samples (Bollen, 1989) (2162 = 415.3, p < 0.001; CFI = 0.95, TLI = 0.91; RMSEA = 0.05; SRMR = 0.04; AIC = 17.819; BIC = 18.609; adj BIC = 18.035). The model explained 43.9% o he a ia ion in happiness, 40.1% in i ali y, 46.3% in igo , and 53.8% in sense o c ea i i y a T2. Examina- ion o he s anda dized coe ficien s (Table 2) showed ha equen physical ac i i y in na u al su oundings a T1 was associa ed wi h s onge eelings o i ali y a T2. F equency o using one’s back ya d wi h na u e elemen s a T1 was ma ginally ela ed o happi- ness a T2. O he con ol a iables, wo kload and au onomy a T1 p edic ed highe c ea i i y a wo k a T2 and suppo p edic ed i ali y. Au o eg ession coe ficien s we e subs an ial o happi- ness (ˇ = 0.63), i ali y (ˇ = 0.54), igo (ˇ = 0.64), and c ea i i y (ˇ = 0.70), all p < 0.001, sugges ing no able s abili y. All o he coe - ficien s we e non-significan . The “ e e se causa ion model” had a nea ly iden ical fi wi h he da a (2162 = 420.5, p < 0.001; CFI = 0.95, TLI = 0.90; RMSEA = 0.05; SRMR = 0.04; AIC = 17.820; BIC = 18.611; adj BIC = 18.036) al hough he in o ma ion c i e ia we e sligh ly be e in he “no mal cau- sa ion model”. The e e sed model explained 53.6% o a iance in numbe o plan s, 34.7% in ype o iew om he window, 33.9% in equency o looking ou o he window a wo k, 26.8% in e- quency o looking ou o he window a home, 39.9% in use o ga den, balcony, pa io o ya d wi h na u e elemen s a home and 44.6% in equency o physical ac i i y in na u al su oundings a T2. S anda dized coe ficien s (Table 3) showed ha sense o c e- a i i y a wo k a T1 p edic ed mo e equen use o one’s back ya d wi h na u e elemen s a T2. Con e sely, i ali y a T1 p e- dic ed less equen use o one’s back ya d wi h na u e elemen s a T2. Happiness was ma ginally posi i ely ela ed o he equency o physical ac i i y in na u e. Au o eg ession coe ficien s we e sub- s an ial; o indoo plan s ˇ = 0.73, o ype o iew om window ˇ = 0.59, o equency o looking ou ˇ = 0.57, o looking ou a na u e a home ˇ = 0.51, use o ya d/ga den a home ˇ = 0.61 and o physical ac i i y in na u e ˇ = 0.66, all p < 0.001. All o he coe - ficien s we e non-significan . The “ ecip ocal model” showed no imp o emen in he model fi ela i e o he hypo hesized no mal (2di (40) = 42.1, p =0.38) o e e se model (2di (40) = 46.8, p =0.21). Fu he mo e, he ecip o- cal model e ealed no new significan pa hs. 4. Discussion O e all, we ob ained limi ed suppo o he link be ween pe - cei ed na u e exposu e and employee well-being o e ime. The link be ween na u e and well-being was no clea ly unidi ec ional o e ime. While acknowledging he limi a ions in ou design and measu es, hese “conse a i e” findings show ha in eal-li e con- K. Ko pela e al. / Landscape and U ban Planning 160 (2017) 38–47 43 Table 1 Means, s anda d de ia ions and Pea son co ela ions be ween na u e elemen s a wo k and home, backg ound, job cha ac e is ic and well-being a iables. 