Different results Gamma generalized linear model R and SPSS

Question

Update: The p-values and S.E.'s are similar between SPSS and R if I change the parameter estimation method in SPSS to 'Hybrid' and the scale parameter method to 'Pearson Chi-square'. Does anyone now how to change these settings in R and what these settings actually mean?

I am trying to perform an GLM with a gamma log link function in R, to analyse a multiple imputation dataset.

However, when I compare the results from the same analysis in R and SPSS they are very different. This example is in a non-imputation dataset to make things easier to interpret. The SPSS result is as follows:

Parameter Estimates                         
Parameter   B   Std. Error  95% Wald Confidence Interval          Hypothesis Test       
        Lower   Upper   Wald Chi-Square df  Sig.
(Intercept) 3,263   ,2499   2,774   3,753   170,571 1   ,000
[Comorb=1]  -,631   ,1335   -,893   -,369   22,331  1   ,000
[Comorb=2]  -,371   ,1473   -,660   -,083   6,358   1   ,012
[Comorb=3]  0a  .   .   .   .   .   .
PAIDhoog     ,257   ,1283   ,006    ,509    4,023   1   ,045
PHQhoog    ,039 ,1504   -,256   ,334    ,068    1       ,794
[etndich=1,00]  -,085   ,1125   -,306   ,135    ,575    1   ,448
[etndich=2,00]  0a  .   .   .   .   .   .
Leeftijd    ,009    ,0035   ,002    ,016    6,588   1   ,010
(Scale) ,613b   ,0470   ,528    ,712            
Dependent Variable: totaalhealthcareutilization
Model: (Intercept), Comorb, PAIDhoog, PHQhoog, etndich, Leeftijd                            
a Set to zero because this parameter is redundant.                          
b Maximum likelihood estimate.

While the same analysis in R yields this result:

  Call:
glm(formula = (totaalhealthcareutilization) ~ PAIDhoog + PHQhoog + 
   Comorb + Leeftijd + etndich, family = Gamma(link = log), 
   data = F)

Deviance Residuals: 
   Min       1Q   Median       3Q      Max  
-2.1297  -0.7231  -0.3018   0.2075   3.1365  

Coefficients:
         Estimate Std. Error t value             Pr(>|t|)    
(Intercept)  3.006208   0.273817  10.979 < 0.0000000000000002 ***
PAIDhoog     0.201881   0.131777   1.532               0.1264    
PHQhoog      0.126989   0.157416   0.807               0.4203    
Comorbgeen  -0.638842   0.144459  -4.422            0.0000128 ***
Comorb1     -0.348187   0.158484  -2.197               0.0286 *  
Leeftijd     0.007311   0.003534   2.069               0.0392 *  
etndich      0.151836   0.118872   1.277               0.2023    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for Gamma family taken to be 0.9432289)

    Null deviance: 286.49  on 381  degrees of freedom
Residual deviance: 243.01  on 375  degrees of freedom
  (71 observations deleted due to missingness)
AIC: 3156

Number of Fisher Scoring iterations: 6

How is this possible? The results keep differing, even if I use na.omit or na.exclude in R. I have used the function 'relevel' in R, to make sure that the same reference category is used for the categoric variables.

I hope you have any idea what I am doing wrong in R.

 This is what a sample of my data looks like: 

