Husky Den Signage Analysis - The UW Garbology...

13
Jack Johnson UW Garbology Project Final Report: Husky Den Waste Bin Signage 6/21/2013 Abstract This document summarizes a pilot study undertaken by the UW Garbology Project (UWGP) in collaboration with the UW Climate Action Plan (UW CAP) to examine which (if any) types of posted wastesorting signs common to UW campus waste disposal bins are most effective in terms of encouraging waste diversion through increased recycling and composting. Project background and objectives are discussed first. This discussion is followed by a detailed description of methods, data, and resultant interpretations and conclusions. It is argued that collected data suggest an underlying difference in average waste diversion rates as a result of posted sign types, with signs produced by UW CAP ranking most effective at diverting waste, signs produced by UW Recycling ranking 2 nd most effective, signs produced by UW Housing and Food Services (UW HFS) ranking 3 rd most effective, and bins absent guides for sorting ranking least effective. Importantly, however, this difference between sign types is statistically modest in comparison with background variance in accumulations within bins, meaning these results should be considered provisional until further study can be performed. Finally, a short series of recommendations for such study are provided, along with a discussion of possible next steps in developing, testing, and implementing more effective signs on campus waste bins. Background Previous work by the UW Garbology Project (UWGP) has documented the composition of UW campus waste in various settings as a means of identifying key areas for improvement in UW’s ongoing efforts to divert more of our waste from ending up in landfills. Through this work it has become clear that campus users often dispose of waste incorrectly, even in cases where infrastructure is readily available for proper waste disposal. For example, we have shown that up to 88% (by mass) of items placed in landfill bound trash bins are compostable or recyclable even when these trash bins are directly adjacent to recycling and compost bins 1 . Thus if our goal is to divert more waste from landfills – thereby improving both UW’s fiscal bottom line as well as its environmental sustainability – we clearly need to do a better job of encouraging the UW community to properly use the waste infrastructure provided for them. Such improvement could be initiated a variety of ways (and we are currently exploring a number of them), but signage placed on bins is a good place to start for a host of reasons. Foremost amongst these reasons is the fact that bin signs are, by dint of location, the visual reference upon which users are most likely to rely at the moment of discard, and as such the degree to which these signs are effective at communicating the nuances of proper disposal is likely to strongly impact how waste is discarded in aggregate. Importantly, these nuances are not trivial, as properly sorting the broad and confusing array of materials discarded on campus can be quite difficult in practice, especially for campus users who have been educated in one of the vast constellation of differing municipal systems in place elsewhere. We therefore need good signs on bins to aid users in proper sorting. Our current infrastructure recognizes this need, and sorting signs on bins are relatively common on campus, especially in indoor areas. Unfortunately, a broad array of types of signage are currently 1 http://uwgarbology.weebly.com/uploads/1/3/0/1/13017489/fall_2012_final_report_v1.1.pdf

Transcript of Husky Den Signage Analysis - The UW Garbology...

Page 1: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

Jack  Johnson  UW  Garbology  Project  Final  Report:    Husky  Den  Waste  Bin  Signage  6/21/2013      Abstract  This  document  summarizes  a  pilot  study  undertaken  by  the  UW  Garbology  Project  (UWGP)  in  collaboration  with  the  UW  Climate  Action  Plan  (UW  CAP)  to  examine  which  (if  any)  types  of  posted  waste-­‐sorting  signs  common  to  UW  campus  waste  disposal  bins  are  most  effective  in  terms  of  encouraging  waste  diversion  through  increased  recycling  and  composting.    Project  background  and  objectives  are  discussed  first.    This  discussion  is  followed  by  a  detailed  description  of  methods,  data,  and  resultant  interpretations  and  conclusions.    It  is  argued  that  collected  data  suggest  an  underlying  difference  in  average  waste  diversion  rates  as  a  result  of  posted  sign  types,  with  signs  produced  by  UW  CAP  ranking  most  effective  at  diverting  waste,  signs  produced  by  UW  Recycling  ranking  2nd  most  effective,  signs  produced  by  UW  Housing  and  Food  Services  (UW  HFS)  ranking  3rd  most  effective,  and  bins  absent  guides  for  sorting  ranking  least  effective.    Importantly,  however,  this  difference  between  sign  types  is  statistically  modest  in  comparison  with  background  variance  in  accumulations  within  bins,  meaning  these  results  should  be  considered  provisional  until  further  study  can  be  performed.    Finally,  a  short  series  of  recommendations  for  such  study  are  provided,  along  with  a  discussion  of  possible  next  steps  in  developing,  testing,  and  implementing  more  effective  signs  on  campus  waste  bins.      Background  Previous  work  by  the  UW  Garbology  Project  (UWGP)  has  documented  the  composition  of  UW  campus  waste  in  various  settings  as  a  means  of  identifying  key  areas  for  improvement  in  UW’s  ongoing  efforts  to  divert  more  of  our  waste  from  ending  up  in  landfills.    Through  this  work  it  has  become  clear  that  campus  users  often  dispose  of  waste  incorrectly,  even  in  cases  where  infrastructure  is  readily  available  for  proper  waste  disposal.    For  example,  we  have  shown  that  up  to  88%  (by  mass)  of  items  placed  in  landfill-­‐bound  trash  bins  are  compostable  or  recyclable  even  when  these  trash  bins  are  directly  adjacent  to  recycling  and  compost  bins1.    Thus  if  our  goal  is  to  divert  more  waste  from  landfills  –  thereby  improving  both  UW’s  fiscal  bottom  line  as  well  as  its  environmental  sustainability  –  we  clearly  need  to  do  a  better  job  of  encouraging  the  UW  community  to  properly  use  the  waste  infrastructure  provided  for  them.    Such  improvement  could  be  initiated  a  variety  of  ways  (and  we  are  currently  exploring  a  number  of  them),  but  signage  placed  on  bins  is  a  good  place  to  start  for  a  host  of  reasons.    Foremost  amongst  these  reasons  is  the  fact  that  bin  signs  are,  by  dint  of  location,  the  visual  reference  upon  which  users  are  most  likely  to  rely  at  the  moment  of  discard,  and  as  such  the  degree  to  which  these  signs  are  effective  at  communicating  the  nuances  of  proper  disposal  is  likely  to  strongly  impact  how  waste  is  discarded  in  aggregate.    Importantly,  these  nuances  are  not  trivial,  as  properly  sorting  the  broad  and  confusing  array  of  materials  discarded  on  campus  can  be  quite  difficult  in  practice,  especially  for  campus  users  who  have  been  educated  in  one  of  the  vast  constellation  of  differing  municipal  systems  in  place  elsewhere.    We  therefore  need  good  signs  on  bins  to  aid  users  in  proper  sorting.        Our  current  infrastructure  recognizes  this  need,  and  sorting  signs  on  bins  are  relatively  common  on  campus,  especially  in  indoor  areas.    Unfortunately,  a  broad  array  of  types  of  signage  are  currently  

