Recovery during lunch breaks : Testing long-term relations with energy levels at work

Sianoja, M., e al. (2016). Reco e y du ing Lunch B eaks: Tes ing Long-Te m Rela ions wi h Ene gy Le els a Wo k. Scandina ian Jou nal o Wo k and O ganiza ional Psychology , 1(1): 7, 1–12, DOI: h p://dx.doi.o g/10.16993/sjwop.13 * Uni e si y o Tampe e, FI † Radboud Uni e si y Nijmegen, NL Co esponding au ho : Ma jaana Sianoja (ma jaana.siano[email p o ec ed]) ORIGINAL ARTICLE Reco e y du ing Lunch B eaks: Tes ing Long-Te m Rela ions wi h Ene gy Le els a Wo k Ma jaana Sianoja*, Ulla Kinnunen*, Jessica de Bloom*, Kale i Ko pela* and Sabine Geu s† This s udy had wo aims. Fi s , we examined whe he lunch b eak se ings, ac i i ies, and eco e y expe iences we e associa ed wi h lunch ime eco e y c oss-sec ionally. Second, we in es iga ed whe he lunch ime eco e y was ela ed o ene gy le els (i.e., exhaus ion and igo ) ac oss a 12-mon h pe iod. We collec ed longi udinal ques ionnai e da a among 841 Finnish wo ke s (59% emale, mean age 47 yea s) om 11 di e en o ganiza ions in a ious ields a wo ime poin s (sp ing 2013 and 2014). We used hie a chical eg ession analysis o es ou hypo heses. We ound ha eco e y expe iences, ha is, psychological de achmen om wo k and con ol du ing he lunch b eak, we e ela ed o success ul lunch ime eco e y. A e con olling o backg ound ac o s, main job cha ac e is ics (wo kload and au onomy), and he ou comes a baseline, success ul lunch ime eco e y was ela ed o a dec ease in exhaus ion and o an inc ease in igo one yea la e . To conclude, lunch b eaks o e an impo an se ing o in e nal eco e y du ing wo king days and seem o ela e o ene gy le els a wo k o e ime. Keywo ds: lunch b eaks; eco e y; de achmen ; con ol; exhaus ion; igo In oduc ion Reco e y om wo k s ess, ha is, psycho-physiological unwinding a e e o expendi u e a wo k ha es o es employees’ ene gy and men al esou ces, is a mechanism explaining how employees can p o ec hei well-being and heal h in demanding wo king condi ions (C aig & Coope , 1992; Geu s & Sonnen ag, 2006; Meijman & Mulde , 1998). Reco e y plays an in e ening ole in he ela ionship be ween s ess ul job cha ac e is ics and he de elopmen o ch onic load eac ions, such as p olonged a igue, sleep diso de s, and ca dio ascula diseases (Geu s & Sonnen ag, 2006). The e o e, a mo e p o ound unde s anding o eco e y p ocesses is essen ial in p o- mo ing sus ainable wo king li e. Reco e y occu s du ing b eaks om wo k when job demands a e no longe p esen (Meijman & Mulde , 1998). Di e en o ms o b eaks ange om sabba icals and aca ions o sho mic o-b eaks wi hin he wo king day. Reco e y wi hin wo king days, e e ed o as in e nal eco e y, has ecei ed a less a en ion in he eco e y esea ch li e a u e han o -job eco e y, e e ed o as ex e nal eco e y (Sonnen ag & F i z, 2015). Al hough eco e y du ing b eaks wi hin he wo king day may no be as sel -e iden as eco e y du ing leisu e ime, in e nal eco e y has po en ial in p e en ing s ess om accumula ing ea ly on, helping o main ain pe o - mance h oughou he day and p e en ing high need o eco e y a he end o he wo king day (Co eng, an Sluijs, Hend iksen, an Mechelen, & Boo , 2015; Geu s, Becke s, & Tucke , 2014). As wo ke s ypically spend a hi d o a hal o hei day a he wo kplace i is impo an o ecognize he eco e y po en ial o wi hin wo king day b eaks, and especially o he lunch b eak, which is ypically he longes and mos common o b eaks in he cou se o he wo king day. Fu he mo e, o ganiza ions ha e a g ea e oppo uni y o in luence employees’ in e nal eco e y han ex e nal eco e y and, he e o e, lunch b eaks as a eco e y se - ing may be o special in e es o employe s. Fo example, o ganiza ions may encou age egula lunch b eaks and p o ide es o a i e en i onmen s (e.g., quie ooms o elaxa ion). The ques ion o how o eco e success ully du ing lunch b eaks has ecen ly gained some esea ch a en ion (B own, Ba on, P e y, & Gladwell, 2014; K ajewski, Wieland, & Saue land, 2010; T ougakos, Hideg, Cheng, & Beal, 2014). Ne e heless, esea ch on in e nal eco e y is s ill sca ce (Sonnen ag & F i z, 2015). In his s udy we examine which lunch ime se ings, ac i i ies, and eco e y expe iences a e ela ed o lunch- ime eco e y (i.e., how o en employees ecupe a e success ully om wo k du ing lunch b eaks) in a c oss- sec ional sample (S udy 1). Fu he mo e, we es whe he lunch ime eco e y is ela ed o ene gy le els a wo k, ha is, exhaus ion and igo , o e a 12-mon h pe iod (S udy 2). Ou s udy con ibu es o he li e a u e on wo k s ess Sianoja e al: Lunch ime Reco e y and Long-Te m Ou comesA . 