Moderation in ordered regression

Jacob Joris

Join Date: Jul 2018

Posts: 12
#1

Moderation in ordered regression

10 Oct 2019, 03:34

Hello!

For my thesis, I have to do some regressions in STATA. I am not very good with STATA at the moment, so I would like to check if some things are correct with you. Help is greatly appreciated.

My regression is testing is 'green attitude' is positively related to 'green buying' (if you care about the environmental impact of the products you buy, you buy more green products). Secondly, I am testing whether the impact of attitude on green buying will be moderated by the trust in ecolabels. (If you trust that ecolabels are reliable, it's easier to differentiate products in the supermarket and convert your attitude into actual behaviour).

Control variables: Age, gender, country

I have two questions.

1. As my dependent variable (green buying) is on a likert scale from 'I often buy green products' to 'I never buy green products', I assume that I have to use an ordinal logistic regression?
Would this be the correct regression?
. ologit greenbuying attitude age i.gender i.country

2. To test for moderation, I am not sure what I have to do. For 'Trust in ecolabels' I have 4 answers: (Fully confident, Fairly confident, Not very confident, Not at all confident). Would that mean I have to make 3 interaction terms: Attitude*Fully confident, Attitude*Fairly confident and Attitude*Not very confident? How would this be tested for moderation?

I really hope I explained this as clear as possible. If it's not clear please let me know.
Tags: None
Jacob Joris

Join Date: Jul 2018

Posts: 12
#2

10 Oct 2019, 06:57

If I run . ologit greenbuying i.attitude##i.trust age i.gender i.country, my results become hard to interpret. I don't think that is correct.

. ologit greenbuying i.attitude##i.trust age i.gender i.country

Iteration 0: log likelihood = -29908.635
Iteration 1: log likelihood = -27708.184
Iteration 2: log likelihood = -27593.268
Iteration 3: log likelihood = -27586.562
Iteration 4: log likelihood = -27586.561

Ordered logistic regression Number of obs = 25,296
LR chi2(53) = 4644.15
Prob > chi2 = 0.0000
Log likelihood = -27586.561 Pseudo R2 = 0.0776