   verrichtingen verpleegkanders Leeftijd HbA1c  BMI Type_Treat DurationDM
1               0               0       26    69 26.7    Insulin          5
2               0               0       69    75 34.5    Insulin         17
3               0               0       67    62 24.3    Insulin          1
4               6               0       38    96   NA    Insulin         10
5               0               0       29    NA 19.1    Insulin         25
6               0               0       50    86 37.9       Both          9
7               1               0       29    44 29.1       Both         33
451             4               0       68   113 37.9       Both         11
452            21               1       57    62 21.5    Insulin          1
453             0               0       37    54 25.4       Both         14
         Socstatus PAID1 PAID2 PAID3 PAID4 PAID5 PAIDtot PHQ1 PHQ2 PHQ3
1    wel achterstandsw     0     1     2     1     0       4    0    0    0
2   geen achterstandsw     2     1     1     2     0       6    0    0    0
3                 <NA>     0     0     0     0     0       0    0    0    0
4   geen achterstandsw     0     0     1     1     0       2    1    0    3
5   geen achterstandsw     0     0     0     0     0       0    1    1    3
6    wel achterstandsw     0     1     0     2     0       3    2    0    3
7   geen achterstandsw     1     1     2     3     0       7    1    1    3
451 geen achterstandsw     0     0     1     0     0       1    0    0    0
452  wel achterstandsw     1     0     4     1     0       6    2    0    3
453  wel achterstandsw     1     1     2     3     2       9    1    0    1
PHQ4 PHQ5 PHQ6 PHQ7 PHQ8 PHQ9 Geslacht        Etnicit HAPOH Bedrijfsarts MW
1      1    1    0    0    0    0    vrouw     Overigwest    NA           NA  NA
2      1    0    0    1    1    0      man            Mar    NA           NA NA
3      0    0    0    0    0    0      man     Overigwest    NA           NA NA
4      3    1    1    1    1    0    vrouw Overignietwest    NA           NA NA
5      0    0    0    3    0    0      man     Overigwest    NA           NA NA
6      1    1    1    0    0    0      man           Turk    NA           NA NA
7      3    0    0    2    0    0    vrouw     Overigwest    NA           NA NA
451    0    0    0    0    0    0      man              4    NA           NA NA
452    3    0    0    1    0    0    vrouw            Mar    NA           NA NA
453    2    2    0    0    0    0    vrouw            Mar    NA           NA NA
FysioErgo Diet Psychiat Psychol Dvk VPtot Internist Specialist ICUopname
1          NA    5        0       0   5     5         2          3         0
2          NA    2        0       0   2     2         3          8         0
3          NA    0        0       0   1     1         2          3         0
4          NA    0        1       2  11    11         6         25         0
5          NA    0        0       0   4     4         2          6         0
6          NA    1        0       0   2     2         2          0         0
7          NA    3        0       0   4     4         2          3         0
451        NA    0        0       0   1     1         3          7         0
452        NA    2        0       0   4     5         0         25         4
453        NA    1        0       0   2     2         0          5         0
Opnamegewoon SEH Comorb DMtype PAIDtotaal PHQtotaal PAIDhoog PHQhoog
1              0   0   geen    DM1          4         2        0       0
2              0   0   geen    DM2          6         3        0       0
3              0   0   geen    DM1          0         0        0       0
 4              1   0   geen    DM2          2        NA        0      NA
5              0   0   geen    DM1          0         8        0       0
6              0   0   geen    DM2          3         8        0       0
7              0   0   geen    DM2          7        10        0       0
451           18   2   <NA>    DM2          1         0        0       0
452           34   3   <NA>    DM1          6         9        0       0
453            0   0   <NA>    DM2          9         6        1       0
interactPHQPAID paidtotaalimp PHQtotaalimp GADtotaalimp PAIDhoogimp
1                 0             4            2            1           0
2                 0             6            3            0           0
3                 0             0            0            0           0
4                 0             2           11            2           0
5                 0             0            8            0           0
6                 0             3            8            0           0
7                 0             7           10            3           0
451               0             1            0            0           0
452               0             6            9            0           0
453               0             9            6            1           1
PHQhoogimp GADimphoog kostenopnames kosteninternist kostenspecialist
1            0          0             0             160              240
2            0          0             0             240              640
3            0          0             0             160              240
4            0          0           443             480             2000
5            0          0             0             160              480
6            0          0             0             160                0
7            0          1             0             160              240
451          0          0          7974             240              560
452          0          0         15062               0             2000
453          0          0             0               0              400
kostenhuisarts kostenMW kostenfysioergo kostendvk kostendietist
1               NA       NA              NA       240           240
2               NA       NA              NA        96            96
3               NA       NA              NA        48             0
4               NA       NA              NA       528             0
5               NA       NA              NA       192             0
6               NA       NA              NA        96            48
7               NA       NA              NA       192           144
451             NA       NA              NA        48             0
452             NA       NA              NA       192            96
453             NA       NA              NA        96            48
totaalkosten jaarHAPOH jaarbedrijfsarts jaarMW jaarfysioergo
1             NA        NA               NA     NA            NA
2             NA        NA               NA     NA            NA
3             NA        NA               NA     NA            NA
4             NA        NA               NA     NA            NA
5             NA        NA               NA     NA            NA
6             NA        NA               NA     NA            NA
7             NA        NA               NA     NA            NA
451           NA        NA               NA     NA            NA
452           NA        NA               NA     NA            NA
453           NA        NA               NA     NA            NA
totaalverbruikjaar kostenHAjaar kostenMWjaar kostenjaarfysioergo
1                   NA           NA           NA                  NA
2                   NA           NA           NA                  NA
3                   NA           NA           NA                  NA
4                   NA           NA           NA                  NA
5                   NA           NA           NA                  NA
6                   NA           NA           NA                  NA
7                   NA           NA           NA                  NA
451                 NA           NA           NA                  NA
452                 NA           NA           NA                  NA
453                 NA           NA           NA                  NA
kostenopnameICU kostenpsycholoog kostenpsychiater kostenvpanders
1                 0                0                0              0
2                 0                0                0              0
3                 0                0                0              0
4                 0              188               94              0
5                 0                0                0              0
6                 0                0                0              0
7                 0                0                0              0
451               0                0                0              0
452            8060                0                0             48
453               0                0                0              0
kostenverrichtingen totaalutilization kostenseh totaalkostennieuw hypoangst
1                     0                NA         0               880         1
2                     0                NA         0              1072         1
3                     0                NA         0               448         0
4                   876                NA         0              4609         0
5                     0                NA         0               832         0
6                     0                NA         0               304         1
7                   146                NA         0               882         5
451                 584                NA       518              9924         0
452                3066                NA       777             29301         0
453                   0                NA         0               544         3
contactprimarycare contactsecondarycare totaalhealthcareutilization
1                   NA                   15                          15
2                   NA                   15                          15
3                   NA                    6                           6
4                   NA                   52                          52
5                   NA                   12                          12
6                   NA                    5                           5
7                   NA                   13                          13
451                 NA                   35                          35
452                 NA                   94                          94
453                 NA                    8                           8
kostenprimarycare kostensecondarycare totaalkostenhealthcare etndich
1                  NA                 880                     NA       1
2                  NA                1072                     NA       2
3                  NA                 448                     NA       1
4                  NA                4609                     NA       2
5                  NA                 832                     NA       1
6                  NA                 304                     NA       2
7                  NA                 882                     NA       1
451                NA                9924                     NA       1
452                NA               29301                     NA       2
453                NA                 544                     NA       2