7, page 2 o 12 eco e y by ex ending he unde s anding o how o p o- mo e in e nal eco e y and explo ing i s ela ion wi h main aining ene gy a wo k. Figu e 1 p esen s he model o he s udy wi h hypo hesized ela ionships. Reco e y du ing lunch b eak: Theo e ical and empi ical pe spec i es In he e o - eco e y (E-R) model (Meijman & Mulde , 1998) eco e y has been de ined as a p ocess o he psycho-biological sys em e u ning o i s p e-s esso le el. Reco e y occu s when he demands causing s ain a e no longe p esen (Meijman & Mulde , 1998). When eco e y is insu icien , an indi idual has o in es addi- ional e o a wo k, which may cause s ain and lead o accumula ing s ain eac ions in he long e m. Reco e y he e o e plays a signi ican ole in coun e ac ing s ain caused by job demands and helps in main aining well-being and ene gy a wo k. Besides seeing eco e y as a passi e p ocess (i.e., caused by me e absence o demands), ac i e pe spec i es on eco e y ha e also been in oduced. Acco ding o conse - a ion o esou ces (COR) heo y (Hob oll, 2002; Hob oll, 1989), people a e mo i a ed o gain new esou ces and p o ec hei exis ing esou ces. Resou ces a e de ined b oadly as “objec s, pe sonal cha ac e is ics, condi ions, o ene gies ha a e alued by he indi idual” (Hob oll, 1989, p. 516). When esou ces a e los , h ea ened wi h loss, o new esou ces a e no gained a e e o in es - men , s ain occu s. Du ing b eaks om wo k, people ha e he oppo uni y o engage in pleasan ac i i ies and o egain esou ces (e.g., ene gy and posi i e mood). Thus, o eco e du ing a b eak, a b eak mus ensu e absence o job demands and p o ide an oppo uni y o employees o egain alued esou ces (Hob oll, 2002; Meijman & Mulde , 1998). This also implies ha b eaks should be egula and long enough o allow enough ime o eco e y. Addi ionally, b eak loca ion ( he place whe e he b eak is spen ), b eak ac i i ies, and expe iences du ing he b eak may in luence i s eco e y po en ial as hey a e closely ela ed o he absence o job demands and oppo - uni ies o esou ce gain. These aspec s ha e been a gued (Sonnen ag & Na e , 2004) and shown (Sonnen ag & F i z, 2015; Sonnen ag & Zijls a, 2006; an Hoo & Baas, 2013) o be o impo ance in e ms o eco e y du ing lei- su e ime. Some o hese (e.g., ac i i ies) ha e also been iden i ied in ea lie esea ch as impo an aspec s o in e - nal eco e y (see Sianoja, Kinnunen, De Bloom, & Ko pela, 2015). When looking a eco e y esea ch on whe e b eaks a e spen , a ecen 5-day dia y s udy ound no di e ence be ween spending b eaks inside o ou side he o ice in e ms o esou ce eco e y (Hun e & Wu, 2016). Howe e , in his s udy ca ied ou among 95 uni e si y s a mem- be s, he ou side condi ion also included di e en spaces inside he o ice building (e.g., a b eak oom), which may no o e as bene icial condi ions o de achmen om wo k as spaces ou side he o ice building (e.g., a ca é o es au an ). O he s udies ha e been speci ically in e es ed in na u al en i onmen s. Acco ding o an in e en ion s udy by B own e al. (2014), spending one’s lunch ime walking in a na u al en i onmen was bene icial in e ms o imp o ed men al heal h when compa ed o walking in buil en i onmen s. Acco dingly, his s udy sugges s ha Figu e 1: Model o he s udy. Sianoja e al: Lunch ime Reco e y and Long-Te m Ou comes A . 7, page 3 o 12 b eak loca ion may be signi ican in e ms o eco e y. To conclude, spending he lunch b eak ou side he o ice building should, in heo y, aid eco e y, as i ensu es be e men al de achmen om wo k o e ing a “change o scene y” whe e job demands a e no p esen (e.g., Ko pela, De Bloom, & Kinnunen, 2015). Conce ning b eak ac i i ies, ea lie s udies on in e nal eco e y ha e associa ed elaxing, physical, and social ac i i ies wi h posi i e eco e y ou comes (Co eng e al., 2015; K ajewski e al., 2010; T ougakos e al., 2014). O hese, we ocused on social ac i i ies. Wendsche e al. (2014) showed ha collec i e es b eaks (i.e., b eaks including social ac i i ies) we e associa ed wi h less u no e han b eaks spen alone. In addi ion, a s udy by T ougakos e al. (2014) ocusing on di e en lunch b eak ac i i ies e ealed ha social ac i i ies ha we e based on indi iduals’ own choice, we e conduci e o eco e y. In sum, in addi ion o absence o job demands, as sug- ges ed by he E-R model, ea lie esea ch shows ha eco e y may be also enhanced by engaging in ac i i ies ha enable esou ce eplenishmen , as sugges ed by he COR heo y. Hypo hesis 1: a) Ha ing lunch b eaks egula ly, b) ha - ing longe lunch b eaks, c) spending lunch b eaks ou side he o ice building and d) spending lunch b eaks wi h o h- e s a e posi i ely associa ed wi h eco e y du ing lunch b eaks. Fu he mo e, i has been a gued ha a eco e ing b eak should p omo e eco e y expe iences (Co eng e al., 2015; T ougakos e al., 2014). Acco ding o Sonnen ag and F i z (2007), he e a e ou such mecha- nisms: psychological de achmen , elaxa ion, mas e y, and con ol. O hese, we examined de achmen , ha is, no hinking abou wo k, and con ol, ha is, ge ing o choose how o spend one’s ee ime (e.g., lunch b eaks). These wo expe iences we e chosen as hey ha e gained mos suppo in ea lie s udies. In s udies ocusing on eco e y du ing leisu e ime, de achmen has been iden i ied as a co e eco e y expe ience (Sonnen ag & F i z, 2015). Psychological de achmen om wo k, in addi ion o physical de achmen , is c ucial, as con inuing o hink abou job demands du ing b eaks may esul in s ain (Sonnen ag & F i z, 2007). In ac , in