CHEN-BO ZHONG on in­for­ma­tion struc­ture and cre­ativ­ity

Rotman Management Magazine - - FROM THE EDITOR -

MANY DAY-TO-DAY work­place ac­tiv­i­ties are char­ac­ter­ized by ‘hi­er­ar­chi­cal struc­ture’. For ex­am­ple, the lean-man­u­fac­tur­ing sys­tem cat­e­go­rizes all com­po­nents in­volved into clearly-de­fined cat­e­gories, so that work­ers can eas­ily dis­tin­guish be­tween items and use the cor­rect com­po­nents on man­u­fac­tur­ing lines.

In many cases, in­for­ma­tion is also highly struc­tured in the work­place. Since em­ploy­ees are clus­tered around jobs and roles, both ex­plicit in­for­ma­tion (as com­piled in a job man­ual) and im­plicit in­for­ma­tion (which is im­plied or un­der­stood by the worker) are cat­e­go­rized by job func­tion.

With­out dis­put­ing the rec­og­nized ben­e­fits of hi­er­ar­chi­cal struc­ture, I re­cently con­ducted re­search with Rot­man PHD Can­di­date Yeun Joon Kim to de­ter­mine if such struc­ture might also come with a high cost: re­duced cre­ativ­ity.

For our work, we de­fined cre­ativ­ity as ‘com­bi­na­tions of in­for­ma­tion that are both novel and use­ful’. We used the term in­for­ma­tion broadly, per the def­i­ni­tion of ‘declar­a­tive in­for­ma­tion’, which in­cludes ‘chunks’ of in­for­ma­tion such as ob­jects, sym­bols and facts that pos­sess dis­tin­guish­able at­tributes. For ex­am­ple, a chair con­tains at least three pieces of declar­a­tive in­for­ma­tion: a seat, legs and a back — each of which refers to a spe­cific part with unique at­tributes that are dis­tin­guish­able from the oth­ers. In this sense, in the realm of pro­duc­tion, raw ma­te­ri­als can be con­sid­ered to be declar­a­tive in­for­ma­tion.

We sus­pected that a hi­er­ar­chi­cal struc­ture fea­tur­ing declar­a­tive in­for­ma­tion might be a dou­ble-edged sword: On the one hand, as re­search shows, it clearly in­creases ef­fi­ciency; on the other, we felt that it may re­duce the gen­er­a­tion of creative ideas, be­cause the pres­ence of higher-or­der cat­e­gories could re­duce ‘dis­tal’ or un­com­mon as­so­ci­a­tions. For ex­am­ple, the in­ven­tor of the wheel­chair needed to con­nect two dis­tal pieces of declar­a­tive in­for­ma­tion: ‘wheel,’ which typ­i­cally be­longs to the ve­hi­cle cat­e­gory, and ‘chair,’ which be­longs to the fur­ni­ture cat­e­gory. We posited that such as­so­ci­a­tions are less likely to take place when the in­for­ma­tion pro­vided to work­ers is highly struc­tured.

Hier­ar­chichal vs. Flat In­for­ma­tion Struc­tures

‘In­for­ma­tion struc­ture’ refers to the way in which units of in­for­ma­tion are associated with one another within an in­for­ma­tion set, and it can be ei­ther hi­er­ar­chi­cal or flat.

In a hi­er­ar­chi­cal in­for­ma­tion struc­ture, in­for­ma­tion is or­ga­nized by higher-or­der cat­e­gories, whereby units of in­for­ma­tion within each cat­e­gory have strong con­cep­tual re­la­tion­ships, but those be­tween cat­e­gories have weak con­cep­tual re­la­tion­ships. In a flat in­for­ma­tion struc­ture, in­for­ma­tion is pre­sented with­out higher-or­der cat­e­gories and units of in­for­ma­tion have weak con­cep­tual re­la­tion­ships with each other.

To il­lus­trate, an in­for­ma­tion set that in­cludes ‘cat’, ‘dog’, ‘cow’, ‘mouse’ and ‘tiger’ is hi­er­ar­chi­cally or­ga­nized un­der the higher-or­der cat­e­gory of ‘an­i­mal’. On the other hand, a set of in­for­ma­tion such as ‘pud­ding’, ‘Ukraine’, ‘cheque’, ‘mouse’ and ‘sym­phony’ has a flat in­for­ma­tion struc­ture, be­cause these terms do not share any ob­vi­ous higher-or­der cat­e­gory. We posited that the lat­ter would lead to higher lev­els of cre­ativ­ity, mainly due to the ‘cog­ni­tive flex­i­bil­ity’ it en­gen­ders. Cog­ni­tive flex­i­bil­ity is the ex­tent to which in­di­vid­u­als can eas­ily switch their fo­cus be­tween dif­fer­ent cat­e­gories or per­spec­tives — mak­ing it more likely that they will in­te­grate dis­tal in­for­ma­tion in unique ways. Pre­vi­ous stud­ies have found a pos­i­tive re­la­tion­ship be­tween cog­ni­tive flex­i­bil­ity and cre­ativ­ity.