                                                                                                                         1  http://uwgarbology.weebly.com/uploads/1/3/0/1/13017489/fall_2012_final_report_v1.1.pdf  

Page 2: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

employed,  and  it  is  not  uncommon  to  find  several  distinct  types  of  signs  in  close  proximity.    Figures  1-­‐4,  for  example,  show  three  different  types  of  signs  within  the  Husky  Den  alone,  and  the  result  of  this  inconsistency  is  greater  complexity  for  the  user  to  interpret,  and  therefore  greater  effort  on  the  part  of  the  user  to  engage  in  good  disposal  behavior.    In  other  words,  the  existence  of  multiple  sign  types  is  likely  to  be  sub-­‐optimal  in  terms  of  user-­‐friendliness,  since  it  entails  a  slightly  greater  potential  for  confusion  or  outright  disregard  than  a  system  which  helps  build  good  habits  through  consistent,  easily-­‐recognizable  standardized  visual  references.    To  make  matters  worse,  some  of  the  signs  currently  in  use  on  campus  are  incomplete  (omitting  certain  items  very  common  to  campus;  see  Fig.  5)  or  downright  outdated  and  misleading  (See  Fig.  6),  which  can  only  further  frustrate  or  confuse  users  looking  to  quickly  and  easily  discard  their  waste.    Still  needed,  then,  is  a  set  of  signs  which  is  1)  accurate  and  up-­‐to-­‐date,  2)  tailored  to  the  waste  items  campus  users  are  most  likely  to  discard,  and  3)  implemented  across  campus  to  the  exclusion  (where  possible)  of  other,  competing  sign  types.    Fortunately,  many  campus  organizations,  including  UW  Recycling,  the  UW  Climate  Action  Plan  (UW  CAP),  and  UW  Housing  and  Food  Services  (UW  HFS)  have  designed  distinct  sets  of  signs  which  easily  meet  the  first  two  criteria  (again,  see  Figs.  1-­‐4),  and  there  exists  a  plentiful  crop  of  candidates  from  which  to  select  one  to  promote  to  campus  ubiquity.    In  selecting  one  sign  type  to  predominate,  however,  we  should  seek  data  which  allow  us  to  evaluate  which  (if  any)  of  our  current  options  works  best.    In  other  words,  we  should  find  a  way  to  measure  and  document  which  type  of  signs  actually  improves  waste  diversion  by  1)  increasing  rates  of  composting  and  recycling  relative  to  landfilling  and  2)  reducing  contamination  in  waste  bins  of  all  types.      Objectives  Simply  put,  the  primary  aim  of  the  work  detailed  here  was  to  begin  to  meet  this  need  through  a  study  of  the  effects  of  different  signage  upon  accumulation  rates  in  waste  bins  within  the  Husky  Den  (the  dining  area  of  the  Husky  Union  Building,  or  HUB).    Secondary  goals  included  1)  building  working  relationships  with  UW  CAP  (who  collaborated  on  the  project)  and  UW  HFS  staff  working  in  the  Husky  Den  to  form  the  basis  for  potential  future  collaborations,  and  2)  field-­‐testing  new  methods  for  measuring  waste  accumulation  rates  with  laser  distance  measurers  (see  below)  before  applying  these  methods  to  broader  UWGP  initiatives  across  campus.    The  Husky  Den  itself  was  chosen  for  several  reasons.      First,  it  is  a  high-­‐traffic  dining  area  which  produces  a  high  waste  load  which  operates  well  above  the  average  campus-­‐wide  waste  diversion  rate2,3,  and  as  such  it  was  expected  to  provide  both  a  dynamic  environment  for  sampling  as  well  as  insight  into  a  relatively  ideal  portion  of  UW’s  waste  system.    Second,  the  area  contains  11  total  sets  of  waste  bins  (see  Fig.  7  for  a  map),  allowing  examination  and  comparison  of  a  healthy  sample  of  distinct  bin  locations  within  a  single  dining  area.    Third,  this  area  contains  4  different  signage  types  among  its  11  waste  bin  locations  –  3  locations  display  UW  CAP  signs,  3  display  UW  Recycling  signs,  4  display  UW  HFS  signs,  and  1  displays  no  sorting  signs  (bins  here  are  labeled  as  compost,  recycling,  or  trash,  but  no  guide  is  posted  as  to  which  items  belong  in  which  bins)  –  allowing  a  good  basis  for  comparison  between  each  of  3  potentially  viable  signage  types  and  a  control  location.    Lastly,  upon  inception  of  this  project  UW  CAP  