a c oss- sec ional s udy de achmen du ing wo k b eaks was connec ed o less need o eco e y a he end o he day (Co eng e al., 2015). Fu he mo e, au onomy (i.e., con ol) du ing lunch b eaks has p e iously been linked o bene icial ou comes (T ougakos e al., 2014). Mo e speci ically, au onomy du ing lunch b eaks was ecog- nized as a mode a o be ween lunch b eak ac i i ies and eco e y ou comes: au onomy s eng hened he posi i e e ec s o he ac i i ies. In addi ion, p e e ed wo k b eak ac i i ies ha e been associa ed wi h inc eased esou ces a e he b eak (Hun e & Wu, 2016). The e o e b eak cha ac e is ics ha enhance psychological de achmen om wo k and allow con ol, may p o ide bene icial se ing o eco e y. Hypo hesis 2: Reco e y expe iences (de achmen and con ol) du ing lunch b eaks a e posi i ely associa ed wi h eco e y du ing lunch b eaks. Long- e m associa ions be ween lunch ime eco e y and ene gy le els a wo k As long- e m ou comes o eco e y we ocused on ene gy, speci ically on exhaus ion and igo a wo k. Acco ding o he E-R model (Meijman & Mulde , 1998), when eco e y is insu icien , high and con inuous demands lead o nega- i e load e ec s and deple ion o ene gy, which in he long e m can lead o emo ional exhaus ion. Emo ional exhaus- ion is one o he co e bu nou dimensions and e e s o “ eelings o being o e ex ended and deple ed o one’s emo ional and physical esou ces” (Maslach, Schau eli, & Lei e , 2001, p. 399). Resea ch has shown ha emo ional exhaus ion p edic s men al and physical illness, such as dep ession and ca dio ascula diseases (Ahola, 2007), as well as inc eased sickness absence (Toppinen-Tanne , Ojajä i, Väänänen, Kalimo, & Jäppinen, 2005). Hun e and Wu (2016) ound ha esou ce eco e y du ing wo kday b eaks ac oss one wo king week was asso- cia ed wi h lowe le els o exhaus ion a he end o he week. As a as we know, he long- e m e ec s be ween poo eco e y du ing lunch b eaks and exhaus ion ha e no ye been examined. Howe e , o e ime employees go h ough nume ous cycles o daily lunch ime eco e y p o- cesses, which may ul ima ely esul in ei he gain o loss o ene gy depending on whe he eco e y is success ul o incomple e. The e o e, insu icien eco e y may, o e ime, esul in cumula i e esou ce loss in e ms o highe exhaus ion. Hypo hesis 3: Insu icien eco e y du ing lunch b eaks is ela ed o high le el o emo ional exhaus ion o e ime. In con as , success ul eco e y ensu es ha ene gy le els a e su icien o people o expe ience igo a wo k (Meijman & Mulde , 1998). Vigo is one o he co e dimensions o wo k engagemen and is cha ac e ized by high ac i a ion, ene gy, and men al esilience while wo king (Schau eli, Salano a, González-Romá, & Bakke , 2002). Wo k engagemen , and pa icula ly igo , has been shown o be impo an in e ms o mo i a ion and pe o - mance a wo k (Bakke , Deme ou i, & Sanz-Ve gel, 2014). I has also been shown ha exhaus ion and igo a e no endpoin s o he same ene gy cons uc (Deme ou i, Mos e , & Bakke , 2010; Mäkikangas, Feld , Kinnunen, & Tol anen, 2012). Thus we canno conclude ha absence o exhaus ion au oma ically implies high le els o igo . I is he e o e impo an o measu e bo h when examining he ene gy le els o indi iduals. To he bes o ou knowledge, s udies on in e nal eco e y and i s ela ion o igo a e so a lacking. Howe e , on a daily le el aking mic o-b eaks a wo k has been associa ed wi h i ali y, a concep ela ed o igo (Zache , B ails o d, & Pa ke , 2014). Fu he mo e, ea lie esea ch has es ablished a posi i e link be ween ex e nal eco e y and wo k engagemen (Kühnel, Sonnen ag, & Wes man, 2009; Sonnen ag, 2003). I eco e y is epea edly insu icien du ing lunch b eaks, i may lead o loss o ene gy and o e ime educe igo . In addi ion, eco e y is associa ed wi h esou ce gain (e.g., ene gy), and esou ces end o accumula e and gene a e o he esou ces in he long e m (Hob oll, 2002). Acco dingly, success ul Sianoja e al: Lunch ime Reco e y and Long-Te m Ou comesA . 7, page 4 o 12 eco e y du ing lunch b eaks may esul in ene gy gain, esul ing in highe le els o igo o e ime. Hypo hesis 4: Success ul eco e y du ing lunch b eaks is ela ed o high le el o igo o e ime. Me hods Pa icipan s and P ocedu e The da a we e collec ed as a pa o la ge p ojec on eco e y om wo k (see Kinnunen e al., 2016). The pa - icipan s o his s udy we e Finnish employees wo king in 11 di e en o ganiza ions in a ious ields, mos ly wo k- ing in cogni i ely o emo ionally demanding jobs. The mos common ields we e educa ion, in o ma ion ech- nology, and media. The ques ionnai e da a we e collec ed in wo phases. Fi s , 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 o he employees by ou con ac pe sons (in ou o ganiza ions). O he employees con ac ed (N = 3,593), 1,347 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) he elec onic ques ionnai e was sen o hose employees’ e-mail add esses who esponded in 2013 and who we e s ill employed in he same o ganiza ions (N = 1,192). 