Analysis of the questionnaire

require(nnet)

d<-read.table("table_reponses_qu.csv", h=T, sep=";")

summary(d)

       X              Copie       Note          age          augmenter      changer_hab    changer_hab_alim
 Min.   :  0.00   1:10   :  5   0   : 41   Min.   :0.000   Min.   :0.000   Min.   :0.000   Min.   :0.000   
 1st Qu.: 33.75   1:11   :  5   1,5 : 73   1st Qu.:2.000   1st Qu.:1.000   1st Qu.:4.000   1st Qu.:3.000   
 Median : 67.50   1:12   :  5   15  :104   Median :3.000   Median :1.000   Median :4.000   Median :4.000   
 Mean   : 74.66   1:13   :  5   16,5: 84   Mean   :2.914   Mean   :1.676   Mean   :4.103   Mean   :3.968   
 3rd Qu.:113.25   1:14   :  5   18  : 28   3rd Qu.:4.000   3rd Qu.:2.000   3rd Qu.:5.000   3rd Qu.:5.000   
 Max.   :184.00   1:16   :  5   3   : 58   Max.   :6.000   Max.   :6.000   Max.   :6.000   Max.   :6.000   
                  (Other):378   4,5 : 20                                                                   
  communiquer    comprendre_logo conscience_empreinte deja_renseigner    diminuer         env       
 Min.   :0.000   Min.   :0.000   Min.   :0.000        Min.   :0.000   Min.   :0.00   Min.   :0.000  
 1st Qu.:2.000   1st Qu.:1.000   1st Qu.:3.000        1st Qu.:2.000   1st Qu.:3.00   1st Qu.:2.000  
 Median :3.000   Median :1.000   Median :4.000        Median :3.000   Median :4.00   Median :3.000  
 Mean   :3.145   Mean   :1.172   Mean   :3.642        Mean   :3.282   Mean   :3.98   Mean   :3.265  
 3rd Qu.:4.000   3rd Qu.:1.000   3rd Qu.:4.000        3rd Qu.:4.000   3rd Qu.:5.00   3rd Qu.:4.000  
 Max.   :6.000   Max.   :3.000   Max.   :6.000        Max.   :6.000   Max.   :6.00   Max.   :6.000  
                                                                                                    
     envie       favorable_affichage      freq       fruits_legumes   futurs_logos        genre      
 Min.   :0.000   Min.   :0.000       Min.   :0.000   Min.   :0.000   Min.   :     0   Min.   :0.000  
 1st Qu.:3.000   1st Qu.:1.000       1st Qu.:1.000   1st Qu.:1.000   1st Qu.:   100   1st Qu.:1.000  
 Median :4.000   Median :1.000       Median :1.000   Median :1.000   Median : 10100   Median :2.000  
 Mean   :3.426   Mean   :1.289       Mean   :1.735   Mean   :1.662   Mean   : 45754   Mean   :1.632  
 3rd Qu.:4.000   3rd Qu.:1.000       3rd Qu.:2.000   3rd Qu.:2.000   3rd Qu.:100110   3rd Qu.:2.000  
 Max.   :6.000   Max.   :3.000       Max.   :5.000   Max.   :5.000   Max.   :111110   Max.   :3.000  
                                                                                                     
      gout         infos_com         infos_logo     mal_retrouver   notion_budget_carbone
 Min.   :0.000   Min.   :      0   Min.   :     0   Min.   :0.000   Min.   :0.000        
 1st Qu.:4.000   1st Qu.: 100000   1st Qu.:    11   1st Qu.:2.000   1st Qu.:1.000        
 Median :4.000   Median :1000000   Median :   111   Median :3.000   Median :2.000        
 Mean   :4.049   Mean   : 556225   Mean   : 29339   Mean   :2.983   Mean   :1.583        
 3rd Qu.:5.000   3rd Qu.:1000000   3rd Qu.:100000   3rd Qu.:4.000   3rd Qu.:2.000        
 Max.   :6.000   Max.   :1111111   Max.   :111111   Max.   :6.000   Max.   :3.000        
                                                                                         