                                                                                                                         2  According  to  data  provided  by  Emily  Newcomer,  Manager  of  UW  Recycling,  over  77  tons  of  waste  were  collected  from  the  HUB  between  September  2012  and  the  end  of  January  2013.    Composition  of  that  waste  was  about  23.7%  recyclables,  43.1%  compost,  and  33.2%  trash,  for  a  total  diversion  rate  of  about  66.8%.  3  UW  Recycling’s  FY  2012  report  estimates  the  current  campus-­‐wide  waste  diversion  rate  to  be  about  57%  (see  http://www.washington.edu/facilities/building/recyclingandsolidwaste/files/report2012/index.html).  

Page 3: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

was  already  working  within  the  Husky  Den  and  could  provide  access  and  logistical  support  for  our  efforts,  lowering  our  investment  of  time  and  resources  in  this  endeavor.    This  is  not  to  say  that  this  space  and  our  access  to  it  were  free  of  logistical  challenges.    In  fact,  the  truth  is  quite  the  opposite;  because  this  is  a  first-­‐time  collaboration,  Husky  Den  staff  members  were  reticent  to  allow  us  to  carry  out  the  full  scope  of  the  study  we  initially  envisioned.    Specifically,  since  our  interest  was  in  ascertaining  the  impacts  of  signage  on  bin  accumulation  rates  and  contamination,  we  were  interested  in  performing  1)  spot  checks  of  bin  accumulations,  2)  measurements  of  bulk  weights  of  waste  produced  at  bin  locations,  and  3)  sorting  and  analysis  of  subsamples  of  waste,  but  Husky  Den  staff  preferred  we  limit  our  initial  investigation  to  only  the  first  of  these  dimensions  until  we  were  able  to  demonstrate  that  we  were  trustworthy,  productive,  and  unobtrusive  collaborators.    This  is  a  perfectly  fair  stance  on  their  part,  but  it  meant  limitations  in  data  collection  for  this  study.    As  such,  our  objectives  have  been  tempered  by  the  type  of  data  we  could  reasonably  generate.    Specifically,  we  were  unable  to  calculate  waste  diversion  in  terms  of  mass  (as  is  the  norm)  and  we  had  to  jettison  an  investigation  of  contamination,  as  we  were  only  able  to  collect  data  relevant  to  how  quickly  waste  volume  of  different  types  accumulates  as  a  function  of  posted  signage.    Still,  this  limited  objective  is  informative  and  worthwhile  as  a  basis  for  both  initial  interpretation  and  future  investigation,  even  if  it  results  in  conclusions  which  are  neither  comprehensive  nor  definitive.      Methods  Given  the  above  objectives  and  considerations,  the  methods  employed  in  this  project  were  generally  aimed  at  enabling  comparison  of  diversion  rates  associated  with  distinct  sign  types  without  the  need  for  either  1)  weighing  any  waste  or  2)  withdrawing  any  samples  of  waste  from  bins.    As  such,  the  methods  used  here  hinged  entirely  on  the  use  of  volumetric  estimates  of  waste  accumulation  rates  within  bins  as  a  basis  of  comparison.        Because  all  waste  bins  in  the  Husky  Den  are  identical  in  their  internal  spatial  dimensions,  we  were  able  to  use  simple  bin  “fullness”  as  a  proxy  for  the  amount  of  accumulation  within  each  individual  bin.    Fullness  was  measured  with  the  use  of  a  Bosch  DLE  40  laser  distance  measurer  (LDM)  by  holding  the  LDM  in  such  a  way  that  its  optics  were  even  with  the  horizontal  plane  at  the  bin  aperture,  its  optics  were  centered  within  the  bin  aperture  in  plan  view,  and  its  beam  emitted  perpendicular  to  the  horizontal  plane.    In  other  words,  we  held  the  device  so  that  it  measured  the  distance  from  the  top  of  each  bin  straight  down  into  the  center  of  the  waste  accumulated  in  the  bin  itself.    This  allowed  us  to  quickly  and  cleanly  measure  the  surface  heights  of  accumulated  waste  in  a  consistent  and  standardized  way,  in  turn  aiding  the  collection  of  abundant  data  which  would  be  directly  comparable  between  bin  measurements.    Measurements  of  bin  fullness  were  repeated  for  each  bin  at  all  11  Husky  Den  bin  locations  between  2  and  3  times  daily  every  Monday  through  Friday  for  a  period  of  5  weeks  during  spring  quarter  2013.    Each  measurement  of  bin  fullness  was  accompanied  by  a  notation  of  time,  bin  type,  sign  type,  and  bin  location,  allowing  data  to  be  segmented  by  these  dimensions  and  normalized  by  elapsed  interval.    Measurements  were  conducted  at  roughly  9:00  AM,  11:30  AM,  and  1:30  PM;  these  times  allowed  us  to  work  around  volunteer  availability  while  capturing  accumulation  rates  during  peak  use  hours  (lunchtime)  in  the  Husky  Den  in  the  face  of  rolling  bin-­‐emptying  episodes  (Husky  Den  custodians  empty  bins  as  needed  rather  than  on  a  set  schedule).    The  resultant  dataset  provided  a  set  of  time-­‐lapse  snapshots  of  accumulations  within  all  Husky  Den  bins  over  a  roughly  4.5  hour  period  for  a  total  of  25  days.    This  dataset  is  not  reported  in  full  here  for  purposes  of  brevity.    