My guess is that this has to do with the encoding of your categorical variables `Comorb` and `etndich`; in SPSS you have 3 levels for `Comorb`, and 2 levels for `etndich`; in R, you have 2 for `Comorb` and 1 for `etndich`. Can you clarify how many levels each categorical variable has? — Maurits Evers, Jul 19 '18 at 12:50
Looking at the SPSS output I suspect your mistake is there. `etndich` seems to be treated as a factor variable although its values are clearly intended to be numeric. — Roland, Jul 19 '18 at 12:50
@Roland I agree with you on the potential issue with `factor`s, but I'm not sure that `etndich` is supposed to be `numeric`; looks like a binary variable to me, with some sort of `1`/`2` encoding for two categories. — Maurits Evers, Jul 19 '18 at 12:53
@ Maurits: You are right: etndich, PAIDhoog and PHQhoog are all binary variables. Comorb has three levels (no comorbidities, 1, 2 or more). — Charlotte, Jul 19 '18 at 12:56
R does not show the category that serves as the reference. For Comorb for instance, the last category (2 or more) is not shown in the output. Is it supposed to be? It does have three levels in R. — Charlotte, Jul 19 '18 at 12:57
I have removed the Comorb and etndich variable, to check whether these explain the issue. Unfortunately the problem remains the same: Spss output: — Charlotte, Jul 19 '18 at 12:59
@user215051 In order for your design matrix to have full rank you need to exclude one category level as the baseline, against which you measure unit changes of your other levels. It's that or have a zero intercept. — Maurits Evers, Jul 19 '18 at 12:59
Your SPSS output suggests that the third `Comorb` level is the actual reference level. So you need to enforce the same reference level in R using `relevel`. Ditto for `etndich`. — Maurits Evers, Jul 19 '18 at 13:01
Thank you very much for your help Maurits. The reference category is the same for both SPSS and R: the last category is the reference (2 or more comorbidities, I have adjusted this via relevel in R). However, even if I only include continous variables in my model the problem remains. I will edit my question to show the output. — Charlotte, Jul 19 '18 at 13:07
@user215051 Hmm I see. Best to share data in a reproducible format. I've got a few ideas but I need actual data to work with. At the moment your dataset contains columns that are not relevant to the problem, and is quite unwieldy on account of not being in an easily copy&paste-able format. Can you revise your post to include a minimal dataset, and include SPSS and R outputs for that data? — Maurits Evers, Jul 19 '18 at 13:11
@user215051 You should include only relevant columns and then use `dput`; then copy&paste the output in your main post. To make your post even better, run SPSS and your R model on the reduced sample data and update results here. Then you have a fully reproducible minimal example:-) — Maurits Evers, Jul 19 '18 at 13:16
@user215051 Thanks for the data and code. Please see my answer below which reproduces your SPSS output. — Maurits Evers, Jul 19 '18 at 14:13

Maurits Evers · Answer 1 · 2018-07-20T15:29:12.113

The following reproduces your SPSS output.

Note, it's all a matter of setting the reference levels of the categorical variables correctly, to match the SPSS encoding. In R the first level will be used as the reference level.

df <- within(F, {
    Comorb <- relevel(Comorb, ref = "2 of meer");         # Reference level = "2 of meer"
    etndich <- factor(etndich, levels = 2:1);             # Reference level = 2
    PAIDhoog <- factor(PAIDhoog, levels = 1:0);           # Reference level = 1
    PHQhoog <- factor(PHQhoog, levels = 1:0);             # Reference level = 1
})

fit <- glm(formula = totaalhealthcareutilization ~ PAIDhoog + PHQhoog +
    Comorb + Leeftijd + etndich, family = Gamma(link = log),
    data = df)

summary(fit)
#
#Call:
#glm(formula = totaalhealthcareutilization ~ PAIDhoog + PHQhoog +
#    Comorb + Leeftijd + etndich, family = Gamma(link = log),
#    data = df)
#
#Deviance Residuals:
#    Min       1Q   Median       3Q      Max
#-2.1297  -0.7231  -0.3018   0.2075   3.1365
#
#Coefficients:
#             Estimate Std. Error t value Pr(>|t|)
#(Intercept)  3.638751   0.267741  13.591  < 2e-16 ***
#PAIDhoog0   -0.201881   0.131777  -1.532   0.1264
#PHQhoog0    -0.126989   0.157416  -0.807   0.4203
#Comorbgeen  -0.638842   0.144459  -4.422 1.28e-05 ***
#Comorb1     -0.348187   0.158484  -2.197   0.0286 *
#Leeftijd     0.007311   0.003534   2.069   0.0392 *
#etndich1    -0.151836   0.118872  -1.277   0.2023
#---
#Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#
#(Dispersion parameter for Gamma family taken to be 0.9432289)
#
#    Null deviance: 286.49  on 381  degrees of freedom
#Residual deviance: 243.01  on 375  degrees of freedom
#  (71 observations deleted due to missingness)
#AIC: 3156
#
#Number of Fisher Scoring iterations: 6