 notion_empreinte_carbone   nutrition        origine       part_chacun    pas_attention     poisson     
 Min.   :0.000            Min.   :0.000   Min.   :0.000   Min.   :0.000   Min.   :0.00   Min.   :0.000  
 1st Qu.:1.000            1st Qu.:3.000   1st Qu.:2.000   1st Qu.:4.000   1st Qu.:1.00   1st Qu.:4.000  
 Median :1.000            Median :4.000   Median :3.000   Median :5.000   Median :2.00   Median :4.000  
 Mean   :1.225            Mean   :3.699   Mean   :3.267   Mean   :4.328   Mean   :1.98   Mean   :4.017  
 3rd Qu.:1.000            3rd Qu.:5.000   3rd Qu.:4.000   3rd Qu.:5.000   3rd Qu.:2.25   3rd Qu.:5.000  
 Max.   :3.000            Max.   :6.000   Max.   :6.000   Max.   :6.000   Max.   :6.00   Max.   :6.000  
                                                                                                        
      prix       produits_laitiers_oeuf   profession        regime      remarquer_logos sensible_climat
 Min.   :0.000   Min.   :0.000          Min.   :0.000   Min.   :0.000   Min.   :0.000   Min.   :0.000  
 1st Qu.:3.000   1st Qu.:2.000          1st Qu.:1.000   1st Qu.:2.000   1st Qu.:1.000   1st Qu.:4.000  
 Median :3.000   Median :2.000          Median :1.000   Median :3.000   Median :1.000   Median :5.000  
 Mean   :3.319   Mean   :2.542          Mean   :1.936   Mean   :3.277   Mean   :1.184   Mean   :4.468  
 3rd Qu.:4.000   3rd Qu.:3.000          3rd Qu.:2.250   3rd Qu.:5.000   3rd Qu.:1.000   3rd Qu.:5.000  
 Max.   :6.000   Max.   :6.000          Max.   :5.000   Max.   :6.000   Max.   :2.000   Max.   :6.000  
                                                                                                       
     temps      tenir_compte_logos viande_blanche   viande_rouge     resto     Nom    
 Min.   :0.00   Min.   :0.000      Min.   :0.000   Min.   :0.000   Domus:185    :223  
 1st Qu.:3.00   1st Qu.:1.000      1st Qu.:3.000   1st Qu.:3.000   ECL  :145   ?:185  
 Median :4.00   Median :2.000      Median :4.000   Median :4.000   Puvis: 78          
 Mean   :3.39   Mean   :1.855      Mean   :3.647   Mean   :3.858                      
 3rd Qu.:4.00   3rd Qu.:3.000      3rd Qu.:4.000   3rd Qu.:5.000                      
 Max.   :6.00   Max.   :3.000      Max.   :6.000   Max.   :6.000                      
                                                                                      

Attempting to use GLMs to study interactions between responses to the questionnaire

Removing the “mal renseigné” from “tenir_compte” and other columns

d<-d[which(d$tenir_compte_logos!=0),]
d$tenir_compte_logos <- factor(d$tenir_compte_logos, levels=sort(unique(d$tenir_compte_logos)))
d<-d[which(d$resto!=0),]
d$resto <- factor(d$resto, levels=sort(unique(d$resto)))
d<-d[which(d$sensible_climat!=0),]
d$sensible_climat <- factor(d$sensible_climat, levels=sort(unique(d$sensible_climat)))
d<-d[which(d$changer_hab_alim!=0),]
d$changer_hab_alim <- factor(d$changer_hab_alim, levels=sort(unique(d$changer_hab_alim)))
d<-d[which(d$profession!=0),]
d$profession <- factor(d$profession, levels=sort(unique(d$profession)))
d<-d[which(d$genre!=0),]
d$genre <- factor(d$genre, levels=sort(unique(d$genre)))
d<-d[which(d$age!=0),]
d$age <- factor(d$age, levels=sort(unique(d$age)))

Defining a reference level for “tenir_compte”

d$tenir_compte_logos_2 <- relevel(d$tenir_compte_logos, ref = "1") #d$tenir_compte_logos  

Function to do a multinomial regression

perform_multinomial_regression <- function (formule, d) {
  formule2<- as.formula(formule)
  tenir_compte <- multinom(formula=formule2, data=d)
  tmp <- summary(tenir_compte)
  print(tmp)
  return(tmp$AIC)
}
### DOES NOT WORK FOR SOME REASON

Which users take into account the logo?