------------------------------------------------------------------------------------------------------------
greenbuying | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------------------------------------+----------------------------------------------------------------
attitude |
Fairly important | .8201081 .0728374 11.26 0.000 .6773493 .9628669
Not very important | 1.875801 .1250258 15.00 0.000 1.630755 2.120847
Not at all important | 3.323528 .2112647 15.73 0.000 2.909457 3.737599
DK/NA | 1.419003 .3087574 4.60 0.000 .8138496 2.024156
|
trust |
Fairly confident | .1142917 .054232 2.11 0.035 .0079989 .2205846
Not very confident | .6896144 .0633293 10.89 0.000 .5654912 .8137376
Not at all confident | 1.20398 .1102265 10.92 0.000 .9879401 1.42002
DK/NA | .9931597 .1831172 5.42 0.000 .6342565 1.352063
|
attitude#trust |
Fairly important#Fairly confident | -.0209008 .0815317 -0.26 0.798 -.1807001 .1388985
Fairly important#Not very confident | -.1864211 .0912053 -2.04 0.041 -.3651802 -.0076621
Fairly important#Not at all confident | -.272597 .1548921 -1.76 0.078 -.5761799 .030986
Fairly important#DK/NA | -.4660487 .2755669 -1.69 0.091 -1.00615 .0740525
Not very important#Fairly confident | -.0688794 .1370753 -0.50 0.615 -.337542 .1997831
Not very important#Not very confident | -.2797692 .1430137 -1.96 0.050 -.5600709 .0005324
Not very important#Not at all confident | -.1287484 .1979626 -0.65 0.515 -.5167481 .2592513
Not very important#DK/NA | .2314287 .3662841 0.63 0.527 -.486475 .9493325
Not at all important#Fairly confident | -.5627099 .2523146 -2.23 0.026 -1.057237 -.0681823
Not at all important#Not very confident | -.8234295 .2519569 -3.27 0.001 -1.317256 -.3296031
Not at all important#Not at all confident | -.3256793 .2883814 -1.13 0.259 -.8908965 .239538
Not at all important#DK/NA | -.8262591 .5158616 -1.60 0.109 -1.837329 .1848109
DK/NA#Fairly confident | -.2344112 .3548998 -0.66 0.509 -.930002 .4611796
DK/NA#Not very confident | .2175433 .3740203 0.58 0.561 -.515523 .9506096
DK/NA#Not at all confident | .0775005 .4761514 0.16 0.871 -.855739 1.01074
DK/NA#DK/NA | .2721843 .4673382 0.58 0.560 -.6437818 1.18815
|
age | -.0065858 .0007687 -8.57 0.000 -.0080924 -.0050792
|
gender |
Female | -.0702839 .0256338 -2.74 0.006 -.1205252 -.0200426
|
country |
BE - Belgium | .2801669 .0899557 3.11 0.002 .1038569 .456477
NL - The Netherlands | .004789 .0897695 0.05 0.957 -.171156 .180734
DE - Germany | -.7021774 .090636 -7.75 0.000 -.8798206 -.5245341
IT - Italy | .990328 .0906194 10.93 0.000 .8127173 1.167939
LU - Luxembourg | -.1178708 .1092301 -1.08 0.281 -.3319579 .0962163
DK - Denmark | -.1765398 .0891486 -1.98 0.048 -.3512678 -.0018117
IE - Ireland | .2581385 .0893675 2.89 0.004 .0829814 .4332956
GB - United Kingdom | .364386 .0900674 4.05 0.000 .1878572 .5409149
GR - Greece | -.1875929 .0911587 -2.06 0.040 -.3662607 -.008925
ES - Spain | -.0963934 .0904821 -1.07 0.287 -.2737351 .0809483
PT - Portugal | .3358904 .0910432 3.69 0.000 .1574489 .5143318
FI - Finland | .200702 .0893958 2.25 0.025 .0254894 .3759146
SE - Sweden | -.1520133 .0887996 -1.71 0.087 -.3260573 .0220308
AT - Austria | -.9390364 .0911822 -10.30 0.000 -1.11775 -.7603225
CY - Cyprus (Republic) | .2144357 .1121542 1.91 0.056 -.0053825 .4342538
CZ - Czech Republic | .4995264 .0897431 5.57 0.000 .3236332 .6754195
EE - Estonia | .0337007 .0923465 0.36 0.715 -.1472951 .2146965
HU - Hungary | -.0934508 .0908474 -1.03 0.304 -.2715085 .0846069
LV - Latvia | .076776 .0917092 0.84 0.402 -.1029707 .2565228
LT - Lithuania | .4364923 .0923275 4.73 0.000 .2555337 .6174509
MT - Malta | .4093672 .1172133 3.49 0.000 .1796333 .6391011
PL - Poland | -.0684919 .0914739 -0.75 0.454 -.2477775 .1107937
SK - Slovakia | -.0050784 .0900455 -0.06 0.955 -.1815644 .1714076
SI - Slovenia | -.1914782 .0898398 -2.13 0.033 -.3675609 -.0153955
BG - Bulgaria | 1.234301 .0922793 13.38 0.000 1.053437 1.415165
RO - Romania | .3980116 .0930486 4.28 0.000 .2156397 .5803836
HR - Croatia | .23088 .0904485 2.55 0.011 .0536042 .4081557
-------------------------------------------+----------------------------------------------------------------
/cut1 | -.428959 .0899918 -.6053397 -.2525783
/cut2 | 2.591908 .0917389 2.412103 2.771712
/cut3 | 3.0219 .0923916 2.840816 3.202984
/cut4 | 3.825578 .0944799 3.640401 4.010755
/cut5 | 4.196897 .0960455 4.008651 4.385142
------------------------------------------------------------------------------------------------------------
Comment
Richard Williams

Join Date: Apr 2014

Posts: 4983
#3

10 Oct 2019, 07:30

Hi Jacob. Your output is very hard to read. You should use code tags. See pt 12 of the Statalist FAQ on answering questions effectively.