We felt that in a hi­er­ar­chi­cal in­for­ma­tion struc­ture, the pres­ence of a higher-or­der cat­e­gory in­flu­ences the in­ter­pre­ta­tion of the in­for­ma­tion in that cat­e­gory, re­duc­ing the pos­si­bil­ity for al­ter­na­tive uses of the in­for­ma­tion; and that in a flat in­for­ma­tion struc­ture, the ab­sence of higher-or­der cat­e­gory al­lows in­di­vid­u­als to dis­cover al­ter­na­tive in­ter­pre­ta­tions of the in­for­ma­tion and in­creases cog­ni­tive flex­i­bil­ity.

In ad­di­tion, we felt that a flat in­for­ma­tion struc­ture might ac­tu­ally in­crease cog­ni­tive flex­i­bil­ity, be­cause it in­tro­duces higher prob­a­bil­i­ties of mak­ing dis­tal con­nec­tions be­tween con­cepts. By def­i­ni­tion, the flat in­for­ma­tion struc­ture has a flat as­so­cia­tive hi­er­ar­chy, mean­ing that each unit of in­for­ma­tion has ap­prox­i­mately equal prob­a­bil­i­ties of be­ing next to any other units of in­for­ma­tion in the set. There­fore, com­pared to those in the hi­er­ar­chi­cal in­for­ma­tion struc­ture con­di­tion, in­di­vid­u­als pre­sented with a flat in­for­ma­tion struc­ture may be more likely to dis­cover serendip­i­tous as­so­ci­a­tions be­tween dis­tal in­for­ma­tion.

Given that our con­scious imag­i­na­tion is bounded and our abil­ity to as­so­ci­ate dis­tal cat­e­gories (i.e., cog­ni­tive flex­i­bil­ity) is lim­ited, serendip­ity can re­fresh ha­bit­ual think­ing and open up new pos­si­ble as­so­ci­a­tions. His­tory pro­vides nu­mer­ous in­stances where serendip­i­tous dis­cov­er­ies — such as the Archimedes prin­ci­ple or the X-ray — have en­riched our lives. Based on these pre­vi­ous find­ings, we felt that a flat in­for­ma­tion struc­ture would in­crease the chances for serendip­i­tous, flex­i­ble uses of in­for­ma­tion.

Our Re­search

We tested our pre­dic­tions in three ex­per­i­ments, us­ing a sen­tence con­struc­tion task and a LEGO task. In the sen­tence­con­struc­tion task, par­tic­i­pants were given a set of words and asked to con­struct mean­ing­ful sen­tences out of the them. In the LEGO task, sub­jects were asked to con­struct an alien fig­ure from a set of LEGO bricks. Both tasks in­volved as­sem­bling com­po­nents, but there was no one ‘cor­rect’ way of do­ing things, and hence ef­fi­ciency was not a rel­e­vant cri­te­rion.

One-hun­dred-and-sixty un­der­grad­u­ate stu­dents vol­un­tar­ily par­tic­i­pated in this ex­per­i­ment in ex­change for one course credit. Upon ar­rival, par­tic­i­pants were ran­domly as­signed to ei­ther a hi­er­ar­chi­cal in­for­ma­tion con­di­tion or a flat in­for­ma­tion struc­ture con­di­tion. Each re­ceived two sheets of let­ter-size pa­per and a pen­cil; the first sheet con­tained 100 nouns, and the other sheet was for writ­ing down sen­tences.

In the hi­er­ar­chi­cal con­di­tion, par­tic­i­pants were pro­vided with a sheet of let­ter-size pa­per con­tain­ing the 100 English nouns or­ga­nized by 20 cat­e­gories. Each cat­e­gory con­tained five nouns that were con­cep­tu­ally re­lated. We did not pro­vide spe­cific names for each cat­e­gory. In the flat in­for­ma­tion struc­ture con­di­tion, the same 100 English

We felt that a flat in­for­ma­tion struc­ture might ac­tu­ally in­crease cog­ni­tive flflex­i­bil­ity.

nouns were pre­sented on a sheet of pa­per with­out any cat­e­go­riza­tion. Par­tic­i­pants were in­structed to gen­er­ate as many sen­tences as they could by com­bin­ing the nouns, tak­ing as much time as they needed within a 60-minute time limit.