Page 4: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

When  measurement  finished,  this  dataset  was  used  to  derive  relative  volumetric  diversion  rates  for  each  of  the  11  bin  locations.    This  was  done  by  first  removing  data  in  cases  where  negative  bin  accumulations  were  observed  for  an  interval,  as  these  instances  represent  bin-­‐emptying  episodes  and  therefore  cannot  provide  reliable  indices  of  accumulation  rates.    After  these  data  were  removed  from  consideration,  accumulation  rates  for  each  bin  at  each  location  were  derived  and  expressed  as  meters  per  hour  to  normalize  all  bin  measurements  within  each  interval.    This  was  necessary  to  correct  for  variance  in  apparent  bin  accumulations  as  a  product  of  the  ~10-­‐15  minutes  needed  to  complete  each  measurement  episode.    Accumulation  rates  were  then  ranked  for  each  bin  type  and  each  interval.    This  step  was  necessary  for  two  primary  reasons.    First,  it  allowed  data  from  different  temporal  intervals  to  be  collapsed  into  a  common  basis  of  comparison;  if  this  hadn’t  been  done,  results  from  the  early  (9:00-­‐11:30)  interval  would  not  have  been  directly  comparable  with  those  of  the  later  (11:30-­‐1:30)  interval  due  to  the  fact  that  the  later  interval  witnessed  much  more  rapid  accumulations  of  waste,  thereby  rendering  it  a  different  waste  system  (in  a  statistical  sense)  at  the  ratio  scale  of  measurement.    Second,  it  allowed  us  to  minimize  the  effects  of  any  inaccuracies  incurred  by  using  single-­‐point  fullness  measurements  to  represent  waste  surfaces  of  uneven  topography.    In  sum,  then,  using  rank-­‐order  data  rather  than  ratio  scale  data  allowed  us  to  boost  sample  size  by  making  full  use  of  all  measurements  while  helping  to  ensure  data  and  conclusions  were  relatively  robust  to  individual  measurement  errors.    Once  these  rankings  were  derived,  they  were  then  pooled  by  location  and  averaged  for  each  bin  type  to  yield  average  ranks  of  each  location’s  accumulation  rates  within  each  bin  type.    In  other  words,  we  were  at  this  stage  able  to  determine  which  location  typically  exhibited  the  highest  (and  2nd  highest,  3rd  highest,  etc.)  accumulation  of  compost,  recyclables,  and/or  landfilled  trash.    This  was  repeated  for  data  pooled  by  signage  type  rather  than  by  location  alone.    These  average  rankings  are  provided  in  Tables  1  (by  signage  type)  and  2  (by  bin  location).    Taken  alone,  each  of  these  rankings  means  little  in  terms  of  waste  diversion,  since  they  could  in  part  be  driven  by  raw  accumulation  rates  as  a  result  of  location;  one  bin  set  might  rank  highly  in  all  areas  simply  because  it’s  next  to  a  major  vendor,  seating  area,  or  exit,  for  example.    From  these  rankings,  however,  average  differences  in  rank  between  trash  and  compost  (T-­‐C  scores)  and  between  trash  and  recyclables  (T-­‐R  scores)  were  derived  for  each  location  and  signage  type;  these  scores  are  a  more  useful  measure  of  relative  diversion  rates,  as  they  highlight  which  bins  within  a  given  location/signage  type  accumulate  waste  more  rapidly  than  other  bins  with  the  same  location/sign  type.    In  other  words,  these  scores  indicate  the  degree  to  which  accumulation  associated  with  a  given  signage  type  or  location  was  on  average  dominated  by  trash  (negative  scores)  or  by  recycling/compost  (positive  scores)  compared  to  other  signage  types/locations,  and  as  such  these  scores  may  stand  as  a  proxy  for  relative  volumetric  waste  diversion.      Discussion  of  Results  Upon  consideration  of  the  data  in  Table  1,  it  is  clear  that  all  signage  types  are  on  average  fairly  similar  in  their  performance,  as  all  signage  types  exhibit  moderate  average  rankings  of  between  5.3  and  6.1  for  all  bin  types.    If  signage  type  were  strongly  and  consistently  affecting  waste  deposition  outcomes,  average  rankings  would  exhibit  a  greater  range  of  values  and  standard  deviations  in  average  rankings  would  be  lower  on  the  whole,  reflecting  greater  convergence  on  a  pattern  of  clear  difference  in  performance.    As  it  stands,  differences  in  performance  between  signage  types  are  not  statistically  significant,  indicating  that  any  differences  among  the  signage  types  examined  are  not  a  primary  determining  factor  in  how  waste  is  categorized  upon  disposal  in  the  Husky  Den.    Instead,  random  or  external  factors  –  such  as  individuals’  inherent  preconceptions  and/or  biases  –  currently  seem  to  be  the  predominant  forces  shaping  how  waste  is  discarded  at  a  given  location.  