I am not sure what exactly it is you don't like about your results. What exactly is it you find hard to interpret? But I find marginal effects and adjusted predictions often aid interpretation. See

https://www3.nd.edu/~rwilliam/xsoc73994/Margins05.pdf

Note that this is is the 5th handout on margins. If you need more background first, the related handouts are at

https://www3.nd.edu/~rwilliam/xsoc73994/index.html

If that isn't really what you want, I suggest reposting your code and output using code tags and being more specific about what you find is hard to interpret.

-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
StataNow Version: 19.5 MP (2 processor)
EMAIL: [email protected]
WWW: https://www3.nd.edu/~rwilliam
Comment

Jacob Joris

Join Date: Jul 2018
Posts: 12

11 Oct 2019, 07:07

Code:

. ologit greenbuying i.attitude##i.trust age i.gender i.country 

Iteration 0:   log likelihood = -29908.635  
Iteration 1:   log likelihood = -27708.184  
Iteration 2:   log likelihood = -27593.268  
Iteration 3:   log likelihood = -27586.562  
Iteration 4:   log likelihood = -27586.561  

Ordered logistic regression                     Number of obs     =     25,296
                                                LR chi2(53)       =    4644.15
                                                Prob > chi2       =     0.0000
Log likelihood = -27586.561                     Pseudo R2         =     0.0776