Compare with SPSS output

Parameter Estimates                         
Parameter   B   Std. Error  95% Wald Confidence Interval           Hypothesis Test      
        Lower   Upper   Wald Chi-Square df  Sig.
(Intercept) 3,639   ,2177   3,212   4,065   279,350 1   ,000
[PAIDhoog=0]    -,202   ,1056   -,409   ,005    3,657   1   ,056
 [PAIDhoog=1]   0a  .   .   .   .   .   .
[PHQhoog=0] -,127   ,1260   -,374   ,120    1,015   1   ,314
[PHQhoog=1] 0a  .   .   .   .   .   .
[Comorb=1]  -,639   ,1148   -,864   -,414   30,940  1   ,000
[Comorb=2]  -,348   ,1250   -,593   -,103   7,758   1   ,005
[Comorb=3]  0a  .   .   .   .   .   .
[etndich=1,00]  -,152   ,0936   -,335   ,032    2,633   1   ,105
[etndich=2,00]  0a  .   .   .   .   .   .
Leeftijd    ,007    ,0028   ,002    ,013    6,599   1   ,010
(Scale) ,581b   ,0387   ,510    ,662            
Dependent Variable: totaalhealthcareutilization
Model: (Intercept), PAIDhoog, PHQhoog, Comorb, etndich, Leeftijd                            
a Set to zero because this parameter is redundant.                          
b Maximum likelihood estimate.

Further comments on differences in the SPSS and `glm` output

The first thing to note is that parameter estimates from SPSS and R are identical: Both parameter sets correspond to the (unique) set of maximum likelihood (ML) estimates given the model and data.
In R, the standard errors are simply given as the square root of the diagonal elements of the estimated covariance matrix
```
sqrt(diag(vcov(fit)))
#(Intercept)   PAIDhoog0    PHQhoog0  Comorbgeen     Comorb1    Leeftijd
#0.267740656 0.131776659 0.157416176 0.144458874 0.158484265 0.003534017
#   etndich1
#0.118871533
```
Note that these values are identical to the reported se’s in summary(fit).

I don’t know SPSS, but it seems that SPSS' se's correspond to scaled square roots of the diagonal elements of the variance-covariance matrix.

Confidence intervals are based on parameter and variance-covariance estimates; as explained in the previous points, parameter estimates are identical, but SPSS uses a scaled variance-covariance matrix, so confidence intervals for the parameters in the SPSS and R output will be different according to said scaling factor.

SPSS' documentation is regrettably diffuse, so I'm not sure how SPSS scales its variance-covariance matrix.