Page 5: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

 Still,  some  moderate  differences  in  signage  performance  are  likely  affecting  bin  accumulations,  as  some  sign  types  are  on  average  ranked  slightly  higher  than  others  in  terms  of  compost  and  recyclables  accumulated  relative  to  trash  accumulated.    UW  CAP  signs,  for  example,  were  associated  with  the  lowest  average  accumulations  of  trash  and  the  highest  average  accumulations  of  recyclables.    This  same  set  of  signs  was  also  on  average  more  highly  ranked  in  its  compost  accumulations  than  in  its  accumulations  of  trash,  and  taken  together  these  facts  indicate  that  the  overall  waste  stream  moving  through  locations  displaying  UW  CAP  signs  is  slightly  more  weighted  towards  recyclables  and  compost  (and  therefore  more  weighted  away  from  trash)  than  locations  displaying  other  sign  types.    Locations  displaying  UW  Recycling  signs  did  almost  as  well  as  UW  CAP  signs  in  terms  of  producing  overall  waste  streams  which  were  more  weighted  towards  recycling  and  compost.    UW  HFS  signs  performed  relatively  poorly  overall,  as  they  were  associated  with  a  waste  stream  that  was  on  average  ranked  more  highly  in  trash  accumulations  than  in  accumulations  of  either  compost  or  recycling.    The  lone  location  without  signs  to  aid  in  sorting  produced  the  waste  stream  that  was  most  strongly  associated  with  more  trash  and  less  compost  or  recycling  (exhibiting  T-­‐R|C  scores  (see  Tables  1  and  2)  which  average  a  full  rank  lower  than  UW  CAP  signs),  although  just  barely,  as  UW  HFS  signs  performed  only  a  hair’s  breadth  better.    Again,  these  results  are  not  statistically  significant,  but  they  are  suggestive  that  some  of  these  signs  “worked”  slightly  better  than  others  over  the  25  days  studied,  and  as  such  they  represent  potentially  fertile  ground  for  future  investigation,  perhaps  under  more  controlled  experimental  conditions  in  which  true  ratio-­‐scale  comparisons  are  possible,  sample  sizes  are  larger,  masses  of  accumulations  can  be  measured,  and/or  statistical  variance  associated  with  estimates  can  be  otherwise  minimized.    By  contrast  to  patterns  in  signage  type,  higher  average  rankings  of  compost  and  recyclables  relative  to  trash  do  not  exhibit  an  apparent  spatial  pattern  (see  Fig  8).    In  fact,  in  some  cases  (for  example,  bin  locations  1  and  4  and  bin  locations  8  and  9)  bin  locations  exhibiting  markedly  different  T-­‐R|C  scores  sit  immediately  adjacent  to  one  another  within  the  same  Husky  Den  enclosure;  geography  therefore  clearly  plays  no  role  in  differences  between  these  bins.    Consequently,  the  present  evidence  suggests  that  –  as  far  as  diversion  rates  are  concerned  –  proximity  to  a  certain  vendor,  exit,  or  seating  area  within  the  Husky  Den  is  a  non-­‐factor  compared  to  signage  type4.    Since  geography  and  signage  type  are  the  only  dimensions  in  which  these  bins  differ  in  terms  of  user  experience,  the  (provisional,  at  present)  elimination  of  geography  as  a  causal  factor  in  waste  diversion  strengthens  the  assertion  that  observed  differences  between  diversion  rates  (to  the  extent  they  are  evident)  are  probably  due  to  differences  in  signage  type  alone.    In  sum,  then,  the  results  of  this  study  are  encouraging.    While  there  is  room  for  improvement  upon  our  methods  and  findings,  the  results  are  generally  consistent  with  the  assertion  that  good  signage  plays  some  role  in  aiding  waste  diversion,  even  in  a  sampling  environment  already  operating  well  above  average  campus-­‐wide  waste  diversion  rates,  and  that  this  role  is  greater  than  that  played  by  bin  location.    Given  the  exploratory  nature  of  the  project,  we  therefore  count  this  work  as  a  success,  and  hope  it  will  form  the  foundation  of  future  efforts  to  investigate  the  role  of  signage  further.        

                                                                                                                         4  This  supposition  could  be  further  tested,  of  course,  by  rotating  signage  to  different  locations  to  determine  whether  accumulations  are  altered  according  to  signage  or  remain  tied  to  location  itself;  such  an  experiment  would  be  worthwhile  in  any  case,  as  results  could  amplify  observed  patterns,  obscure  these  patterns,  or  illuminate  new  patterns  altogether.  