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 we e in o med abou he goals o he s udy, assu ed ha esponses would be ea ed con iden ially and eminded ha pa icipa ion was olun a y. In S udy 1, we used he c oss-sec ional sample col- lec ed a T2, because no all a iables (i.e., spending lunch b eaks ou side, spending b eaks wi h o he s, lunch ime de achmen , and lunch ime con ol) we e measu ed a T1. A c oss-sec ional design was conside ed app op ia e because we we e in e es ed in he immedia e ela ions o b eak se ings, ac i i ies, and expe iences wi h lunch ime eco e y. S udy 2 was based on he longi udinal sample co e ing bo h measu emen s wi h a 12-mon h ime lag be ween he measu emen s. I is di icul o heo e ically de e mine he mos app op ia e ime lag as we lack he- o ies o change, and he e o e e en desc ip i e esea ch on he ime cou ses o impo an ela ionships has been ecommended (Kelloway & F ancis, 2013). We conside one yea o be an app op ia e ime lag, as i is so a he mos ypical ime pe iod used in ea lie eco e y s ud- ies showing long- e m e ec s (Kinnunen & Feld , 2013; Sil aloppi, Kinnunen, Feld , & Tol anen, 2011; Sonnen ag, Binnewies, & Mojza, 2010). Addi ionally, he eali y o da a collec ion in o ganiza ions imposed ce ain limi a- ions. We we e no able o schedule measu emen s mo e equen ly because we had o conside he o ganiza ions’ wishes and ime cons ain s. O he sample used in bo h s udies (N = 841), 58.6% we e women. The pa icipan s’ a e age age was 47.1 yea s ( ange 21–67, SD = 10.0). Mos o he pa icipan s (76.4%) we e li ing wi h a pa ne (ei he ma ied o cohabi - ing), and 45.6% had child en (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 quali ica ion o less. O he pa icipan s, 8.3% we e blue-colla wo ke s (e.g., cleane s), 30.0% lowe whi e-colla wo ke s (e.g., o ice wo ke s), 57.8% senio whi e-colla wo ke s (e.g., eache s) and 3.8% senio -le el manage s (e.g., chie execu i e o ice 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 weekly wo king hou s we e 39.1 (SD = 5.9). O he pa icipan s, 53.6% wo ked in he public sec o as eache s o adminis a i e s a in oca ional o uppe seconda y schools, o in a poly echnic (uni e si y o applied sciences). The es (46.4%) wo ked in he p i a e sec o in a ious jobs. In analyzing sample a i ion we compa ed he espond- en s (n = 841) o he longi udinal sample wi h he non- esponden s. The esul s indica ed ha he esponden s did no di e om he non- esponden s in e ms o gen- de , age, ha ing a pa ne , numbe o child en o le el o educa ion. They also did no di e in e ms o he s udy a iables measu ed a bo h ime poin s ( egula i y o aking lunch b eaks, lunch b eak leng h, lunch ime eco - e y, exhaus ion, o igo ). Howe e , he esponden s we e mo e o en employed as senio whi e-colla wo ke s (58% s. 50%) han he non- esponden s (p < .05) and mo e o en on a pe manen job con ac (89% s. 79%) han he non- esponden s (p < .001). Also, he esponden s wo ked mo e hou s pe week (39.1 s. 37.9 hou s, p < .01) and mo e o en on egula day shi s (90% s. 83%, p < .01) han he non- esponden s. As we used he da a collec ed a T2 in ou c oss-sec ional s udy, his sample a i ion conce ns bo h S udy 1 and S udy 2. Measu es Reco e y du ing lunch b eaks To measu e he deg ee o eco e y du ing lunch b eaks a T1 and T2, we used one i em “I ecupe a e om wo k du ing my lunch b eak” om he Reco e y a e B eaks Scale (Deme ou i, Bakke , Sonnen ag, & Fullaga , 2012) aiming o cap u e speci ically how well and egula ly employees eco e du ing hei lunch b eaks. The i em was a ed on a scale om 1 ( e y seldom o ne e ) o 5 ( e y o en o always). Ea lie s udies ha e p o ided sup- po o he alidi y o single i em measu es (e.g., D ole & Mo ison, 2001; Elo, Leppänen, & Jahkola, 2003). Con- ce ning eco e y, i has been shown ha eco e y om wo k measu ed wi h one i em co ela ed highly wi h longe eco e y scales, such as need o eco e y (Kinnunen, Feld , Sil aloppi, & Sonnen ag, 2011). B eak se ings, ac i i ies and expe iences O b eak se ings and ac i i ies, we measu ed egula i y o lunch b eaks [dicho omized o 0 = occasionally (1–3 imes a week), 1 = egula ly (4–5 imes a week)] and leng h o he lunch b eak (in minu es). Those pa icipan s (n = 36 a T1 and n = 32 a T2) who epo ed no aking lunch b eaks, we e no asked o answe any u he lunch b eak ela ed ques ions ( eco e y du ing lunch b eak, b eak ac i i ies, o expe iences) and as lunch b eak eco e y was he main ocus in ou s udy, hey we e excluded om he analyses. Sianoja e al: Lunch ime Reco e y and Long-Te m Ou comes A . 