1 Plan s 2 View ype 3 Look ou 4 View, home 5 Ya d, home 6 Na u e, PA 7 Gende a 1 emale 2 male 8 Age a 12 B eaks 13 Au o- nomy 14 Sup- po 15 Wo k load 16 C ea i i y T1 T2 17 Vigo T1 T2 18 Happiness T1 T2 19 Vi ali y T1 T2 1. Plan s .69*** - .05 .01 .04 .08* .00 - .18** * - .01 -.12** - .07* .04 .19** * - .00 .08 * .00 - .0 4 2. View, u ban s na u e -.07 .58 *** .08* .03 .03 .02 .02 - .03 .00 .07 .04 - .03 .03 .05 .06 .09 ** 3. Looking ou o he window -.004 .12** .55*** .16** * .09* * .12** * - .10** - .01 .09** .09* * .09* * .00 .10 ** .09 ** .04 .07 4. Na u e iew, home .03 .004 .21*** .57*** .52** * .13** * - .05 .24** * .04 .04 .02 .04 .15 *** .15 *** .12 *** .12 *** 5. Use o home ga den / ya d .03 .02 .17*** .51*** .66*** .16*** - .04 .23** * .06 - .03 .02 .11** * .14 *** .12 *** .13 *** .12 *** 6. Na u e, phys.ac i i y .06 .05 .10** .14*** .20*** .66*** - .18** * .00 .02 .01 .03 .02 .05 .08 * .14 *** .16 *** 7. Gende a -.22*** -.01 -.14*** -.08* -.05 -.17*** .04 .23*** .16*** - .02 -.15*** .07 * - .09 ** - .08 * .03 8. Age a .001 -.03 .01 .25*** .28*** .03 .04 .16*** - .06 - .05 - .07 - .01 .0 0 - .02 .04 12. F eq o > 10 min. b eaks -.10** .02 .11** .04 .004 .07* .26*** .07 .67*** .06 .05 -.22*** .03 .01 .09 ** .09 ** 13. Au onomy -.05 .003 .04 -.03 -.03 .03 .16*** -.08* .07* .74*** .34*** -.24*** . 20 *** .26 *** .1 2*** .29 *** 14. Suppo .11** .05 .01 .00 .02 .05 -.02 -.10** .02 .32*** .63*** -.05 .17 *** .3 9 *** .32 *** .39 *** 15. Wo kload .10** -.09* -.01 .05 .11** .02 -.16*** .03 -.24*** -.30*** -.08* .69*** .18 *** .0 6 - .0 3 - .0 7* 16. C ea i i yT1 -.02 .08* .08* .11*** .12*** .04 .05 .01 -.01 .15*** .15*** .15*** C ea i i yT2 .00 .05 .06 .09** .08* .05 .07* -.01 -.01 .16*** .12*** .15*** .72*** .30*** .11** .27*** 17. Vigo T1 .09* .07 .04 .10** .12*** .19*** -.06 .01 .04 .26*** .39*** .01 .35*** Vigo T2 .07* .03 .05 .09* .11*** .15*** -.09** .00 .00 .19*** .29*** .04 .24*** .33*** .68*** .34*** .56*** 18. HappinessT1 .05 .03 .03 .14*** .16*** .17*** -.12*** .004 .08* .13*** .26*** .01 .14*** .41*** Happiness T2 .04 .05 .03 .10** .14*** .16*** -.08* -.02 .10** .11*** .24*** -.02 .15*** .16*** .37*** .37*** .68*** .50*** 19. Vi ali y T1 .04 .06 .07 .11*** .12*** .20*** -.01 .002 .11** .27*** .40*** -.07 .29*** .65*** .59*** Vi ali y T2 .00 .07 .05 .11*** .12*** .19*** .03 .04 .11** .22*** .30*** -.05 .26*** .30*** .50*** .63*** .47*** .57*** .67*** Response scale 0-20 1-2 1-5 0-3 0-3 1-6 1-2 21 -67 0-8 1-5 1-5 1-5 1-5 0-6 1-10 1-5 Mean T1 (T2) 2.28 (2.12) 20% na u e 3.02 (3.30) 2.40 (2.57) 2.29 (2.32) 4.30 (4.30 ) 59% emale 47.13 2.44 (2.54) 3.18 (3.20) 3.99 (3.96) 3.89 (3.82) 3.45 (3.37) 4.53 (4.37) 7.77 (7.83) 3.37 (3.41) SD T1 (T2) 3.50 (3.51) 22% na u e 1.20 (1.13) .88 (.77) .96 (.90) 1.30 (1.30 ) 10.02 1.12 (1.18) .82 (.80) .66 (.69) .82 (.79) .77 (.77) 1.21 (1.32) 1.59 (1.52) .79 (.78) *, **, ***Significan a p ≤ 0.05, 0.01, o 0.001 espec i ely. N = 683–841. Uppe iangle (g ey a ea): in e co ela ions a T2. Lowe iangle: in e co ela ions a T1 excep well-being a iables a T2. Au oco ela ions be ween T1 and T2 on he diagonal, unde lined. a: Gende and age a e om T1– he same esponden s answe ed a T2. 