To eval­u­ate lev­els of cre­ativ­ity in the sen­tences gen­er­ated, three un­der­grad­u­ates from the Lin­guis­tics depart­ment of a North Amer­i­can univer­sity were re­cruited. They eval­u­ated the level of cre­ativ­ity of each sen­tence from 1 (not at all creative) to 7 (ex­tremely creative).

In Study 1, par­tic­i­pants pre­sented with dis­or­ga­nized RE­SULT: in­for­ma­tion — a flat in­for­ma­tion struc­ture — were more creative than those pre­sented with in­for­ma­tion or­ga­nized by cat­e­gories. The ben­e­fi­cial ef­fect of the flat in­for­ma­tion struc­ture on cre­ativ­ity was me­di­ated by cog­ni­tive flex­i­bil­ity.

For Study 2, in ex­change for a course credit, 117 un­der­grad­u­ate stu­dents were re­cruited. We used the same task as in Study 1; the only dif­fer­ence was that we pro­vided 45 English nouns in­stead of 100. Par­tic­i­pants were ran­domly as­signed to the con­di­tions and re­ceived a sheet of pa­per con­tain­ing 45 nouns ei­ther or­ga­nized by nine cat­e­gories or un­or­ga­nized. They were asked to con­struct as many sen­tences as they could by com­bin­ing pro­vided nouns. Un­like Study 1, which used a pa­per-and-pen­cil sur­vey, these par­tic­i­pants en­tered their sen­tences on­line.

To eval­u­ate the sen­tences cre­ated, we re­cruited three PHD can­di­dates to eval­u­ate (1) cre­ativ­ity (‘how creative is this sen­tence?’); and (2) creative use of the pro­vided nouns (‘how cre­atively has the par­tic­i­pant used the pro­vided noun(s) in this sen­tence?’).

Study 2 sup­ported our hy­pothe­ses. Par­tic­i­pants in RE­SULT: the flat-in­for­ma­tion struc­ture con­di­tion gen­er­ated more creative sen­tences than those in the hi­er­ar­chi­cal in­for­ma­tion struc­ture con­di­tion us­ing dif­fer­ent mea­sures of cre­ativ­ity. In ad­di­tion, the ef­fects of in­for­ma­tion struc­ture on cre­ativ­ity were me­di­ated via cog­ni­tive flex­i­bil­ity such that those in the for­mer con­di­tion used nouns from more cat- egories than those in the lat­ter con­di­tion.

In Study 3, we at­tempted to show that our pre­dic­tions are not lim­ited to ab­stract con­structs, but ap­ply to an in­stance that in­volves com­bin­ing units to cre­ate new ob­jects. LEGO bricks are anal­o­gous to units of in­for­ma­tion in many ways and, sim­i­lar to in­for­ma­tion, there are al­most an in­fi­nite num­ber of al­ter­na­tive com­bi­na­tions of LEGO bricks.

Just like new in­for­ma­tion can be cre­ated by com­bin­ing ex­ist­ing in­for­ma­tion, LEGO bricks can be com­bined to make com­plex shapes and struc­tures (e.g., houses, ro­bots, and crea­tures). Also, sim­i­lar to other declar­a­tive in­for­ma­tion, LEGO bricks can be cat­e­go­rized by higher-or­der cat­e­gories such as colour and shape.

In ex­change for one course credit or ten dol­lars, 182 un­der­grad­u­ate stu­dents vol­un­tar­ily par­tic­i­pated in this ex­per­i­ment. We used the Alien Task de­vel­oped by Univer­sity of Alabama Pro­fes­sor Thomas B. Ward. Orig­i­nally, this task asked par­tic­i­pants to imag­ine that they are visit­ing a planet in another galaxy and en­counter an alien who lives on that planet—and to draw that alien. In­stead of draw­ing an alien, we asked par­tic­i­pants to build one out of LEGO bricks.