7, page 5 o 12 In addi ion a T2, we asked whe he he employees habi ually spen hei lunch b eaks ou side he o ice building [“I spend my lunch b eak ou side my company building (e.g., in a es au an o in a ca é)”] o wi h o he s [“I spend my lunch b eak wi h o he s (e.g., wi h colleagues, acquain ances, iends o amily membe s)”]. The answe s we e dicho omized [0 = no (ha dly e e o once a week), 1 = yes (2–5 imes a week)]. O eco e y expe iences, we measu ed de achmen and con ol du ing lunch ime a T2. Bo h de achmen and con ol we e measu ed wi h one i em ( espec i ely: “I dis- ance mysel men ally om my wo k du ing lunch b eaks” and “I decide mysel how o spend my lunch b eaks”) om he Finnish e sion o he Reco e y Expe ience Ques ionnai e (Kinnunen e al., 2011; Sonnen ag & F i z, 2007). The i ems we e adap ed o conce n lunch b eaks and measu ed on a scale om 1 ( e y seldom o ne e ) o 5 ( e y o en o always). Po en ial long- e m ou comes Emo ional exhaus ion was measu ed a T1 and T2 wi h he i e-i em scale (e.g., “I eel emo ionally d ained om my wo k”) om he Maslach Bu nou In en o y (Kalimo, Hakanen, & Toppinen-Tanne , 2006; Maslach, Jackson, & Lei e , 1996) wi h esponse op ions on a se en-poin esponse scale om 0 (ne e ) o 6 (e e y day). The C on- bach’s alphas we e .93 a T1 and .93 a T2. Vigo was measu ed a T1 and T2 wi h he h ee-i em sho ened scale (e.g., “A my wo k, I eel bu s ing wi h ene gy”) om he U ech Wo k Engagemen Scale (Schau eli, Bakke , & Salano a, 2006) using a se en-poin esponse scale anging om 0 (ne e ) o 6 (e e y day). The C onbach’s alphas we e .89 a T1 and .90 a T2. Con ol a iables O he backg ound ac o s, we con olled o age (in yea s), gende (1 = woman, 2 = man) and wo king hou s pe week, as hese may play a ole in eco e y (e.g., Moh en, Jansen, & Kan , 2010; Sil aloppi e al., 2011). Wo king hou s we e measu ed wi h a single ques ion: “How many hou s do you ac ually wo k pe week? (Include paid and unpaid o e ime, bu no you commu ing ime)”. We also con olled o main job cha ac e is ics, namely wo kload and au onomy, measu ed a T1 and T2 as hey may ac as con ounding a iables in ou s udy. Fi s , app op ia e job design may abo e all p omo e in e nal eco e y (Geu s e al., 2014) as i enables he employees o adjus hei wo k acco ding o hei cu en need o eco e y. Fu he mo e, job demands and esou ces play a pi o al ole in main aining ene gy as job demands may s a a heal h de e io a ing p ocess leading o exhaus ion, and job esou ces, in u n, o a heal h p omo ing p o- cess leading o an inc ease in igo (Bakke e al., 2014). 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 es- su e?”, C onbach’s alphas .88 a T1 and .87 a T2) om he Quan i a i e Wo kload In en o y (Spec o & Jex, 1998). Au onomy was measu ed wi h i e i ems (e.g., “I can in lu- ence decisions ha a e impo an o my wo k”, C onbach’s alphas .77 a T1 and .78 a T2) om he QPS No dic-ADW (Dallne e al., 2000). All job cha ac e is ics we e meas- u ed wi h a i e-poin scale om 1 ( e y seldom o ne e ) o 5 ( e y o en o always). Analyses In bo h s udies (S udies 1 and 2), we used hie a chical eg ession analyses o es ou hypo heses. In he c oss- sec ional S udy 1 lunch ime eco e y se ed as a depend- en a iable. A he i s s ep, we added he con ol a i- ables (age, gende , weekly wo king hou s, wo kload, and au onomy). A he second s ep we added a iables desc ib- ing lunch b eak se ings and ac i i ies ( egula i y o he lunch b eaks, leng h o he lunch b eak, b eak ou side, and wi h o he s). Finally, a he hi d s ep we added eco - e y expe iences (de achmen and con ol) du ing lunch b eaks. In he longi udinal S udy 2, we ollowed simila s eps wi h bo h ou comes (exhaus ion and igo ). A he i s s ep, we con olled o he ou come a Time 1. A he sec- ond s ep, we added con ol a iables (age, gende , weekly wo king hou s, wo kload, and au onomy). Lunch ime eco e y a Time 1 was added a he inal s ep, as we we e in e es ed in i s explana o y powe a e con olling o he ou come a Time 1, backg ound ac o s and wo k cha ac e is ics. Resul s Desc ip i e esul s Means, s anda d de ia ions, and ze o-o de co ela ions o he s udy a iables a e p esen ed in Table 1 (S udy 1) and Table 2 (S udy 2). We i s looked a he equencies o lunch b eak cha ac e is ics examined in S udy 1, in which all a iables we e measu ed a T2. To ha e a egula lunch b eak was common in ou sample, as 86% o he pa ici- pan s epo ed aking a lunch b eak 4–5 imes a week. O hose pa icipan s who ook lunch b eaks a leas once a week, 37% epo ed habi ually spending he b eak ou - side he o ice building and 71% wi h o he people. In S udy 1, o he lunch b eak se ings and ac i i ies, egula lunch b eaks ( = .21), longe lunch b eaks ( = .16), b eaks ou side he o ice building ( = .17), and b eaks wi h o h- e s ( = .08) showed posi i e associa ions wi h lunch ime eco e y. Howe e , bo h eco e y expe iences – de ach- men and con ol – du ing lunch b eaks showed he s onges co ela ions: high le el o de achmen ( = .59) and con ol ( = .30) du ing lunch b eaks we e associa ed wi h success ul lunch ime eco e y. In addi ion, o he con ol a iables, wo kload was nega i ely ( = −.12) and au onomy posi i ely ( = .33) associa ed wi h eco e y du - ing lunch b eaks. In S udy 2 he e we e signi ican longi udinal co ela- ions be ween lunch ime eco e y and bo h po en ial long- e m ou comes (Table 2). Lunch ime eco e y a T1 was nega i ely ela ed o exhaus ion a T2 ( = −.35) and posi i ely ela ed o igo a T2 ( = .36). O he con ol a - iables gende ( emale), long weekly wo king hou s, high wo kload and low au onomy we e ela ed o exhaus ion a T2, and gende ( emale) and high le el o au onomy we e ela ed o igo a T2. In addi ion, lunch ime eco e y ( = .48) and bo h ou comes ( = .69 o exhaus ion and Sianoja e al: Lunch ime Reco e y and Long-Te m Ou comesA . 