Page 6: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

Recommendations  While  the  preceding  discussion  leaves  room  for  some  interpretation  and  debate  in  terms  of  the  strength  of  observed  patterns,  for  the  present  these  patterns  are  at  least  consistent  enough  that  we  are  in  a  position  to  make  recommendations  moving  forward.    These  recommendations  are  detailed  below.    Recommendation  1:    Repeat  the  experiment  under  more  controlled  conditions.  It  may  ultimately  be  the  case  that  signage  type  matters  little  compared  to  other  factors  affecting  waste  diversion,  and  as  such  it  may  be  the  case  that  even  the  most  controlled  study  will  fail  to  produce  statistically  significant  differences  between  the  performance  of  different  signage  types.    On  the  other  hand,  we  have  not  yet  achieved  the  experimental  control  needed  to  support  this  hypothetical.      It  would  therefore  be  premature  to  presume  we  can’t  do  better  next  time,  particularly  because  over  the  course  of  this  pilot  study  our  challenges  included:    1)  an  inability  to  control  custodians’  waste  collection  times,  leading  to  discarded  observations  and  extra  variance  in  data,  and  therefore  extra  error  terms  in  results,  2)  an  inability  to  sort  waste,  constraining  our  ability  to  either  a)  translate  volumetric  estimates  into  either  mass  estimates,  or  b)  evaluate  contamination,  and  3)  an  inability  to  implement  a  mid-­‐study  movement  of  sign  types    to  other  bin  locations  –  say,  swapping  UW  CAP  signs  with  UW  HFS  signs  –  as  a  means  of  testing  whether  patterns  in  accumulations  follow  signage  type  or  remain  tied  to  a  set  location.    Given  these  difficulties,  our  ability  to  discern  even  marginal  differences  between  signage  types  is  a  promising  achievement,  and  likely  merits  a  future  study  which  addresses  these  unmet  needs  by:    1)  working  with  custodians  to  coordinate  a  set  daily  interval  for  measurement,  to  be  preceded  by  the  universal  emptying  of  all  bins  so  that  a  common  baseline  for  accumulations  is  achieved,  2)  combining  spot-­‐checks  with  sub-­‐sampling  for  sorting  and  weighing,  3)  moving  signs  to  other  bin  locations  mid-­‐study,  and  4)  conducting  measurements  over  a  much  larger  sample  of  days,  allowing  better  characterization  of  variance  and  therefore  (potentially)  lowering  error  terms,  as  well  as  facilitating  sign-­‐swapping  mid-­‐study.    This  strategy  will  obviously  require  more  work  (sorting,  longer  study  duration),  greater  collaboration  with  Husky  Den  staff  (implementing  coordinated  emptying  and  sampling),  and  more  infrastructure  (facilities  and  supplies  for  collection  and  sorting),  but  it  would  carry  the  benefit  of  a  more  limited  daily  engagement  on  the  part  of  volunteers  (spot-­‐checks  would  only  need  to  be  performed  once  daily,  at  the  end  of  the  designated  interval)  as  well  as  more  definitive  results.    The  UW  Garbology  Project  has  the  resources  to  accommodate  these  needs  in  terms  of  labor,  facilities,  and  supplies;  all  that  remains  is  to  procure  slightly  greater  access  and  collaboration  with  Husky  Den  management  and  custodial  staff.    Given  the  potential  benefits  of  this  work  in  terms  of  increased  waste  diversion,  we  think  pushing  to  procure  this  access  and  collaboration  is  a  worthwhile  endeavor.    Recommendation  2:    Expand  the  scope  of  inquiry  to  incorporate  qualitative  data.  Everything  we  have  done  so  far  explores  whether  signage  type  makes  a  difference  in  terms  of  waste  accumulations.    Missing,  however,  is  any  investigation  as  to  why  one  particular  type  of  sign  might  perform  better  than  another  in  practice.    Does  one  sign  type  grab  attention  better  than  others,  making  users  more  likely  to  use  it  as  a  reference?    Does  one  sign  type  communicate  sorting  criteria  more  effectively  than  others,  aiding  users  (once  engaged)  to  more  easily  make  better  sorting  decisions?    Addressing  these  dimensions  will  be  critical  to  efficiently  refining  our  signage  moving  forward,  since  it  will  allow  the  deliberate  creation  and  implementation  of  signage  with  attributes  known  to  be  effective.        In  broad  terms,  such  systematic  investigation  must  begin  by  hypothesizing  which  particular  factors  enhance  one  sign’s  communicative  power  over  another  –  for  example,  perhaps  UW  HFS  signs  perform  worse  because  the  text  on  them  is  small  and  hard  to  read  –  but  ultimately  these  hypotheses  will  need  to  be  tested  with  evidence  if  concrete  progress  is  to  be  made.    Because  this  testing  process  will  ultimately  