Sample data

F <- structure(list(HbA1c = c(69, 75, 62, 96, NA, 86, 44, 49, NA, 63, 43, 75, 48, 56, 79, 78, 67, 66, 75, 67, 65, 66, 34, 62, 79, 60, 91, 51, 84, 72, 65, NA, NA, 62, 61, 69, 63, NA, 85, 38, 42, 80, 59, 96, 59, 49, 62, 98, 71, 78, 50, 43, 44, 69, 56, 38, 59, 74, 115, 69, 67, 51, NA, 107, 71, 86, 78, 41, 60, 59, 74, 73, 49, 34, 71, 57, 55, 74, 67, 61, 48, 59, 70, NA, 55, 72, 69, 82, 40, 58, NA, 53, 46, 69, 60, 39, 76, 69, 61, 86, 58, 63, 66, 103, 73, 54, 59, 46, 58, 70, 57, 53, 49, 53, 58, 71, 60, 76, 64, 97, 60, 49, 53, 44, 53, 73, 59, 75, 61, 55, 68, 56, 51, 91, 92, 76, 51, 55, 61, 83, 52, 62, 71, 75, 54, 64, 90, 65, NA, 69, 70, 70, 59, 62, 60, 63, 58, 58, 63, 60, 49, 62, 95, 42, 99, 67, 117, 68, 55, 55, 70, 60, 61, 91, 33, 89, 60, 47, 62, 72, 40, 88, 59, 56, 57, 59, 74, 41, 53, 76, 48, 73, 65, 96, 58, 55, 67, 45, 45, 69, 72, 44, 59, 43, 90, 69, 69, 71, 93, 42, 87, 54, 83, 60, 48, NA, 53, 56, 57, 77, 63, NA, 63, 60, 68, 51, 48, 65, 61, 79, 63, 62, 53, 67, 53, 53, 63, 55, 61, 51, 53, 46, NA, 78, 76, 73, 51, 49, 68, 86, 71, 55, 57, 113, 63, 68, 94, NA, 38, 50, NA, 42, 60, 57, 49, 60, 81, 69, 55, 82, 64, 55, 74, 71, 56, 60, NA, 47, 49, 98, 55, 80, 71, 69, 35, 53, 90, 64, 82, 132, 64, 70, 65, 34, 65, 54, NA, 68, 58, 76, 82, 66, 74, 66, NA, 54, 53, 78, 62, 88, 69, 49, 83, 54, 55, 56, 66, 84, 47, 82, 53, 62, 163, 41, 55, 89, 76, 81, 45, 50, 89, 72, 90, 47, 38, 83, NA, 53, 74, 55, 47, 49, 56, 74, 107, 86, 48, 59, 86, 44, 55, 64, 81, 66, 63, 98, 51, NA, 60, 50, 55, 52, 79, 58, 50, 89, NA, 36, 50, 70, NA, 86, 57, 60, 78, 53, 70, 79, 49, 78, 83, 66, 57, 62, 80, 70, NA, 67, 80, 46, 79, 47, 145, 87, 53, 65, 73, 75, 53, 50, 71, NA, 65, 106, 123, 51, 55, 43, 48, 86, 61, 64, 55, 71, 61, 96, 80, 69, 66, 74, 88, 48, 68, 55, 52, 58, 69, 66, 44, 45, 64, 84, 72, 49, NA, 71, 70, 104, 78, 73, 47, 75, 45, 57, 88, 86, 55, 72, 47, 53, 113, 62, 54), BMI = c(26.7, 34.5, 24.3, NA, 19.1, 37.9, 29.1, 27.1, NA, 21.1, 48.5, 26.2, 26.9, NA, 25.5, 25.3, 44.3, 25.2, 26.7, NA, 25.5, 25.9, 31.2, 33, 21.8, 23.7, 32, 23.6, 32.4, 29.7, NA, 22.9, 24.4, 33.9, 35.4, 41.2, 20.4, NA, 30.1, 21, NA, NA, 29.5, 16.6, 38.1, 23.9, 19.1, 35.4, 24.2, NA, 26.1, 20, 28.7, 30.7, 25.4, 29.6, 25.4, 26.2, 18.3, 31, NA, NA, 31.5, 32, 35.6, 24.3, 33.3, 35.5, NA, 24.1, NA, 33.4, 28.4, NA, 25.9, 26.7, 35.5, 31.6, 25, 25.5, 22.2, 22.3, 23.4, 35.3, 26.1, 32.6, 20.9, 35.9, 29.1, 32.8, 32.2, 28.9, 28.9, 28.8, 19.7, 29.4, 28.8, 28.2, 20.9, 33.5, 17.6, 38.6, 27.1, NA, 29, 25.6, 22.5, 30.6, 35.6, 32.5, 23.4, 27.2, 23.6, 26.6, 23.5, 30.3, 30.6, 26.4, 38.1, 34.7, NA, 24.6, 22.2, 39.8, 23, 35.8, 31.4, 22.8, 29.3, 27, 31.1, NA, NA, 32.4, 36, NA, 52.8, 22, 27.1, 23.3, 22.7, 25, 42.6, 30.2, 25.3, 30.5, 25.3, 28.4, 30.1, 32.4, NA, 32, 18.8, 23.1, 28.5, 25.1, 22.8, 23.6, 18.5, NA, 27.1, 25.3, 19.8, 20.8, 32.7, 30.1, 34.8, 37.5, NA, 28.1, 46, 23.5, 26.3, 22.2, 28.2, 29.3, 24.2, 29.7, 28.9, 28, 31.3, 28.6, 29.1, 28.4, 23.1, 34.9, 22.7, 26.9, 28.9, 35.9, 23, 25.8, 22.8, 19.2, 27.9, 29.2, 35, 25.1, 20.5, 23.9, 34.3, 23.1, 25.1, 20.5, 24.6, 24.4, 23.7, 22.4, 40.1, 21.9, 50, 34.2, 30.5, 20.7, 29.3, 32.6, 32.1, 23.9, NA, 34, 22.6, 30.2, 28.6, 27.5, 33, 24, 28.8, NA, 32.8, 21.8, NA, 37.8, 26.4, 36.2, 20.8, 24.4, 31, 31.9, 27.6, 25.4, 22.7, NA, 27.7, 32.4, 34, 26.2, 26.7, 23.7, 32, 24.1, 35.8, 23.5, 38.9, 35.3, NA, 23.9, 30.2, 24.4, 24.4, 27.9, NA, 25.7, 25.6, 25.8, 47.9, 25.6, 36.1, NA, 24.2, 24.8, 21.4, 22.3, 24.3, 24.7, 22.5, 25.9, 30.1, 27.4, 27.8, 22.6, 24.4, NA, 33.8, 41.9, 21.4, 32.5, 41.1, 27.2, NA, 37.8, 29, 23.2, 28.7, 25.2, 32.6, 29, 24.4, 23.1, 22.8, 23.1, 39.8, 26.6, 25.3, 53.5, 25, 22.9, 22.2, 30.2, 27.4, 27.4, NA, 25.2, 22.4, 20.2, 23.9, 23.3, 31.2, 24, 23.5, 38.8, 30, 30.6, 28.9, 23.1, 34.4, 28.7, 30.8, 21.6, 24.1, 25.5, 39.2, 29.3, 36.2, 28.3, NA, NA, NA, 29.5, 33.1, 23.4, 23.5, 25.1, 34.4, 24.5, 29.7, 22.2, 