44 K. Ko pela e al. / Landscape and U ban Planning 160 (2017) 38–47 Vi ali y T2 Happiness T2 Vigo T2 C ea i i y T2 Vi ali y T1 Happiness T 1 Vigo T1 C ea i i y T1 Indoo plan s T1 Window iew T1 Looking ou , wo k T1 Looking ou , home T1 Indoo plan sT2 Use o ya d, home T1 PA in na u e T1 Job au onomy T1 Social suppo T1 Wo kload T1 Numbe o b eaks T1 Gende T1 Age T1 Windo w iew T2 Looking ou , wo k T2 Looking ou , home T2 Use o ya d, home T2 PA in na u e T2 Job au onomy T2 Social suppo T2 Wo kload T2 Numbe o b eaks T2 Fig. 1. The hypo he ical “no mal causa ion“ wo-wa e au o eg essi e c oss-lagged panel model. To simpli y he figu e, he ou sho a ows om each o he na u e exposu e and con ol a iables ep esen pa hs o ou well-being a iables (happiness, i ali y, igo , c ea i i y). Table 2 “No mal causa ion” model, s anda dized pa h coe ficien s be ween na u e exposu e, con ol and well-being a iables in an au o eg essi e c oss-lagged panel model. Significan coe ficien s a e bolded, ma ginally significan (p ≤ 0.06) in i alics. T2: Happiness Vi ali y Vigo C ea i i y T1: ˇ S.E. p ˇ S.E. p ˇ S.E. p ˇ S.E. p 1. Plan s 0.005 0.029 0.849 −0.006 0.027 0.837 0.024 0.028 0.384 0.010 0.025 0.689 2. View, u ban s. na u e 0.031 0.028 0.283 0.028 0.031 0.365 0.004 0.029 0.897 −0.008 0.027 0.780 3. Looking ou o he window −0.005 0.030 0.870 −0.003 0.031 0.920 0.022 0.029 0.435 0.000 0.030 0.998 4. Na u e iew, home −0.017 0.038 0.658 0.050 0.031 0.110 −0.017 0.031 0.574 0.006 0.032 0.853 5. Use o home ga den/ya d 0.076 0.041 0.061 0.032 0.035 0.362 0.051 0.032 0.107 −0.032 0.035 0.359 6. Na u e, physical ac i i y 0.036 0.037 0.321 0.070 0.033 0.031 0.005 0.029 0.855 0.013 0.029 0.648 7. Gende −0.009 0.035 0.794 0.042 0.033 0.205 −0.053 0.030 0.072 0.039 0.031 0.206 8. Age −0.046 0.034 0.177 0.045 0.032 0.157 −0.015 0.030 0.623 0.001 0.031 0.982 12. F eq o > 10 min b eaks 0.001 0.036 0.982 0.051 0.034 0.135 0.013 0.030 0.650 −0.016 0.031 0.607 13. Au onomy −0.004 0.039 0.912 0.049 0.036 0.173 0.024 0.031 0.447 0.086 0.032 0.008 14. Suppo 0.023 0.038 0.552 0.088 0.035 0.012 0.046 0.034 0.174 −0.005 0.030 0.875 15. Wo kload 0.014 0.037 0.694 0.024 0.032 0.455 0.010 0.031 0.745 0.078 0.029 0.008 ex s only a small po ion o he a iance in employee well-being migh be explained wi h pe cei ed na u e exposu e a e con- olling o he s abili y o well-being and job cha ac e is ics. Ou esul s highligh ha he inclusion o physical ac i i y in he in e - ac ion wi h na u e migh be mo e e ec i e han mo e passi e o ms o in e ac ion. Mo e specifically, he p esen esul s sugges ha o p omo e employee i ali y and happiness independen ly o job cha ac e is ics, age o gende , i may be beneficial o specifically s imula e and plan o na u e-based ee- ime ac i i ies. K. Ko pela e al. / Landscape and U ban Planning 160 (2017) 38–47 45 Table 3 “Re e se causa ion” model, s anda dized coe ficien s be ween well-being and na u e exposu e a iables in an au o eg essi e c oss-lagged panel model. Significan coe