Page 7: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

involve  waste  analysis  (the  only  relevant  endpoint  when  waste  diversion  is  the  goal)  it  will  be  relatively  resource-­‐intensive  compared  to  the  process  of  generating  hypotheses,  thereby  placing  a  premium  on  the  generation  of  good  hypotheses  at  the  outset.    In  other  words,  in  investigating  why  some  signs  work  better  than  others  it  will  pay  to  prioritize  testing  of  only  those  hypothesized  factors  which  are  most  likely  to  matter  to  users.    A  smart  way  to  engage  the  process  of  hypothesis  generation  would  therefore  be  to  begin  by  basing  hypotheses  on  qualitative  observations  derived  from  actual  users  of  UW’s  waste  infrastructure.    Controlled  surveys  or  interviews  with  users  would,  for  example,  serve  to  highlight  how  users  interact  with  signs  of  different  types,  and  might  provide  fertile  ground  for  identifying  which  (if  any)  attributes  of  signs  might  lead  to  greater  effectiveness.    Targeted  testing  of  these  attributes  could  then  proceed  based  on  these  insights,  accelerating  the  process  of  refining  the  effectiveness  of  our  signage.    None  of  this  is  novel,  of  course  –  market  research  firms  have  been  conducting  focus  groups  for  decades  –  but  its  systematic  application  to  effective  waste  diversion  has  yet  to  reach  its  potential.    Recommendation  3:    In  the  meantime,  use  UW  CAP  signs  where/when  it  makes  sense  to  do  so.    Use  UW  Recycling  signs  in  other  locations.  Statistical  caveats  notwithstanding,  we  cannot  ignore  the  facts  that  1)  the  results  of  this  study  derive  from  an  extensive  dataset  (50  measurement  intervals;  25  days,  2  intervals  per  day),  and  are  therefore  substantial,  and  2)  these  results  suggest  that  UW  CAP  signage  performs  better  on  average  than  any  of  the  tested  alternatives.    At  present,  we  are  not  in  a  position  to  turn  a  blind  eye  to  strategies  which  enjoy  some  empirical  support  for  improved  waste  diversion  –  improvement  to  our  campus-­‐wide  waste  diversion  rates  has  slowed  considerably  over  the  last  few  years5  and  these  rates  remain  far  from  UW  Recycling’s  year  2020  goals6  –  if  our  only  alternative  is  to  simply  wait  idly  for  something  better  to  come  along.    Instead,  we  should  press  any  potential  improvement  at  our  disposal,  especially  those  that  incur  implementation  costs  which  are  essentially  negligible.    As  such,  until  fully  conclusive  data  are  available,  the  prudent  course  of  action  is  to  implement  UW  CAP  signs  where  practical.    Because  their  material  composition  means  that  these  signs  cannot  be  mass-­‐produced  and  will  really  only  function  in  indoor  locations,  “practical”  at  present  means  that  these  signs  should  only  be  implemented  in  high-­‐traffic  areas  (like  campus  cafes)  if/when  other  signage  types  are  absent  or  already  in  the  process  of  being  replaced.    If  this  is  done,  small  average  improvements  in  waste  diversion  could  result  in  multiple  tons’  worth  of  additional  waste  diversion  each  year  in  these  limited  contexts  alone.    Additionally,  UW  Recycling  signs,  which  also  perform  reasonably  well  in  terms  of  encouraging  waste  diversion,  are  weatherproof  and  can  be  cheaply  mass-­‐produced.    These  signs  therefore  can  and  should  be  implemented  far  and  wide  in  the  short  term.    In  particular,  these  signs  should  replace  the  range  of  outdated  or  misleading  signs  scattered  across  campus.    Given  the  results  of  this  pilot  study,  it  would  likely  benefit  campus  waste  diversion  efforts  if  these  signs  were  to  broadly  replace  UW  HFS  signs  as  well.      Acknowledgements  Thanks  are  due  to  UW  CAP  staff  members  Elise  Glassman  and  Steve  Smith  for  collaborating  on  all  stages  of  this  project,  Husky  Den  staff  for  allowing  implementation  of  UW  CAP  signs  for  purposes  of  this  study,  

                                                                                                                         5  http://www.washington.edu/facilities/building/recyclingandsolidwaste/files/2011Report.pdf  6  UW  Recycling  aims  for  70%  waste  diversion  by  2020.  

Page 8: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

and  to  the  long  list  of  volunteers  who  performed  daily  measurements,  including  Elise  and  Steve  as  well  as  Johnny  Cheng,  Megan  Rue,  Amy  Mandin,  Angela  Battle,  Claudia  Frere,  Melanie  Wade,  and  Adam  Fahlstrom.      Figures  and  Tables  

   

 Figure  1:    An  example  of  UW  Recycling  bin  signage  on  Husky  Den  waste  bins.  

 

Page 9: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

 Figure  2:    An  example  of  UW  Housing  and  Food  Services  (HFS)  bin  signage  on  Husky  Den  waste  bins.  

 

 Figure  3:    An  example  of  UW  Climate  Action  Plan  (CAP)  bin  signage  on  Husky  Den  waste  bins.  

Page 10: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

 

 Figure  4:    Close-­‐up  examples  of  each  type  of  compost  sign  on  display  in  the  Husky  Den.    From  left  to  right:    UW  

Recycling  sign,  UW  HFS  sign,  UW  CAP  sign.    

 Figure  5:    A  sign  on  a  campus  compost  bin.    This  sign  makes  no  mention  of  compostable  plastics  whatsoever,  

leaving  user  to  categorize  their  plastics  on  their  own.    

Page 11: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

 Figure  6:      Another  sign  on  a  different  campus  compost  bin.    The  sign  prohibits  plastic  take-­‐out  containers  and  utensils  even  though  all  plastic  utensils  and  take-­‐out  containers  sold  on  campus  are  now  compostable  at  Cedar  

Grove.    