25.5, 23.3, 37.5, 26.8, 44.5, 32.4, 26.1, 21.4, 26.5, 32.7, 26.9, NA, 27.4, 36.3, 25.1, 37.7, NA, 27.6, 24.2, 46.9, 30.8, 29.3, 25.4, 35.7, 36.8, 35, 22.3, 28.3, 20.4, 25, 35, NA, 39.4, 25.2, 22.5, 34.5, NA, 21.6, 30.1, 25, NA, 28.3, 19.7, 22.3, 33.2, NA, 24.6, 23.9, 22.8, 24.1, 31.7, 28.4, 34.5, 30.1, 33.3, 28, 38, 35.9, 30.6, 33.5, 29.5, 21.4, 24.4, 27.5, 31.7, 23.8, NA, 21.8, 28.7, 33.5, 23.5, 27.3, 28.7, NA, 25.6, 26.7, 44.8, 26.2, 27.1, 39.7, 24.1, 21.3, 29.5, 30, NA, 27, NA, 23.6, 22.3, 32.6, 51.9, 27.7, 28.7, 35.2, 27.2, 29.6, 22.8, 19.6, 25.7, 28.3, 31.2, 21.7, 36.2, 26.9, 37.9, 21.5, 25.4), Comorb = structure(c(2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, NA, NA, NA), .Label = c("2 of meer", "geen", "1"), class = "factor"), PAIDhoog = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, NA, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0, 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 0, NA, NA, NA, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 1, 0, 0, 1, 0, 0, 1, 0, NA, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1), PHQhoog = c(0, 0, 0, NA, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, 0, 0, NA, 0, 0, 0, 0, 1, NA, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, NA, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, NA, 0, 0, 0, NA, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, NA, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, 0, NA, 0, 0, 0, 0, 0, 0, 0, NA, 0, 0, NA, 0, NA, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, NA, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, NA, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, NA, NA, 1, 0, 1, NA, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, NA, 0, 0, NA, NA, 0, NA, 0, 0, NA, 0, 0, 0, 0, 0, NA, 0, 0, NA, 0, 0, 1, 0, 0, NA, 0, 0, 0, 0, 1, 1, 0, 1, NA, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, NA, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, NA, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, NA, NA, 0, 0, 0, 1, 1, 0, 0, 0, 1, 1, 0, 0, 0, NA, 0, NA, 0, 0, 0, 0, NA, 0, NA, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, 1, 1, 1, 0, 1, NA, NA, 0, 1, 0, 0, 1, 1, NA, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, NA, 0, 0, 0, 0, 0, 0, 1, NA, 0, 0, NA, 0, 0, 0, 0, 0, 1, NA, 0, 1, 1, 0, 0, NA, 0, 0, 1, 1, 0, 0, 0, NA, 0, 1, 0, 0, 0, 0), totaalhealthcareutilization = c(15, 15, 6, 52, 12, 5, 13, 15, 13, 8, 10, 4, 9, 8, 6, 5, 8, 42, 15, 21, 6.3, 9, 5, 5, 14, 24, 8, 15, 25, 12, 29, 21, 6, 11, 8, 7, 29, 7, 7, 19, 14, 25, 16, 7, 20, 13, 17, 12, 5, NA, 9, 11, 14, 57, 12, 10, 37, 8, 12, 57, 8, 11, 14, 11, 49, 10, 10, 11, 19, 20, 21, 5, 1, 2, 2, 3, 3, 6, 4, 3, 4, 6, 5, 4, 4, 5, 7, 6, 6, 8, 5, 7, 8, 5, 6, 6, 6, 8, 7, 6, 6, 9, 11, 7, 9, 7, 7, 7, 7, 8, 10, 10, 10, 9, 9, 9, 11, 8, 10, 9, 9, 11, 13, 8, 12, 12, 9, 11, 7, 8, 10, 10, 9, 10, 10, 12, 12, 16, 9, 5, 10, 7, 13, 13, 13, 15, 16, 11, 11, 17, 13, 12, 22, 19, 15, 14, 11, 12, 19, 13, 15, 13, 14, 11, 17, 12, 17, 10, 13, 15, 12, 13, 13, 20, 16, 21, 17, 25, 22, 18, 18, 17, 15, 19, 10, 15, 20, 33, 22, 26, 23, 27, 20, 21, 21, 13, 24, 45, 27, 27, 19, 19, 25, 43, 16, 16, 13, 24, 29, 17, 24, 25, 32, 27, 29, 22, 35, 56, 26, 45, 23, 54, 26, 33, 23, 39, 35, 24, 36, 37, 37, 74, 53, 36, 60, 33, 35, 26, 44, 78, 22, 26, 77, 62, 121, 51, 28, 68, 63, 43, 64, 81, 120, 95, 98, 23, 11, 21, 10, 7, 41, 7, 33, 6, 40, 20, 2, 31, 23, 23, 13, 68, 9, 8, 41, 19, 27, 29, 46, NA, 35, 16, 12, 9, 14, 20, 7, 2, 4, 6, 6, 6, 4, 9, 6, 8, 9, 12, 9, 7, 8, 12, 11, 11, 14, 12, 14, 12, 16, 15, 22, 23, 19, 11, 12, 13, 17, 18, 19, 27, 15, 9, 17, 18, 19, 17, 19, 12, 16, 54, 21, 30, 23, 25, 24, 37, 35, 27, 47, 22, 27, 27, 30, 32, 32, 31, 39, 28, 36, 54, 50, 45, 42, 88, 56, 63, 82, 60, 70, 139, 122, 71, 130, 84, 33, 111, 111, 246, 157, 54, 24, 41, 22, 7, 33, 15, 9, 6, 16, 