ficien s a e bolded, ma ginally significan (p ≤ 0.06) in i alics. Fo b e i y pa h coe ficien s o con ol a iables a e no epo ed. T2: Plan s View, u ban s. na u e Looking ou o he window Na u e iew, home Use o home ga den/ya d Na u e, physical ac i i y T1: Happiness ˇ −0.008 0.030 0.047 −0.011 0.090 0.073 S.E. 0.036 0.038 0.040 0.059 0.049 0.038 p 0.829 0.427 0.237 0.858 0.067 0.057 Vi ali y ˇ −0.026 −0.051 −0.042 −0.013 −0.130 −0.004 S.E. 0.037 0.052 0.051 0.052 0.054 0.046 p 0.481 0.330 0.403 0.804 0.016 0.928 Vigo ˇ −0.010 −0.013 0.042 0.066 0.006 −0.015 S.E. 0.048 0.046 0.043 0.047 0.045 0.044 p 0.832 0.780 0.328 0.158 0.895 0.732 C ea i i y ˇ 0.025 −0.002 0.060 0.024 0.082 0.022 S.E. 0.029 0.036 0.038 0.036 0.034 0.034 p 0.396 0.964 0.114 0.512 0.016 0.516 4.1. Hypo hesis ega ding exposu e o he na u al wo ld, “no mal causa ion model” On he basis o p e ious findings, we an icipa ed ha (H1) he pe cep ions o di e en ypes and deg ees o exposu e o he na u al wo ld would be posi i ely and independen ly ela ed o psychological well-being ac oss ime. Acco dingly, mo e equen , sel - epo ed physical ac i i y in na u e p edic ed s onge eel- ings o i ali y one yea la e . Hence, physical ac i i ies in na u e e ealed independen powe in “explaining” i ali y. Howe e , he size o his associa ion emained ela i ely small, possibly because we placed exposu e o he na u al wo ld in a s ingen es by including impo an job cha ac e is ics and he s abili y o well- being in ou model. Au o eg essions emained he mos powe ul de e minan s in he longi udinal models. Thus, i ali y o happiness a e qui e s able in e indi idually ac oss one yea which cons ains he size o he associa ion ha na u e exposu e can ha e wi h such eelings longi udinally. Mo eo e , he mo e equen use o one’s ga den, back ya d o pa io was ela ed o inc eased happiness bu only ma ginally. Thus only he mos in ensi e o ms o exposu e o he na u al wo ld − sel - epo ed physical ac i i y ( o i ali y) and he use o he ga den ( o happiness) − showed ela ions o some aspec s o well-being o e one yea . Bo h exposu es we e ela ed o home and ee ime en i onmen s. E en hough he mean le els o sel - epo ed equencies o looking ou o he window a wo k and a home inc eased om he fi s measu emen poin o he second and he es emained qui e s able, none o he na u e exposu e a iables a wo k p e- dic ed well-being. We specula ed ha na u e exposu e a iables ep esen con inually epea ed beha io s o e he yea bu we e no able o check his empi ically. On he o he hand, i ou non-significan esul s hold ue also in u u e longi udinal s ud- ies, ea lie c oss-sec ional su ey esul s showing se e al posi i e e ec s o wo kplace na u e exposu e may ha e ac ually e ealed only sho - e m associa ions and/o may ha e su e ed om he common me hod a iance bias. The exis ence and he o m o a long- e m dose- esponse ela ionship be ween na u e exposu e and well-being s ill emains an open ques ion. Ou esul s ag ee wi h ea lie s udies s a ing ha he e ec s o na u e exposu e depend on he ou comes (B