Constant model: no variable has an impact on the results

formula_tenir_compte_const <- as.formula(tenir_compte_logos_2 ~ 1 )
tenir_compte_const <- summary(multinom(formula=formula_tenir_compte_const, data=d))

# weights:  6 (2 variable)
initial  value 346.062871 
final  value 337.998183 
converged

AIC_const <- tenir_compte_const$AIC
                    

formula_tenir_compte_genre <- as.formula(tenir_compte_logos_2 ~ genre )
                                   
tenir_compte_genre <- summary(multinom(formula=formula_tenir_compte_genre, data=d))

# weights:  12 (6 variable)
initial  value 346.062871 
iter  10 value 334.633284
final  value 334.569171 
converged

AIC_genre <- tenir_compte_genre$AIC

formula_tenir_compte_resto <- as.formula(tenir_compte_logos_2 ~ resto )
tenir_compte_resto <- summary(multinom(formula=formula_tenir_compte_resto, data=d))

# weights:  12 (6 variable)
initial  value 346.062871 
iter  10 value 331.672383
final  value 331.668634 
converged

AIC_resto <- tenir_compte_resto$AIC

formula_tenir_compte_age <- as.formula(tenir_compte_logos_2 ~ age )
tenir_compte_age <- summary(multinom(formula=formula_tenir_compte_age, data=d))

# weights:  21 (12 variable)
initial  value 346.062871 
iter  10 value 331.178347
iter  20 value 330.885758
final  value 330.884143 
converged

AIC_age <- tenir_compte_age$AIC

formula_tenir_compte_profession <- as.formula(tenir_compte_logos_2 ~ profession )
tenir_compte_profession <- summary(multinom(formula=formula_tenir_compte_profession, data=d))

# weights:  18 (10 variable)
initial  value 346.062871 
iter  10 value 334.906188
final  value 334.904657 
converged

AIC_profession <- tenir_compte_profession$AIC

formula_tenir_compte_sensible_climat <- as.formula(tenir_compte_logos_2 ~ sensible_climat )
tenir_compte_sensible_climat <- summary(multinom(formula=formula_tenir_compte_sensible_climat, data=d))

# weights:  21 (12 variable)
initial  value 346.062871 
iter  10 value 320.483239
iter  20 value 319.983810
final  value 319.981651 
converged

AIC_sensible_climat <- tenir_compte_sensible_climat$AIC

formula_tenir_compte_changer_hab_alim <- as.formula(tenir_compte_logos_2 ~ changer_hab_alim )
tenir_compte_changer_hab_alim <- summary(multinom(formula=formula_tenir_compte_changer_hab_alim, data=d))

# weights:  21 (12 variable)
initial  value 346.062871 
iter  10 value 311.533245
iter  20 value 309.101188
final  value 309.095624 
converged

AIC_changer_hab_alim <- tenir_compte_changer_hab_alim$AIC

Comparison of AICs for individual variables

AICs_single <- c("AIC_const", "AIC_genre", "AIC_resto", "AIC_age", "AIC_profession", "AIC_sensible_climat", "AIC_changer_hab_alim")
for (i in 1:length(AICs_single)) { cat( AICs_single[i], get(AICs_single[i]), sep=" : ", fill=TRUE) }

AIC_const : 679.9964
AIC_genre : 681.1383
AIC_resto : 675.3373
AIC_age : 685.7683
AIC_profession : 689.8093
AIC_sensible_climat : 659.9633
AIC_changer_hab_alim : 642.1912

The variable which best explains whether the logo was used or not is changer_hab_alim, then it is sensible_climat, then resto. The other variables, including profession and age, do not explain whether people used the logo or not.

Model with several variables

formula_tenir_compte_several <- as.formula(tenir_compte_logos_2 ~ changer_hab_alim + sensible_climat + resto )
tenir_compte_several <- summary(multinom(formula=formula_tenir_compte_several, data=d))

# weights:  42 (26 variable)
initial  value 346.062871 
iter  10 value 298.085207
iter  20 value 290.902041
iter  30 value 290.526439
final  value 290.525167 
converged

NaNs produced

AIC_several <- tenir_compte_several$AIC
print(AIC_several)

[1] 629.0503

The model including the variables changer_hab_alim, sensible_climat and resto is the best one in terms of AIC.

print(tenir_compte_several)

Call:
multinom(formula = formula_tenir_compte_several, data = d)

Coefficients:
  (Intercept) changer_hab_alim2 changer_hab_alim3 changer_hab_alim4 changer_hab_alim5 changer_hab_alim6
2   27.436726         -2.759198        -36.736410        -38.115255        -38.894699         -12.88578
3    9.780258         39.778634          6.727016          5.950062          5.222356          31.83942
  sensible_climat2 sensible_climat3 sensible_climat4 sensible_climat5 sensible_climat6   restoECL restoPuvis
2         21.39824         11.71866         13.02542         11.24025                0 -1.1497801  1.1885035
3        -15.71478        -15.82012        -13.71854        -15.18606                0 -0.4577682  0.6418063