67, 3, 22, 48, 15, 57, 25, 48, 74, 40, 25, 18, 21, 3, 6, 7, 7, 14, 9, 11, 16, 14, 14, 14, 28, 18, 22, 21, 26, 39, 24, 22, 18, 22, 19, 19, 45, 15, 13, 22, 31, 29, 46, 37, 23, 35, 68, 39, 51, 35, 50, 80, 69, 51, 41, 90, 43, 32, 48, 34, 53, 25, 66, 39, 83, 70, 237, 81, 126, 95, 170, 35, 94, 8), etndich = c(1, 2, 1, 2, 1, 2, 1, 1, NA, 1, 1, 1, 1, 1, NA, 1, 1, 2, 1, 2, 1, 1, 1, 1, 1, 1, 1, 2, 2, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 2, 1, 1, 2, 2, 1, 1, 2, 1, 1, 2, 1, 1, 2, 2, NA, 2, 1, NA, 1, 1, 1, 1, 1, 1, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, NA, 1, 2, 1, 2, 1, 1, 1, 1, 2, 1, 1, 1, 2, 1, 2, NA, 2, 1, 2, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 2, 2, 1, 2, 1, 1, 1, 2, 1, 1, NA, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 2, 2, 1, 1, 1, 2, 2, 1, 2, 2, 1, 1, 1, 2, 1, 1, 1, 1, 1, 2, 2, 1, 1, 1, 2, 1, 1, 1, 1, NA, 2, 1, 1, 1, 1, 1, 2, 1, 1, 2, 1, 2, 1, 2, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 1, 1, 1, 1, 2, 1, 1, 2, 1, 1, 1, 1, 2, 1, 1, 2, 2, 1, 2, 1, 1, 1, 1, 1, 2, 1, 2, 2, 1, 1, 1, 1, NA, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 2, 1, 2, 2, 1, 1, 2, 1, 1, 2, NA, 1, 2, 2, 1, 1, 2, 1, 2, 1, 1, 1, 1, 1, 1, 1, 2, 1, 2, 1, 1, 1, 1, 1, 2, 1, NA, NA, 2, 1, 2, 1, 2, 1, 1, 1, 2, 1, 2, 2, 1, 1, 2, 2, 1, 2, 1, 2, 1, 2, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, NA, 2, 1, 1, 1, 1, NA, 1, 1, 2, 1, 1, 1, 2, 1, NA, 1, 1, 1, 1, 1, 1, 1, NA, NA, 2, 1, 1, 2, 2, NA, 2, NA, 2, 2, 1, NA, 1, NA, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, NA, 1, 1, NA, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 2, 1, 1, 2, 2), Leeftijd = c(26, 69, 67, 38, 29, 50, 29, 23, 52, 39, 50, 29, 36, 52, 43, 53, 47, 33, 52, 55, 43, 64, 35, 24, 51, 39, 50, 51, 46, 51, 30, 32, 28, 25, 52, 48, 60, 31, 61, 47, 46, 56, 38, 72, 88, 34, 56, 27, 27, 56, 52, 49, 34, 25, 22, 60, 61, 42, 45, 51, 42, 61, 69, 57, 35, 50, 42, 50, 51, 46, 28, 34, 52, 33, 30, 64, 65, 35, 31, 57, 75, 43, 46, 35, 65, 29, 29, 75, 49, 31, 57, 29, 40, 75, 30, 34, 58, 47, 37, 43, 34, 47, 46, 42, 49, 57, 46, 36, 51, 80, 45, 47, 48, 23, 51, 53, 44, 64, 44, 33, 40, 42, 29, 60, 28, 47, 47, 39, 25, 41, 39, 27, 57, 66, 42, 22, 59, 27, 43, 53, 65, 52, 41, 50, 55, 29, 55, 39, 41, 25, 74, 68, 55, 29, 77, 45, 18, 34, 49, 74, 44, 33, 48, 82, 61, 54, 46, 30, 33, 65, 51, 44, 50, 57, 27, 56, 85, 52, 31, 62, 62, 34, 48, 28, 28, 63, 30, 40, 44, 37, 73, 70, 39, 59, 56, 61, 40, 43, 33, 58, 44, 62, 26, 72, 67, 59, 48, 37, 52, 37, 57, 53, 59, 44, 71, 81, 33, 61, 50, 33, 48, 50, 63, 46, 60, 58, 40, 63, 39, 71, 38, 40, 56, 36, 52, 61, 83, 59, 43, 69, 50, 57, 38, 50, 27, 43, 46, 30, 50, 34, 68, 53, 48, 84, 41, 57, 61, 72, 27, 80, 71, 69, 61, 43, 67, 60, 58, 67, 72, 40, 79, 52, 80, 33, 25, 80, 67, 56, 66, 54, 50, 65, 39, 36, 69, 39, 34, 41, 36, 61, 33, 42, 43, 45, 48, 67, 69, 66, 37, 28, 64, 65, 68, 62, 84, 82, 59, 61, 74, 52, 41, 30, 33, 55, 55, 26, 53, 33, 64, 65, 74, 67, 70, 58, 51, 62, 67, 52, 40, 57, 57, 57, 59, 56, 61, 58, 45, 63, 61, 50, 70, 32, 50, 74, 70, 49, 42, 71, 51, 67, 46, 45, 75, 54, 75, 45, 46, 64, 60, 55, 61, 65, 68, 71, 43, 78, 53, 63, 85, 75, 66, 67, 54, 63, 68, 84, 58, 72, 70, 58, 29, 63, 83, 64, 75, 59, 76, 61, 62, 65, 61, 72, 20, 43, 67, 33, 62, 63, 51, 34, 68, 68, 60, 67, 44, 64, 69, 53, 69, 47, 41, 38, 57, 71, 70, 68, 25, 60, 71, 48, 64, 62, 72, 60, 45, 67, 59, 73, 27, 64, 66, 57, 72, 71, 77, 58, 56, 65, 74, 44, 22, 63, 42, 80, 52, 66, 60, 56, 54, 42, 68, 57, 37)), .Names = c("HbA1c", "BMI", "Comorb", "PAIDhoog", "PHQhoog", "totaalhealthcareutilization", "etndich", "Leeftijd"), row.names = c(NA, -453L), variable.labels = structure(c("HbA1c", "BMI level", "", "", "", "", "", ""), .Names = c("HbA1c", "BMI", "Comorb", "PAIDhoog", "PHQhoog", "totaalhealthcareutilization", "etndich", "Leeftijd")), codepage = 65001L, class = "data.frame")