ingslima k, Ha ig & Pa il, 2009). Vi ali y bu no , o example, happiness was esponsi e o sel - epo ed physical ac i i y in na u e o e ime. This is in e - es ing as he e is e idence ha momen a y happiness in na u al en i onmen s is g ea e han in u ban en i onmen s (MacKe on & Mou a o, 2013). One eason o ou finding may lie in he ac ha i ali y had he lowes s abili y ac oss one yea ; he e o e he e was mo e a iance le o be explained o example by physical ac i i y. I is known ha physical exe cise inc eases i ali y pe se bu , as ye , he e is no clea ly o mula ed heo y ega ding he specific ea u es o he na u al en i onmen influencing i ali y. Vi al- i y may a ise om di e en aspec s o human-na u e in e ac ion han in a en ion o s ess es o a ion. I is no ewo hy ha he associa ion o exposu e o he na u al wo ld wi h i ali y was s a is- ically significan (and ma ginally also o happiness), independen o wo k- ela ed suppo , au onomy, wo kload, and he numbe o b eaks. Thus hey a e likely based on di e en psychological mechanisms han ela edness, compe ence, and au onomy a wo k. Again, he p ocesses o a en ional and s ess es o a ion may be ope a ional bu as we could no measu e hese mechanisms in he p esen s udy, he sugges ions emain specula i e. Fu u e he- o e ical con ibu ions and longi udinal s udies ocusing on hese di e ences in a ec i e e ec s a e called o . 4.2. Hypo heses ega ding he “ e e se” and “ ecip ocal models” The e e se model showed an almos equal fi wi h ou da a sup- po ing H2 and sugges ing ha in his field o s udy, i is impo an o emembe ha mood and a ec i e s a es may a ec he choice be ween en i onmen s and hus na u e exposu e. In he e e se model, sel - epo ed c ea i i y a wo k p edic ed mo e equen sel - epo ed use o one’s ya d/ga den. I seems possible ha people who eel mo e c ea i e a e inclined o selec g een housing en i on- men s o use hem mo e o en. In con as o he “no mal causa ion” whe e he mo e equen use o one’s ya d ma ginally p edic ed mo e happiness, mo e happiness, in u n, p edic ed mo e equen physical ac i i y in na u e (ma ginally). Vi ali y p edic ed he less equen use o one’s ya d al hough mo e equen physical ac i i y in na u al se ings p edic ed s onge i ali y in he “no mal causa- ion model”. I can be specula ed ha as equen physical ac i i y seems o be ela ed o s onge i ali y and happiness, mo e i al and happie people equi e and use mo e dis an and la ge na u al a eas han hei home ya d o be able o be su ficien ly ac i e. The ecip ocal model (H3) go no suppo . 46 K. Ko pela e al. / Landscape and U ban Planning 160 (2017) 38–47 4.3. Limi a ions A limi a ion o he p esen s udy includes using sel - epo measu es o all a iables, gi ing ise o conce ns abou common me hod a iance (CMV). Tempo ally sepa a ing p edic o and c i- e ion a iables, as in he p esen s udy, is one accep able way o educing he isk o CMV (Podsako , MacKenzie, Lee, & Podsako , 2003). Despi e his and he ac ha we modelled simul aneously se e al na u e exposu e a iables while con olling o heo e i- cally impo an a