Std. Errors:
  (Intercept) changer_hab_alim2 changer_hab_alim3 changer_hab_alim4 changer_hab_alim5 changer_hab_alim6
2    51.72707          10.07722          10.06206          10.05715          10.05683          10.09059
3    51.84498          10.07722          10.10186          10.09654          10.09618          10.09059
  sensible_climat2 sensible_climat3 sensible_climat4 sensible_climat5 sensible_climat6  restoECL restoPuvis
2         235.6312         61.78386         61.78018         61.77914              NaN 0.3724295  0.4912884
3         235.6312         61.94302         61.93747         61.93648                0 0.3437635  0.5187916

Residual Deviance: 581.0503 
AIC: 629.0503 

As the users declare that they are willing to change their food habits, we see that they are less likely to have never used the logos. Users at the Puvis restaurant are less likely to use the logo than those at ECL.

Do people who care about the climate eat less red meat?

d<-read.table("table_reponses_qu.csv", h=T, sep=";")

d<-d[which(d$viande_rouge!=0),]
d$viande_rouge <- factor(d$viande_rouge, levels=sort(unique(d$viande_rouge)))
d<-d[which(d$sensible_climat!=0),]
d$sensible_climat <- factor(d$sensible_climat, levels=sort(unique(d$sensible_climat)))
d<-d[which(d$changer_hab_alim!=0),]
d$changer_hab_alim <- factor(d$changer_hab_alim, levels=sort(unique(d$changer_hab_alim)))
d<-d[which(d$resto!=0),]
d$resto <- factor(d$resto, levels=sort(unique(d$resto)))
d<-d[which(d$profession!=0),]
d$profession <- factor(d$profession, levels=sort(unique(d$profession)))
d<-d[which(d$genre!=0),]
d$genre <- factor(d$genre, levels=sort(unique(d$genre)))
d<-d[which(d$age!=0),]
d$age <- factor(d$age, levels=sort(unique(d$age)))
# d<-d[which(d$viande_blanche!=0),]
# d$viande_blanche <- factor(d$viande_blanche, levels=sort(unique(d$viande_blanche)))
# d<-d[which(d$produits_laitiers_oeuf!=0),]
# d$produits_laitiers_oeuf <- factor(d$produits_laitiers_oeuf, levels=sort(unique(d$produits_laitiers_oeuf)))
# d<-d[which(d$poisson!=0),]
# d$poisson <- factor(d$poisson, levels=sort(unique(d$poisson)))

Defining a reference level for “viande_rouge”

d$viande_rouge <- relevel(d$viande_rouge, ref = "1") 

The reference is “eating red meat at all meals”.

Constant model

formula_tenir_compte_constant <- as.formula(viande_rouge ~ 1 )
tenir_compte_constant <- summary(multinom(formula=formula_tenir_compte_constant, data=d))

# weights:  12 (5 variable)
initial  value 596.655903 
iter  10 value 468.997833
final  value 468.698118 
converged

AIC_constant <- tenir_compte_constant$AIC
print(AIC_constant)

[1] 947.3962

Models with one variable at a time

formula_tenir_compte_age <- as.formula(viande_rouge ~ age )
tenir_compte_age <- summary(multinom(formula=formula_tenir_compte_age, data=d))

# weights:  42 (30 variable)
initial  value 596.655903 
iter  10 value 463.717558
iter  20 value 452.620490
iter  30 value 451.696034
iter  40 value 451.682402
final  value 451.682342 
converged

AIC_age <- tenir_compte_age$AIC
print(AIC_age)

[1] 963.3647

formula_tenir_compte_genre <- as.formula(viande_rouge ~ genre )
tenir_compte_genre <- summary(multinom(formula=formula_tenir_compte_genre, data=d))

# weights:  24 (15 variable)
initial  value 596.655903 
iter  10 value 462.381838
iter  20 value 460.796991
iter  30 value 460.775659
final  value 460.775475 
converged

AIC_genre <- tenir_compte_genre$AIC
print(AIC_genre)

[1] 951.551

formula_tenir_compte_profession <- as.formula(viande_rouge ~ profession )
tenir_compte_profession <- summary(multinom(formula=formula_tenir_compte_profession, data=d))

# weights:  36 (25 variable)
initial  value 596.655903 
iter  10 value 464.058501
iter  20 value 457.277260
iter  30 value 456.490542
iter  40 value 456.420038
final  value 456.419545 
converged

AIC_profession <- tenir_compte_profession$AIC
print(AIC_profession)

[1] 962.8391

formula_tenir_compte_resto <- as.formula(viande_rouge ~ resto )
tenir_compte_resto <- summary(multinom(formula=formula_tenir_compte_resto, data=d))