Upon ar­rival, par­tic­i­pants were ran­domly as­signed to ei­ther the hi­er­ar­chi­cal or flat in­for­ma­tion struc­ture con­di­tion. In the hi­er­ar­chi­cal con­di­tion, a to­tal of 442 LEGO bricks, which con­sisted of nine colours and 11 quad­ran­gu­lar shapes (99 pos­si­ble cat­e­gories), were cat­e­go­rized into two large boxes. Each box had 24 cells (four rows, six col­umns) par­ti­tioned by plas­tic walls. The bricks were cat­e­go­rized by a to­tal of 48 cells. In the flat in­for­ma­tion struc­ture con­di­tion, the same LEGO bricks were con­tained in two large boxes of iden­ti­cal size to those in the hi­er­ar­chi­cal in­for­ma­tion struc­ture con­di­tion. How­ever, in the flat con­di­tion there was no par­ti­tion, so all the 442 bricks were mixed and di­vided into the two large boxes.

Par­tic­i­pants were in­structed not to pour the bricks onto the ta­ble and could only take pieces di­rectly from the boxes when they needed them. Three in­de­pen­dent raters were re­cruited to eval­u­ate the cre­ativ­ity of the LEGO aliens.

Par­tic­i­pants in the flat-in­for­ma­tion con­di­tion made RE­SULT: more creative alien fig­ures than those in the hi­er­ar­chi­cal con­di­tion. Fur­ther, the ef­fects of struc­ture on cre­ativ­ity were me­di­ated by both cog­ni­tive flex­i­bil­ity and per­sis­tence.

Im­pli­ca­tions for Man­agers

Across three stud­ies, we found that in­di­vid­u­als pre­sented with a flat in­for­ma­tion struc­ture were more creative com­pared to those pre­sented with a hi­er­ar­chi­cal in­for­ma­tion struc­ture. We also con­firmed that the in­creased cre­ativ­ity in the flat in­for­ma­tion struc­ture con­di­tions was due to the re­sult­ing in­creased level of cog­ni­tive flex­i­bil­ity.

Or­ga­ni­za­tional stud­ies have had a long his­tory ad­vo­cat­ing for hi­er­ar­chi­cal struc­tures that in­crease the ef­fi­ciency of work. This makes sense, be­cause the pro­cess­ing of large amounts of in­for­ma­tion is lim­ited by hu­man ca­pac­ity. As a re­sult, struc­tures have emerged to re­duce com­plex­ity and en­hance ef­fi­ciency. For this rea­son, many or­ga­ni­za­tional ac­tiv­i­ties are built around hi­er­ar­chi­cal struc­tures.

The prin­ci­ple of di­vi­sion of labour, for ex­am­ple, or­ga­nizes labour forces by worker spe­cial­iza­tions, cre­at­ing a hi­er­ar­chi­cal in­for­ma­tion struc­ture where in­for­ma­tion clus­ters around job roles and skills. Such struc­tures are im­por­tant and are of­ten nec­es­sary to pro­mote ef­fi­ciency.

How­ever, when it comes to cre­ativ­ity, rigid walls be­tween cat­e­gories in hi­er­ar­chi­cal in­for­ma­tion struc­tures may be harm­ful, be­cause creative ideas of­ten rise from com­bin­ing of dis­tal in­for­ma­tion. Put sim­ply, a hi­er­ar­chi­cal in­for­ma­tion struc­ture seems to prime the ideation process within par­tic­u­lar cog­ni­tive cat­e­gories, whereas a flat in­for­ma­tion struc­ture frees up flex­i­ble ex­plo­ration over dis­tal cog­ni­tive cat­e­gories.

In clos­ing

Re­searchers have sug­gested that func­tional diver­sity in teams in­creases cre­ativ­ity, be­cause team mem­bers from var­i­ous func­tional and cul­tural back­grounds bring dif­fer­ent reper­toires of in­for­ma­tion to team ideation pro­cesses. Our find­ings pro­vide clues for bet­ter man­ag­ing func­tion­al­ly­di­verse teams.

To reap the ben­e­fits of a cross-func­tional team, our re­search sug­gests that man­agers might need to cre­ate a flat in­for­ma­tion struc­ture. This can be achieved by for­mal pro­ce­dures such as hav­ing team mem­bers jot down as many new ideas as pos­si­ble and then mix­ing them up (so as not to make any or­ga­niz­ing cat­e­gories salient) be­fore com­bin­ing them for new prod­uct devel­op­ment.

As in­di­cated herein, in ad­di­tion to mak­ing in­for­ma­tion avail­able to em­ploy­ees, how that in­for­ma­tion is pre­sented is a crit­i­cal fac­tor that af­fects creative out­put.

di sor­ga­niz par­tic­i­pants par­tic­i­pants pre­sented with dis­or­ga­nized in­for­ma­tion were more creative.

Newspapers in English

Newspapers from Canada

© PressReader. All rights reserved.