iables and e en es ed e e se and ecip ocal models, causal claims equi e u he esea ch as ou models we e co ela i e. The one-yea ime lag be ween ou measu emen s may no ha e been op imal and should be a ied in u u e s udies. Ou selec ion was based on ea lie esea ch bu also on he p ac ical eali y o busy o ganiza ional se ings no allowing da a collec ions wi hin sho ime spans. A u he limi a ion o he p esen s udy and an op ion o u u e s udies is ha we we e no able o measu e he in ensi y o na u e- based physical ac i i y o he amoun o quali y o in e ac ion wi h indoo plan s du ing he wo king day. The es - e es eliabili y o ou one-i em measu es was sa is ac o y (co ela ions be ween T1 and T2 anging om 0.55 o 0.69). Howe e , he alidi y o epo - ing, e.g., he use o he ga den may be ques ioned, pa icula ly i ha use a ies by season. Al hough he pa icipan s knew ha he s udy was abou eco e y om wo k, he ques ionnai e included so many ques ions ha i is unlikely ha demand cha ac e is ics (guessing he ela ionships ha we e s udied) bias he esul s significan ly. The e was some selec i e a i ion in ou sample. Hence, he findings canno be gene alized o he wo king popula ion as a whole; hose mo e commi ed o hei companies may ha e pa - icipa ed mo e likely. The esul s migh look di e en in samples expe iencing hea ie wo kloads o ope a ing in di e en sec o s o cul u al con ex s. Acknowledgemen This p ojec was suppo ed by he Academy o Finland (g an no.: 257682). Re e ences Aal onen, S., Leskinen, T., Mo is, T., Alen, M., Kap io, J., Liukkonen, J., e al. (2012). Mo i es o and ba ie s o physical ac i i y in win pai s disco dan o leisu e ime ac i i y o 30 yea s. In e na ional Jou nal o Spo s Medicine, 33, 157–163. Abdel-Khalek, A. M. (2006). Measu ing happiness wi h a single-i em scale. Social Beha io and Pe sonali y, 34, 139–150. Amabile, T. M., Ba sade, S. G., Muelle , J. S., & S aw, B. M. (2005). A ec and c ea i i y a wo k. Adminis a i e Science Qua e ly, 50, 367–403. Anne s ed , M., Os e g en, P.-O., Bjö k, J., G ahn, P., Skä bäck, E., & Wäh bo g, P. (2012). G een quali ies in he neighbou hood and men al heal h − Resul s om a longi udinal coho s udy in Sou he n Sweden. BMC Public Heal h, 12(1), 337. A ies, M. B. C., Vei ch, J. A., & Newsham, G. R. (2010). Windows, iew, and o fice cha ac e is ics p edic physical and psychological discom o . Jou nal o En i onmen al Psychology, 30, 533–541. Aspinall, P., Ma os, P., Coyne, R., & Roe, J. (2015). The u ban b ain: analysing ou doo physical ac i i y wi h mobile EEG. B i ish Jou nal o Spo s Medicine, 49, 272–276. As ell-Bu , T., Mi chell, R., & Ha ig, T. (2014). The associa ion be ween g een space and men al heal h a ies ac oss he li e cou se: A longi udinal s udy. Jou nal o Epidemiology and Communi y Heal h, 68, 578–583. A chley, R. A., S aye , D. L., & A chley, P. (2012). C ea i i y in he wild: Imp o ing c ea i e easoning h ough imme sion in na u al se ings. Plos One, 7(12), e51474. Beukeboom, C. J., Lange eld, D., & Tanja-Dijks a, K. (2012). S ess- educing