# weights:  24 (15 variable)
initial  value 596.655903 
iter  10 value 469.504075
iter  20 value 465.040223
final  value 464.996426 
converged

AIC_resto <- tenir_compte_resto$AIC
print(AIC_resto)

[1] 959.9929

formula_tenir_compte_sensible_climat <- as.formula(viande_rouge ~ sensible_climat )
tenir_compte_sensible_climat <- summary(multinom(formula=formula_tenir_compte_sensible_climat, data=d))

# weights:  42 (30 variable)
initial  value 596.655903 
iter  10 value 454.838216
iter  20 value 445.389019
iter  30 value 444.359666
iter  40 value 444.346883
final  value 444.346791 
converged

NaNs produced

AIC_sensible_climat <- tenir_compte_sensible_climat$AIC
print(AIC_sensible_climat)

[1] 938.6936

formula_tenir_compte_changer_hab_alim <- as.formula(viande_rouge ~ changer_hab_alim )
tenir_compte_changer_hab_alim <- summary(multinom(formula=formula_tenir_compte_changer_hab_alim, data=d))

# weights:  42 (30 variable)
initial  value 596.655903 
iter  10 value 446.213951
iter  20 value 431.584612
iter  30 value 430.558039
iter  40 value 430.526262
final  value 430.526004 
converged

AIC_several <- tenir_compte_changer_hab_alim$AIC
print(AIC_changer_hab_alim)

[1] 642.1912

Comparison of AICs for individual variables

AICs_single <- c("AIC_constant", "AIC_genre", "AIC_resto", "AIC_age", "AIC_profession", "AIC_sensible_climat", "AIC_changer_hab_alim")
for (i in 1:length(AICs_single)) { cat( AICs_single[i], get(AICs_single[i]), sep=" : ", fill=TRUE) }

AIC_constant : 947.3962
AIC_genre : 951.551
AIC_resto : 959.9929
AIC_age : 963.3647
AIC_profession : 962.8391
AIC_sensible_climat : 938.6936
AIC_changer_hab_alim : 642.1912

The variable with the best explanatory variable for the consumption of red meat is changer_hab_alim, then it is sensible_climat. The other variables do not help.

Model with several variables

formula_tenir_compte_several <- as.formula(viande_rouge ~ changer_hab_alim + sensible_climat  )
tenir_compte_several <- summary(multinom(formula=formula_tenir_compte_several, data=d))

# weights:  72 (55 variable)
initial  value 596.655903 
iter  10 value 440.273626
iter  20 value 420.446825
iter  30 value 415.925200
iter  40 value 414.930011
iter  50 value 414.806924
iter  60 value 414.803020
final  value 414.802998 
converged

NaNs produced

AIC_several <- tenir_compte_several$AIC
print(AIC_several)

[1] 929.606

The combined model is not as good as the model with just changer_hab_alim. Let’s have a look at the parameters for this latter model.

print(tenir_compte_changer_hab_alim)

Call:
multinom(formula = formula_tenir_compte_changer_hab_alim, data = d)

Coefficients:
  (Intercept) changer_hab_alim2 changer_hab_alim3 changer_hab_alim4 changer_hab_alim5 changer_hab_alim6
2    13.80936         -13.80956         -12.71113        -13.810580          12.28553        -15.717679
3    14.09688         -12.71073         -11.00615        -10.919538          14.07735         15.902997
4    13.80931         -12.71089         -11.03701         -9.821037          15.68669         16.190449
5    13.40346         -12.30498         -11.61193        -10.226024          15.82689         -9.992251
6   -11.29461         -17.76082          12.39302         11.987203          39.95439         -2.553339

Std. Errors:
  (Intercept) changer_hab_alim2 changer_hab_alim3 changer_hab_alim4 changer_hab_alim5 changer_hab_alim6
2   0.6741687      1.361328e+00         1.1141406         1.3520570         0.7584673      1.551454e-14
3   0.6425439      1.073178e+00         0.9728213         0.9710066         0.5307285      8.036337e-01
4   0.6741742      1.125470e+00         1.0015110         0.9819046         0.5449056      8.036337e-01
5   0.7334394      1.161934e+00         1.0853931         1.0334026         0.6190845      5.142881e-12
6   0.5151042      1.082418e-12         0.8879311         0.9307621         0.3093012      1.047816e-19

Residual Deviance: 861.052 
AIC: 921.052 

It seems like people who say they are “tout-à-fait d’accord” with changing their food habits tend to eat less red meat.