Results

To  test  the  hypothesis  it  was  necessary  to  create  two  new  variables.  These  were  defined  as   the  score  of  the  difference  between  the  two  different  packaging  rating  of  Bellavista.  With  this   method   we   can   see   if   a   change   in   the   self-­‐image   congruence   and   in   the   expression   of   preference   has   occurred.   The   new   variables   were   then   named   as   SELFCHANGE   and   PREFERENCECHANGE  and  of  course  they  can  have  either  negative  value  in  case  of  change  or   they  can  be  zero  in  the  case  where  no  change  has  occurred.  

I’ve  started  testing  the  hypothesis  that  self-­‐image  congruence  and  preference  were  correlated.  

To   do   so   I   have   employed   the   Person   r   method,   which   is   used   to   measure   the   correlation   between  two  variables;  it  varies  from  0  to  1,  the  higher  it  is  the  more  correlated  the  variables   are.  (Bryman  and  Bell,  pg  347)  Measuring  the  correlation  in  this  case  is  not  enough,  we  have   also  to  check  if  these  correlations  are  statically  significant,  in  other  words  we  have  to  see  if  we   can  trust  the  results  obtained.  All  of  four  brands  hold  a  high  statistical  significance;  the  level   was  at  99%  (alpha  0,01).    

a)  Bellavista  with  the  old  packaging:  0,534.  

b)  Ca’  del  bosco:  0,623   c)  Ferrari:  0,790  

d)  Bellavista  with  the  new  packaging:  0,768  

e)  The  variable  about  the  preference  change  and  the  self-­‐change  where  positively  correlated   with  Pearson’s  R  of  0,626.  

After   checking   the   correlation   between   the   two   principal   variables,   the   competitors’   scores   were   not   included   in   the   analysis   anymore.   Their   images   and   the   related   questions   were   placed  between  the  two  different  Bellavista’s  packaging;  the  reason  was  to  give  respondents  

something  to  think  about  and  to  free  their  mind.  The  mental  process  was  then  more  fluid  and   thus   ensured   the   correct   validity   of   the   results.   Another   reason   for   including   the   two   competitors   was   also   to   force   consumers   to   concentrate   on   the   questions   and   thus   to   give   preference   scores   not   in   absolute   but   in   relation   with   other   two   brands.    

The   four   items   about   Bellavista   were   at   fist   examined   in   order   to   analyse   the   given   hypotheses.  

Firstly,  the  variables  about  the  self-­‐image  congruence  were  tested  with  the  Wilcoxon  signed-­‐

rank  test;  it  is  a  non-­‐parametric  test  that  has  to  be  used  when  it  is  needed  to  compare  two   related  sample  to  assess  if  the  population  has  expressed  a  different  mean.  It  was  found  that   they  were  significant  with  the  p  value  of  0,05.  

Using  the  same  method,  the  two  variables  regarding  the  expression  of  preference  were  found   not  statistically  significant  with  alpha  0,05;  this  shows  us  a  problem  that  has  to  be  evaluated.  

The   Wilcoxon   test   has   evinced   that   the   change   in   the   self-­‐image   variables   is   statistically  

significant   while   the   change   in   the   preference   scores   is   not,   this   suggest   not   accepting   the   hypothesis  and  thus  ending  the  analysis  of  the  survey.  

From  a  different  perspective,  it  is  worth  to  continue  the  research  and  to  test  the  correlation   between   the   self-­‐image   change   and   the   preference   change.   Using   again   the   calculus   of   the   Pearson’s  R  for  the  correlations  it  is  found  that  their  value  is  statistically  significant  with  a  p   value   of   0,01   and   their   correlation   score   is   0,626   which   means   that   the   two   variables   are   extremely   positively   correlated   and   thus   a   change   in   the   self   image   congruence   goes   along   with  a  change  in  the  preference  expressed  by  respondents.  

Furthermore,  if  we  take  into  account  the  frequency  tables  of  these  two  variables  we  can  note   that   the   ties   are   almost   equal   (66   for   the   self   image,   while   70   for   the   preference)   but   the   difference   is   given   by   the   tendency   of   the   respondents   to   have   a   greater   change   in   the   self   than   in   the   preference   variable.   That   is   maybe   due   to   the   fact   that   when   they   answer   the   survey,   they   first   look   at   the   image   thinking   about   the   congruence   between   them   and   the   product  user  image.    

Having  213  valid  responses  we  have  found  that  a  change  for  the  self-­‐image  has  occurred  for   the  69%  of  them,  27,7%  negative  change  and  41,3%  positive  change,  the  ties  represent  the   31%.  

Regarding  the  preference  scores,  we  have  got  the  same  valid  responses  but  the  data  differs   because   the   changes   are   the   67,1%,   negative   28,6%   and   positive   39,5%,   while   the   ties   are   32,9%.  

After   the   acceptation   of   all   the   hypotheses   I   have   conducted   further   studies   trying   to   understand  which  variables  are  linked  and  influence  the  self-­‐image  and  the  preference.  I  have   employed   some   linear   generalized   model   to   find   out   that   the   only   variable   statically   significant   was   the   age.   It   was   found   that   it   influences   the   self-­‐image   positively.  

The  correlation  between  the  age  and  the  variables  of  self-­‐image  are  all  positive  except  for  the   value  that  includes  the  change  between  new  and  old  Bellavista’s  packaging.  The  Pearson’s  R   obtained   was   -­‐0,194   statically   significant   for   the   p   value   of   0,01;   this   means   that   when   the   self-­‐image  decreases  the  age  of  the  respondents  increase  and  vice  versa.  

The   hypotheses   formulated   in   the   third   chapter   were   then   confirmed:  

H1:   A   change   in   the   self-­‐image   congruence   is   correlated   to   the   change   in   the   expression   of   preference.  

H2:  If  the  packaging  change  is  positively  evaluated,  the  consumer’s  self-­‐image  congruence  score   of   the   new   packaging   will   be   higher   than   for   the   old   packaging.   This   will   result   in   a   greater   preference  toward  the  brand.  

H3:  If  the  packaging  change  is  negatively  evaluated,  the  consumer’s  self-­‐image  congruence  score   of   the   new   packaging   will   be   lower   than   for   the   old   packaging;   consequently   the   preference   toward  the  brand  will  diminish.