e ec s o eal and a ificial na u e in a hospi al wai ing oom. The Jou nal o Al e na i e and Complemen a y Medicine, 18(4), 329–333. Bollen, K. A. (1989). S uc u al equa ions wi h la en a iables. New Yo k: Wiley. B a man, G. N., Hamil on, J. P., & Daily, G. C. (2012). The impac s o na u e expe ience on human cogni i e unc ion and men al heal h. Annals o he New Yo k Academy o Sciences, 1249, 118–136. B ingslima k, T., Ha ig, T., & Pa il, G. G. (2007). Psychological benefi s o indoo plan s in wo kplaces: Pu ing expe imen al esul s in o con ex . Ho Science, 42, 581–587. B ingslima k, T., Ha ig, T., & Pa il, G. G. (2009). The psychological benefi s o indoo plan s: A c i ical e iew o he expe imen al li e a u e. Jou nal o En i onmen al Psychology, 29, 422–433. B own, D. K., Ba on, J. L., P e y, J., & Gladwell, V. F. (2014). Walks4Wo k: Assessing he ole o he na u al en i onmen in a wo kplace physical ac i i y in e en ion. Scandina ian Jou nal o Wo k, En i onmen & Heal h, 40, 390–399. Chang, C.-Y., & Chen, P.-K. (2005). Human esponse o window iews and indoo plan s in he wo kplace. Ho Science, 40, 1354–1359. Dallne , M., Elo, A.-L., Gambe ale, F., Ho inen, V., Kna dahl, S., Linds öm, K., e al. (2000). Valida ion o he gene al No dic ques ionnai e (QPSNo dic) o psychological and social ac o s a wo k (No No d 2000 12). Copenhagen: No dic Council o Minis e s. De Bloom, J., Geu s, S. A. E., & Kompie , M. A. J. (2013). Vaca ion (a e -) e ec s on employee heal h and well-being, and he ole o aca ion ac i i ies: Expe iences and sleep. Jou nal o Happiness S udies, 14, 613–633. De Bloom, J., Kinnunen, U., & Ko pela, K. (2015). Reco e y p ocesses du ing and a e wo k: Associa ions wi h heal h, wo k engagemen : and job pe o mance. Jou nal o Occupa ional and En i onmen al Medicine, 57, 732–742. Degenha d , B., F ick, J., Buchecke , M., & Gu sche , H. (2011). Influences o pe sonal, social, and en i onmen al ac o s on wo kday use equency o he nea by ou doo ec ea ion a eas by wo king people. Leisu e Sciences, 33, 420–440. D a igne, A., Waliczek, T. M., Linebe ge , R. D., & Zajicek, J. M. (2008). The e ec o li e plan s and window iews o g een spaces on employee pe cep ions o job sa is ac ion. Ho Science, 43, 183–187. E ensen, K. H., Raanaas, R. K., Häge häll, C. M., Johansson, M., & Pa il, G. G. (2015). Res o a i e elemen s a he compu e wo ks a ion: A compa ison o li e plan s and inanima e objec s wi h and wi hou window iew. En i onmen and Beha io , 47, 288–303. Fjeld, T. (2000). The e ec o in e io plan ing on heal h and discom o among wo ke s and school child en. Ho Technology, 10, 46–52. Ga e sleben, B., & And ews, M. (2013). When walking in na u e is no es o a i e −The ole o p ospec and e uge. Heal h & Place, 20, 91–101. Geo ge, J. M., & Zhou, J. (2001). When openness o expe ience and conscien iousness a e ela ed o c ea i e beha io : An in e ac ional app oach. Jou nal o Applied Psychology, 86, 513–524. Gilch is , K., B own, C., & Mon a zino, A. (2015). Wo kplace se ings and wellbeing: G eenspace use