 

Page 12: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

 Figure  7:    A  map  of  Husky  Den  waste  bin  locations.    Bins  3,  6,  and  9  displayed  UW  CAP  (CAP)  signs.    Bins  1,  2,  and  5  displayed  UW  Recycling  (UWR)  signs.    Bins  4,  8,  10,  and  11  displayed  UW  HFS  (HFS)  signs.    Bin  7  had  no  guides  for  

materials  sorting  (N/A)  on  display.    This  map  is  modified  from  an  image  provided  by  Steve  Smith  (UW  CAP).              

Sign  Type  Average  Rank   Standard  Deviation  in  Rank   Diff.  in  Average  Rank  

Trash   Recycling   Compost   Trash   Recycling   Compost   T  -­‐  R   T  -­‐  C    (T-­‐R|C)  

UW  CAP   6.1   5.3   5.8   2.9   2.8   2.8   0.8   0.3   0.6  UW  Recycling   6.0   5.6   5.7   3.1   2.8   2.5   0.4   0.3   0.4  UW  HFS   5.3   5.7   5.3   2.6   2.6   3.0   -­‐0.5   0.0   -­‐0.2  

None   5.4   5.6   5.8   2.7   2.8   2.5   -­‐0.2   -­‐0.4   -­‐0.3  Table  1:    Results  of  analysis,  displayed  by  signage  type.    Shown  are  average  ranks  of  accumulation  rates  for  each  

bin  type  (left)  and  standard  deviations  in  these  averages  (center).    Higher  numbers  indicate  lower-­‐ranked  bins,  and  therefore  lower  average  accumulations.    At  right  differences  between  average  ranks  of  accumulations  of  trash  and  

those  of  recyclables  (T-­‐R)  and  compost  (T-­‐C)  are  shown;  more  positive  numbers  indicate  higher  rates  of  composting  and  recycling  compared  to  discard  in  landfill-­‐bound  trash  bins,  while  more  negative  numbers  

represent  higher  rates  of  discard  in  landfill  trash  bins  relative  to  composting  and  recycling.    T-­‐R|C  represents  the  average  of  T-­‐R  and  T-­‐C  scores,  and  provides  a  quick  summary  of  a  sign  type’s  relative  overall  waste  diversion  rate.  

   

Page 13: Husky Den Signage Analysis - The UW Garbology Projectuwgarbology.weebly.com/uploads/1/3/0/1/13017489/husky_den_sign… · 1(2(3(% %

Location  Average  Rank   Standard  Deviation  in  Rank   Difference  in  Average  Rank  

Trash   Recycling   Compost   Trash   Recycling   Compost   T  -­‐  R   T  -­‐  C   T-­‐R|C  

1   5.3   5.3   5.6   3.4   3.0   2.7   -­‐0.1   -­‐0.3   -­‐0.2  2   8.0   6.5   6.2   1.9   2.6   2.0   1.5   1.8   1.7  3   7.0   6.2   6.2   2.3   2.5   2.8   0.8   0.8   0.8  4   5.9   5.4   5.3   2.6   2.6   2.9   0.5   0.7   0.6  5   4.7   4.8   5.4   2.9   2.5   2.9   -­‐0.1   -­‐0.7   -­‐0.4  6   5.4   4.8   6.0   3.2   3.1   3.0   0.7   -­‐0.5   0.1  7   5.4   5.6   5.8   2.7   2.8   2.5   -­‐0.2   -­‐0.4   -­‐0.3  8   4.5   4.4   4.7   2.5   2.3   3.6   0.1   -­‐0.2   -­‐0.1  9   5.9   4.9   5.4   2.9   2.8   2.8   1.0   0.6   0.8  

10   4.4   6.0   4.8   2.2   2.5   2.6   -­‐1.6   -­‐0.3   -­‐1.0  11   6.2   7.1   6.3   2.7   2.5   2.5   -­‐0.9   -­‐0.1   -­‐0.5  

Table  2:    Results  of  analysis,  displayed  by  bin  location.    Shown  are  average  ranks  of  accumulation  rates  for  each  bin  type  (left)  and  standard  deviations  in  these  averages  (center).    Higher  numbers  indicate  lower-­‐ranked  bins,  and  

therefore  lower  average  accumulations.    At  right  differences  between  average  ranks  of  accumulations  of  trash  and  those  of  recyclables  (T-­‐R)  and  compost  (T-­‐C)  are  shown;  more  positive  numbers  indicate  higher  rates  of  composting  and  recycling  compared  to  discard  in  landfill-­‐bound  trash  bins,  while  more  negative  numbers  

represent  higher  rates  of  discard  in  landfill  trash  bins  relative  to  composting  and  recycling.    T-­‐R|C  represents  the  average  of  T-­‐R  and  T-­‐C  scores,  and  provides  a  quick  summary  of  a  location’s  relative  overall  waste  diversion  rate.  

 

 Figure  8:    A  map  of  Husky  Den  waste  bin  locations  showing  T-­‐R|C  scores  for  each  location  (the  far  right  column  of  Tables  1  and  2).    Importantly,  high  and/or  low  values  are  not  strongly  associated  with  any  given  region,  implying  location  within  the  Husky  Den  is  not  a  primary  influence  on  the  way  users  categorize  waste  upon  deposition.    This  

map  is  modified  from  an  image  provided  by  Steve  Smith  (UW  CAP).