7, page 6 o 12 M / % SD 1 2 3 4 5 6 7 8 9 10 11 1. Lunch ime eco e y 3.42 0.97 – 2. Age 48.13 10.02 .04 – 3. Gende 158.6% – .03 .04 – 4. Weekly wo king hou s 38.65 5.90 −.03 −.06 .07* − 5. Wo kload 3.82 0.79 −.12*** −.07 −.15*** .27*** – 6. Au onomy 3.20 0.80 .33*** −.06 .16*** .05 −.24*** – 7. Regula i y o lunch b eaks286.2% – .21*** −.13*** −.04 .04 −.05 .10** – 8. Leng h o lunch b eak 29.02 8.55 .16*** −.05 .11** −.02 −.05 .06 .08* − 9. Lunch b eak ou side336.9% – .17*** −.04 .11** .12** .08* .14*** .15*** .15*** – 10. Lunch b eak wi h o he s471.4% – .08* −.09** −.13*** .03 .02 .05 .21*** .07* .12** – 11. De achmen a lunch b eak 3.33 1.06 .59*** .01 .01 −.09* −.16*** .18*** .21*** .18*** .19*** .03 – 12. Con ol a lunch b eak 4.27 0.96 .30*** .03 .11** .03 −.05 .38*** .05 .19*** .22*** .05 .24*** Table 1: Means, s anda d de ia ions, and ze o-o de co ela ions o he s udy a iables in S udy 1. No e. 1Gende : 1 = emale, 2 = male; 2Regula i y o lunch b eaks: 0 = occasionally (1–3 imes a week), 1 = egula ly (4–5 imes a week); 3B eak ou side: 0 = no (ha dly e e o once a week), 1 = yes (2–5 imes a week); 4B eak wi h o he s: 0 = no (ha dly e e o once a week), 1 = yes (2–5 imes a week). The second column shows pe cen ages o ca ego ical a iables: 1 % o emale pa icipan s; 2 % o pa icipan s aking lunch b eaks egula ly; 3 % o pa icipan s ypically spending lunch b eaks ou side; 4 % o pa icipan s ypically spending lunch b eaks wi h o he s. * p < .05; ** p < .01; *** p < .001; 807 < N < 841. Sianoja e al: Lunch ime Reco e y and Long-Te m Ou comes A . 7, page 7 o 12 M / % SD 1 2 3 4 5 6 7 8 9 1. Lunch ime eco e y T1 3.39 1.00 2. Age T1 47.13 10.02 .08* 3. Gende 158.6% – .01 .04 4. Weekly wo king hou s T1 39.09 5.94 −.09* −.01 .06 5. Wo kload T1 3.89 0.82 −.16*** .03 −.16*** .28*** 6. Au onomy T1 3.18 0.82 .30*** −.08* .16*** .05 −.30*** 7. Exhaus ion T1 1.92 1.45 −.41*** .04 −.11** .12** .36*** −.35*** 8. Vigo T1 4.53 1.21 .43*** .01 −.06 .05 .01 .26*** −.45*** 9. Exhaus ion T2 1.92 1.41 −.35*** −.02 −.15*** .08* .31*** −.30*** .69*** −.35*** 10. Vigo T2 4.37 1.32 .36*** .00 −.09** .07 .04 .19*** −.37*** .68*** −.45*** Table 2: Means, s anda d de ia ions, and ze o−o de co ela ions o he s udy a iables in S udy 2. No e. 1Gende : 1 = emale, 2 = male. The second column shows pe cen ages o ca ego ical a iables: 1 % o emale pa icipan s. * p < .05; ** p < .01; *** p < .001; 785 < N < 841. = .68 o igo ) we e ela i ely s able be ween T1 and T2. No mean le el changes occu ed in lunch ime eco e y o exhaus ion be ween T1 and T2. Howe e , igo was signi i- can ly lowe a T2 han a T1 (p < .001). Tes ing he hypo heses S udy 1 The esul s o he hie a chical eg ession analysis con- ce ning he associa ions be ween lunch b eak se ings, ac i i ies, eco e y expe iences, and lunch ime eco e y a e shown in Table 3. A s ep 1, con ol a iables (back- g ound a iables and job cha ac e is ics) explained 12% o he a iance in lunch ime eco e y and au onomy a wo k signi ican ly con ibu ed o lunch ime eco e y. Regula lunch b eaks, longe lunch b eaks and habi u- ally spending lunch b eaks ou side he o ice building con ibu ed o success ul lunch ime eco e y, inc eas- ing he explana ion a e o he model o 17%. Spend- ing lunch b eaks wi h o he s did no con ibu e o lunch ime eco e y. A e adding eco e y expe ience a iables o he model a s ep 3, only egula i y o he lunch b eaks (o he lunch ime cha ac e is ics en e ed a s ep 2) con inued o be associa ed wi h lunch ime eco - e y. Bo h de achmen and con ol we e posi i ely ela ed o eco e y, and hey aised he explana ion a e o he model o 41%. De achmen (β = .51, p < .001) p edic ed lunch ime eco e y mo e s ongly han con ol (β = .09, p < .01). In sum, Hypo hesis 1 was pa ially suppo ed, as mos o he posi i e e ec s o lunch ime se ings and ac i i ies disappea ed when lunch ime eco e y expe iences we e en e ed in o he model. Mo e speci ically, Hypo hesis 1a was ully suppo ed, as aking lunch b eaks egula ly con- ibu ed o success ul lunch ime eco e y. Hypo heses 1b and 1c we e pa ially suppo ed, as longe lunch b eaks and spending b eaks ou side we e only signi ican be o e eco e y expe iences we e en e ed in o he model. Hypo hesis 1d did no ecei e suppo , as spending lunch b eaks wi h o he s did no con ibu e o eco e y. Fu he mo e, Hypo hesis 2 was ully suppo ed, as bo h high le els o de achmen and con ol du ing lunch b eak con ibu ed o success ul lunch b eak eco e y. S udy 2 The esul s o hie a chical eg ession analyses explo ing he longi udinal ela ionships o lunch ime eco e y wi h exhaus ion and igo a e shown in Table 4. Conce ning exhaus ion, a s ep 1, exhaus ion a T1 s ongly p edic ed exhaus ion a T2 explaining 47% o he a iance. A he second s ep, adding he con ol a iables, he explana ion a e o he model inc eased by 1 %, as gende ( emale) was signi ican