Thanks again Maurits! I have followed these steps in R, however the output remains the same as the output you are showing above. Were you able to reproduce the SPSS output? The output above is similar to my output in R before. — Charlotte, Jul 19 '18 at 14:19
@Charlotte What do you mean? The parameter estimates *are* the same now between R and SPSS. Compare the `B` values from the SPSS output with the `Estimate` values from the R summary. — Maurits Evers, Jul 19 '18 at 14:24
The parameter estimates are indeed the same, but the S.E.'s and p-values aren't. Do you know what causes this? — Charlotte, Jul 19 '18 at 14:56
I am now checking all the previous output I have posted and this seems to be the 'problem' every time. The parametes estimates are similar between R and SPSS, but the standard error's and p-values aren't.. — Charlotte, Jul 19 '18 at 14:58
@Charlotte *"The output above is similar to my output in R before."* Hang on, hang on. Parameter estimates being the same means that the ML estimation is the same in both methods. **Estimates are not similar, they are identical**. This answers your original question. As to why standard errors/p-values are slightly different depends on the underlying hypothesis test. I don't know SPSS, so I'm not sure what they do. Note however that stat. significance for every parameter is the same in both methods. Results **are reproducible** between SPSS and R, and lead to **identical parameter estimates**. — Maurits Evers, Jul 19 '18 at 21:38
@Charlotte To reiterate: There exists **only one set of ML parameter estimates** for a given model (and these estimates agree from SPSS and R provided you match reference levels of categorical variables). But there may be **different parameter se's and p-values**, which depends on the hypothesis test. Consider for example a bootstrap-based test: In that case se's and p-values would change slightly *every time* you re-run the analysis. The key here is that the statistical significance of parameters should not change. And it doesn't in our case. — Maurits Evers, Jul 19 '18 at 21:45
@Charlotte: And lastly in response to what you said in your (earlier, now deleted) "answer" post: *"The differences appear to be located in the way the parameters are estimated."* No there is no difference in the way parameters are estimated. Estimates from both `glm` and SPSS are ML estimates and they are identical. — Maurits Evers, Jul 19 '18 at 23:45
You should add the content of your last comments to the answer. — Roland, Jul 20 '18 at 06:03
@ Maurits: Thank you very much for your extensive help. However, I still don't understand why the p-values and S.E.'s are similar between SPSS and R if I change the parameter estimation method in SPSS to 'Hybrid' and the scale parameter method to 'Pearson Chi-square'. @ Everyone: I am so sorry that I am not using this platform in the right manner, it's my first time. I will edit my original post from now on with new information! — Charlotte, Jul 20 '18 at 07:35
@Charlotte I've made an edit to elaborate. This is getting quite technical, but it boils down to SPSS using some *scaled* variance-covariance matrix for its standard error estimation and in turn confidence interval/p-value calculation. R's se and CI calculation is consistent with standard statistical definitions, and I suggest sticking to R (I'm not a fan of black-box stats programs), as it is it easy to verify what's going on; for example I'm demonstrating how you can reproduce se's from the estimated variance-covariance matrix. — Maurits Evers, Jul 20 '18 at 15:35
*"However, I still don't understand why the p-values and S.E.'s are similar between SPSS and R if I change the parameter estimation method in SPSS to 'Hybrid' and the scale parameter method to 'Pearson Chi-square'"* **Again: parameter estimates are identical**. Same ML method. Same parameter estimates. We seem to be going in circles here. — Maurits Evers, Jul 20 '18 at 15:37
@Roland Thanks, I've made an edit. I'm not familiar with SPSS, but it seems that SPSS uses a scaled var-cov matrix to estimate se's and confidence intervals. Regrettably the SPSS documentation is somewhat sparse... — Maurits Evers, Jul 20 '18 at 15:41
Thanks Maurits! It is thus not possible to replicate the results from SPSS (SE's, p-values en confidence intervals) in R? — Charlotte, Jul 24 '18 at 13:57

Different results Gamma generalized linear model R and SPSS

1 Answers1

Further comments on differences in the SPSS and `glm` output

Sample data

Linked

Different results Gamma generalized linear model R and SPSS

1 Answers1

Further comments on differences in the SPSS and glm output

Sample data

Linked

Further comments on differences in the SPSS and `glm` output