and iews con ibu e o employee wellbeing a pe i-u ban business si es. Landscape and U ban Planning, 138, 32–40. Gladwell, V. F., B own, D. K., Ba on, J. L., Ta ainen, M. P., Kuoppa, P., P e y, J., e al. (2012). The e ec s o iews o na u e on au onomic con ol. Eu opean Jou nal o Applied Physiology, 112, 3379–3386. G oenewegen, P. P., an den Be g, A. E., de V ies, S., & Ve heij, R. A. (2006). Vi amin G: E ec s o g een space on heal h, well-being, and social sa e y. BMC Public Heal h, 6, 149. Ha ig, T., E ans, G. W., Jamne , L. D., Da is, D. S., & Gä ling, T. (2003). T acking es o a ion in na u al and u ban field se ings. Jou nal o En i onmen al Psychology, 23, 109–123. Jaussi, K. S., Randel, A. E., & Dionne, S. D. (2007). I am, I hink I can, and I do: The ole o pe sonal iden i y, sel -e ficacy: and c oss-applica ion o expe iences in c ea i i y a wo k. C ea i i y Resea ch Jou nal, 19, 247–258. Jiang, B., Li, D., La sen, L., & Sulli an, W. C. (2014). A dose- esponse cu e desc ibing he ela ionship be ween u ban ee co e densi y and sel - epo ed s ess eco e y. En i onmen & Beha io , h p://dx.doi.o g/10.1177/ 0013916514552321 Kaplan, R. (1993). The ole o na u e in he con ex o he wo kplace. Landscape and U ban Planning, 26, 193–201. Ka asek, R., & Theo ell, T. (1990). Heal hy wo k: s ess, p oduc i i y and he econs uc ion o wo king li e. New Yo k: Basic Books. Ke , J. H., & Tacon, P. (1999). Psychological esponses o di e en ypes o loca ions and ac i i ies. Jou nal o En i onmen al Psychology, 19, 287–294. Kinnunen, U., & Feld , T. (2013). Job cha ac e is ics: eco e y expe iences and occupa ional well-being: Tes ing c oss-lagged ela ionships ac oss 1 yea . S ess and Heal h, 29, 369–382. Knigh , C., & Haslam, S. A. (2010). The ela i e me i s o lean, en iched, and empowe ed o fices: An expe imen al examina ion o he impac o wo kspace managemen s a egies on well-being and p oduc i i y. Jou nal o Expe imen al Psychology: Applied, 16, 158–172. Ko pela, K., & Kinnunen, U. (2011). How is leisu e ime in e ac ing wi h na u e ela ed o he need o eco e y om wo k demands? Tes ing mul iple media o s. Leisu e Sciences, 33, 1–14. Ko pela, K. M., Ha ig, T., Kaise , F. G., & Fuh e , U. (2001). Res o a i e expe ience and sel - egula ion in a o i e places. En i onmen and Beha io , 33, 572–589. La go-Wigh , E., Chen, W. W., Dodd, V., & Weile , R. (2011). Heal hy wo kplaces: The e ec s o na u e con ac a wo k on employee s ess and heal h. Public Heal h Repo s, 126(Suppl. 1), 124–130. Leh o, X. Y. (2013). Assessing he pe cei ed es o a i e quali ies o aca ion des ina ions. Jou nal o T a el Resea ch, 52, 325–339. Lo up, L., S igsdo e , U. K., Meilby, H., & Claudi, A. G. (2013). The wo kplace window iew: A de e minan o o fice wo ke s’ wo k abili y and job sa is ac ion. Landscape Resea ch, h p://dx.doi.o g/10.1080/01426397.2013. 829806 Lucas, R. E., & Donnellan, M. B. (2007). How s able is happiness?: Using he s a s model o es ima e he s abili y o li e sa is ac ion. Jou nal o Resea ch in Pe sonali y, 41, 1091–1098.