ly ela ed o exhaus ion. A he inal s ep, lunch ime eco e y a T1 con ibu ed signi ican ly (β = −.07) o exhaus ion a T2. The inc ease in he explana- ion a e was signi ican , al hough i inc eased only 0.3%. The explana ion a e o he inal model was 48%. Thus, in line wi h Hypo hesis 3, success ul eco e y a lunch b eaks seems o explain – o a mino deg ee – a dec ease in exhaus ion ac oss one yea . In he model p edic ing igo , a s ep 1, igo a T1 s ongly p edic ed igo a T2 explaining 47% o he a iance. A he second s ep, adding he con ol a iables nei he backg ound ac o s no job cha ac e is ics we e signi ican p edic o s o igo . A he inal s ep, lunch ime eco e y a T1 con ibu ed signi ican ly (β = .10) o igo a T2 and added 1% o he explana ion a e. The explana- ion a e o he inal model was 48%. Thus, in line wi h Hypo hesis 4 success ul eco e y a lunch b eaks seems o explain – o a mino deg ee – an inc ease in igo ac oss one yea . Discussion This s udy had wo main aims. Fi s , we in es iga ed whe he ce ain lunch b eak se ings, ac i i ies, and expe- iences we e ela ed o eco e y du ing lunch b eaks. Second, we examined whe he lunch ime eco e y was associa ed wi h ene gy le els a wo k one yea la e . We based ou s udy on he E-R model and he COR heo y. Sianoja e al: Lunch ime Reco e y and Long-Te m Ou comesA . 7, page 8 o 12 Lunch ime eco e y S ep 1 S ep 2 S ep 3 P edic o s β β β Age .06 .09* .06* Gende 1−.03 −.04 −.03 Weekly wo king hou s −.04 −.05 .00 Wo kload −.04 −.04 .01 Au onomy .33*** .30*** .22*** Regula i y o lunch b eaks2.15*** .07* Leng h o lunch b eak .11** .03 Lunch b eak ou side3.11** .01 Lunch b eak wi h o he s4−.00 .02 De achmen a lunch b eak .51*** Con ol a lunch b eak .09** ΔR2.12*** .06*** .24*** R2.12*** .17*** .41*** Table 3: Resul s o hie a chical eg ession analysis o lunch ime eco e y (S udy 1), N = 774. No e. 1Gende : 1 = emale, 2 = male; 2Regula i y o lunch b eaks: 0 = occasionally (1–3 imes a week), 1 = egula ly (4–5 imes a week); 3B eak ou side: 0 = no (ha dly e e o once a week), 1 = yes (2–5 imes a week); 4B eak wi h o he s: 0 = no (ha dly e e o once a week), 1 = yes (2–5 imes a week). * p < .05; ** p < .01; *** p < .001. Model 1 Model 2 Exhaus ion Vigo S ep 1 S ep 2 S ep 3 S ep 1 S ep 2 S ep 3 P edic o s a T1 β β β β β β Dependen a iable a T11.68*** .65*** .63*** .69*** .68*** .64*** Age T1 −.04 −.03 −.03 −.04 Gende 2−.07* −.07* −.05 −.05 Weekly wo king hou s T1 −.02 −.02 .04 .05 Wo kload T1 .05 .06 .02 .03 Au onomy T1 −.03 −.02 .03 .01 Lunch ime eco e y T1 −.07* .10** ΔR2.47*** .01** .003* .47*** .01 .01** R2.47*** .48*** .48*** .47*** .48*** .49*** Table 4: Resul s o hie a chical eg ession analysis o exhaus ion (Model 1) and igo (Model 2) a T2 (S udy 2), N = 745. No e. 1Dependen a iable a T1: Fo he i s model Exhaus ion a T1, o he second model Vigo a T1. 2Gende : 1 = emale, 2 = male. * p < .05; ** p < .01; *** p < .001. Among ou sample o Finnish wo ke s, ha ing lunch b eaks was common, as 86% o he pa icipan s ook hem 4–5 imes a week. On a e age, he pa icipan s el occa- sionally eco e ed a e hei lunch b eaks and no changes in his ega d we e obse ed ac oss one yea . In line wi h ou expec a ions, o he b eak se ings o ac i i ies, egu- la i y o he lunch b eaks, leng h o he lunch b eak and spending lunch b eaks ou side he o ice con ibu ed o success ul lunch ime eco e y. Thus, ou s udy suppo s he impo ance o aking egula lunch b eaks. Howe e , associa ions be ween b eak leng h and b eaks ou side we e no longe signi ican a e aking eco e y expe i- ences in o accoun . As expec ed, we ound ha highe le els o de achmen and con ol du ing lunch b eaks we e ela ed o mo e success ul lunch ime eco e y. This inding concu ed wi h ea lie esea ch on in e nal eco - e y (Co eng e al., 2015; T ougakos e al., 2014). In ligh o ou esul s, i seems ha de achmen is mo e mean- ing ul in e ms o lunch ime eco e y han con ol. This is logical, as de achmen ensu es o al absence o job Sianoja e al: Lunch ime Reco e y and Long-Te m Ou comes A . 7, page 9 o 12 demands, whe eas employees wi h high le el o con ol may s ill choose o engage, o example, in discussing wo k issues. Ou esul he e o e ex ends he ea lie ind- ing ha de achmen om wo k is a powe ul eco e y expe ience du ing non-wo k ime (Sonnen ag & F i z, 2015). Howe e , ou one-i em measu e o con ol did no necessa ily cap u e all dimensions o con ol as a eco e y expe ience, o example con ol o e when o ake lunch b eaks (c . Sonnen ag & F i z, 2007). The measu e used may he e o e ha e unde es ima ed he impo ance o con ol du ing b eaks. We ecommend u u e s udies o assess eco e y expe iences a lunch ime wi h mul iple i ems o cap u e hei ull meaning. Bo h aking longe lunch b eaks and habi ually spend- ing b eaks ou side he wo kplace p emises we e co e- la ed wi h highe le els o de achmen . Thus ou esul s sugges ha lunch b eak leng h and spending lunch b eaks ou side he o ice building may ma e o lunch- ime de achmen , which in u n ela es o lunch ime eco e y. We ecommend ha u u e s udies, wi h longi- udinal designs enabling app op ia e media ion analysis, es whe he lunch ime eco e y expe iences media e he e ec s o lunch ime se ings and ac i i ies on eco e y. One ea lie s udy ound ha spending he b eak inside e sus ou side one’s o ice (ou side = in he same build- ing o ou side he building) did no ha e an e ec on eco e y a e b eaks du ing he wo king day (Hun e & Wu, 2016). As ou esul s sugges ha whe e lunch b eaks a e spen could ma e , i is impo an o no e ha ou measu e (ou side = ou side he o ice building) was di e en om he one used by Hun e and Wu (2016). The e o e we sugges ha u u e s udies use mo e com- p ehensi e measu es in di e en ia ing whe e b eaks a e spen o disen angle hese di e ing esul s. Fo example, spending b eaks in he b eak oom o he depa men could ha e di e en eco e y ou comes om spend- ing b eaks ou side he o ice building in a es au an . Fu he mo e, ou ou side condi ion was qui e gene al, and did no ake speci ic eco e y enhancing en i on- men al ac o s (e.g., na u al se ings) in o accoun . Gi en ha na u al se ings a e mo e likely o a o d es o a i e expe iences han a e buil en i onmen s, compa ing hem would be a good op ion o u u e s udies (B own e al., 2014). Fu he mo e, in ou s udy, spending he lunch b eaks wi h o he s was no associa ed wi h eco e y. This is su p ising, as ea lie esea ch sugges s ha b eaks including social ac i i ies a e mo e bene icial o eco e y han b eaks spen alone (Wendsche e al., 2014). Howe e , ea lie esea ch has also sugges ed ha social ac i i ies a e mo e bene icial when based on one’s own choice (T ougakos e al., 2014). Ou s udy ook no accoun o his issue, which may explain ou non- signi ican inding. Addi ionally, we did no dis inguish be ween spending he b eak wi h colleagues and spend- ing he b eak wi h o he people, like iends and amily. This may be impo an , as in heo y spending he lunch b eak wi h iends o amily may ela e o mo e success- ul de achmen om wo k han spending he b eak wi h colleagues. The e o e we ecommend ha u u e s ud- ies ake in o accoun whe he social ac i i ies a e based on employees’ own choice and wi h whom employees spend hei b eaks. When looking a lunch ime eco e y and i s long- e m ela ionship wi h ene gy le els, we ound ha success ul lunch ime eco e y was associa ed wi h less exhaus ion one yea la e , as expec ed. Al hough he e ec s we ound we e small, i is wo h no ing ha his ela ionship was s ill alid a e con olling o baseline le el o exhaus ion and se e al con ols. Thus success ul lunch ime eco e y explained a mino dec ease in exhaus ion in he long e m. Ou indings lend en a i e suppo o ou expec- a ions de i ed om he E-R model: insu icien eco e y du ing lunch b eaks is ela ed o loss o ene gy. When his loss o ene gy accumula es o e ime due o epea ed episodes o insu icien eco e y, i may pa ly explain inc eased le els o exhaus ion. Fu he mo e, ou esul is in line wi h he conclusions o ea lie s udies linking in e nal eco e y wi h less exhaus ion in he sho e m (Hun e & Wu, 2016). Simila ly, he connec ion be ween lunch ime eco e y and igo was suppo ed. Success ul eco e y was ela ed o a mino inc ease in igo one yea la e a e con olling o baseline le el o igo and se e al o he con ols. These indings en a i ely suppo ou expec a ions de i ed om he E-R and COR heo ies ha success ul eco e y p e en s ene gy loss and inc eases in e nal esou ces (e.g., ene gy). When lunch ime eco e y is epea edly success ul, i accumula es and gene a es new esou ces ac oss ime, ela ing o a small inc ease in igo . As he le els o exhaus ion and igo a wo k we e easonably s able o e one yea (i.e., he T1 le el explained abou hal o hei a iance a T2), ou indings es ima ing he long- e m change in ene gy le els due o lunch ime eco e y can be conside ed p omising. Taken oge he , lunch ime eco e y seems o be o impo ance in e ms o ene gy a wo k o e ime. Limi a ions, s eng hs, and sugges ions o u u e s udies This s udy has ce ain limi a ions ha should be consid- e ed. Fi s , choosing he bes ime lag o s udying lon- gi udinal ela ions be ween in e nal eco e y and ene gy is no sel -e iden and he one-yea ime lag used in ou s udy is deba able. Ou esul s explained a ia ion in ene gy le els only o a mino deg ee. The e ec s would likely be s onge i mo e equen measu es o e sho e ime lags (e.g., e e y couple o mon hs) we e applied. Fu u e esea ch may bene i om es ing simila long- e m e ec s wi h mo e equen measu emen s o e di - e en ime spans. Ne e heless, ou longi udinal analysis suppo ed long- e m ela ionships be ween lunch ime eco e y, exhaus ion and igo , suppo ing he iew ha employees’ deg ee o eco e y du ing hei lunch b eaks may ha e signi icance, no only on a daily le el, bu also in he long- e m. Second, al hough p e ious s udies ha e demons a ed one i em measu es o be alid subs i u es o longe scales (D ole & Mo ison, 2001; Elo e al., 2003; Fishe , Ma hews, & Gibbons, 2016; Kinnunen e al., 2011) u u e esea ch may bene i om using mul iple i em measu es o lunch ime eco e y and eco e y expe iences. Thi d,