------------------------------------------------------------------------------------------------------------
                               greenbuying |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------------------------------------+----------------------------------------------------------------
                                  attitude |
                         Fairly important  |   .8201081   .0728374    11.26   0.000     .6773493    .9628669
                       Not very important  |   1.875801   .1250258    15.00   0.000     1.630755    2.120847
                     Not at all important  |   3.323528   .2112647    15.73   0.000     2.909457    3.737599
                                    DK/NA  |   1.419003   .3087574     4.60   0.000     .8138496    2.024156
                                           |
                                     trust |
                         Fairly confident  |   .1142917    .054232     2.11   0.035     .0079989    .2205846
                       Not very confident  |   .6896144   .0633293    10.89   0.000     .5654912    .8137376
                     Not at all confident  |    1.20398   .1102265    10.92   0.000     .9879401     1.42002
                                    DK/NA  |   .9931597   .1831172     5.42   0.000     .6342565    1.352063
                                           |
                            attitude#trust |
        Fairly important#Fairly confident  |  -.0209008   .0815317    -0.26   0.798    -.1807001    .1388985
      Fairly important#Not very confident  |  -.1864211   .0912053    -2.04   0.041    -.3651802   -.0076621
    Fairly important#Not at all confident  |   -.272597   .1548921    -1.76   0.078    -.5761799     .030986
                   Fairly important#DK/NA  |  -.4660487   .2755669    -1.69   0.091     -1.00615    .0740525
      Not very important#Fairly confident  |  -.0688794   .1370753    -0.50   0.615     -.337542    .1997831
    Not very important#Not very confident  |  -.2797692   .1430137    -1.96   0.050    -.5600709    .0005324
  Not very important#Not at all confident  |  -.1287484   .1979626    -0.65   0.515    -.5167481    .2592513
                 Not very important#DK/NA  |   .2314287   .3662841     0.63   0.527     -.486475    .9493325
    Not at all important#Fairly confident  |  -.5627099   .2523146    -2.23   0.026    -1.057237   -.0681823
  Not at all important#Not very confident  |  -.8234295   .2519569    -3.27   0.001    -1.317256   -.3296031
Not at all important#Not at all confident  |  -.3256793   .2883814    -1.13   0.259    -.8908965     .239538
               Not at all important#DK/NA  |  -.8262591   .5158616    -1.60   0.109    -1.837329    .1848109
                   DK/NA#Fairly confident  |  -.2344112   .3548998    -0.66   0.509     -.930002    .4611796
                 DK/NA#Not very confident  |   .2175433   .3740203     0.58   0.561     -.515523    .9506096
               DK/NA#Not at all confident  |   .0775005   .4761514     0.16   0.871     -.855739     1.01074
                              DK/NA#DK/NA  |   .2721843   .4673382     0.58   0.560    -.6437818     1.18815
                                           |
                                       age |  -.0065858   .0007687    -8.57   0.000    -.0080924   -.0050792
                                           |
                                    gender |
                                   Female  |  -.0702839   .0256338    -2.74   0.006    -.1205252   -.0200426
                                           |
                                   country |
                             BE - Belgium  |   .2801669   .0899557     3.11   0.002     .1038569     .456477
                     NL - The Netherlands  |    .004789   .0897695     0.05   0.957     -.171156     .180734
                             DE - Germany  |  -.7021774    .090636    -7.75   0.000    -.8798206   -.5245341
                               IT - Italy  |    .990328   .0906194    10.93   0.000     .8127173    1.167939
                          LU - Luxembourg  |  -.1178708   .1092301    -1.08   0.281    -.3319579    .0962163
                             DK - Denmark  |  -.1765398   .0891486    -1.98   0.048    -.3512678   -.0018117
                             IE - Ireland  |   .2581385   .0893675     2.89   0.004     .0829814    .4332956
                      GB - United Kingdom  |    .364386   .0900674     4.05   0.000     .1878572    .5409149
                              GR - Greece  |  -.1875929   .0911587    -2.06   0.040    -.3662607    -.008925
                               ES - Spain  |  -.0963934   .0904821    -1.07   0.287    -.2737351    .0809483
                            PT - Portugal  |   .3358904   .0910432     3.69   0.000     .1574489    .5143318
                             FI - Finland  |    .200702   .0893958     2.25   0.025     .0254894    .3759146
                              SE - Sweden  |  -.1520133   .0887996    -1.71   0.087    -.3260573    .0220308
                             AT - Austria  |  -.9390364   .0911822   -10.30   0.000     -1.11775   -.7603225
                   CY - Cyprus (Republic)  |   .2144357   .1121542     1.91   0.056    -.0053825    .4342538
                      CZ - Czech Republic  |   .4995264   .0897431     5.57   0.000     .3236332    .6754195
                             EE - Estonia  |   .0337007   .0923465     0.36   0.715    -.1472951    .2146965
                             HU - Hungary  |  -.0934508   .0908474    -1.03   0.304    -.2715085    .0846069
                              LV - Latvia  |    .076776   .0917092     0.84   0.402    -.1029707    .2565228
                           LT - Lithuania  |   .4364923   .0923275     4.73   0.000     .2555337    .6174509
                               MT - Malta  |   .4093672   .1172133     3.49   0.000     .1796333    .6391011
                              PL - Poland  |  -.0684919   .0914739    -0.75   0.454    -.2477775    .1107937
                            SK - Slovakia  |  -.0050784   .0900455    -0.06   0.955    -.1815644    .1714076
                            SI - Slovenia  |  -.1914782   .0898398    -2.13   0.033    -.3675609   -.0153955
                            BG - Bulgaria  |   1.234301   .0922793    13.38   0.000     1.053437    1.415165
                             RO - Romania  |   .3980116   .0930486     4.28   0.000     .2156397    .5803836
                             HR - Croatia  |     .23088   .0904485     2.55   0.011     .0536042    .4081557
-------------------------------------------+----------------------------------------------------------------
                                     /cut1 |   -.428959   .0899918                     -.6053397   -.2525783
                                     /cut2 |   2.591908   .0917389                      2.412103    2.771712
                                     /cut3 |     3.0219   .0923916                      2.840816    3.202984
                                     /cut4 |   3.825578   .0944799                      3.640401    4.010755
                                     /cut5 |   4.196897   .0960455                      4.008651    4.385142
------------------------------------------------------------------------------------------------------------

Hi Richard, thank you. I didn't realise how to post code correctly, it should work now. The margins document is really helpful.

First off, is ologit the correct regression type here? And if so, is using i.attitude##i.trust the correct way to test whether the impact of attitude on green buying will be moderated by the trust in ecolabels? How do I decide whether it is or not from my results? I looked online and read the articles you sent, but I still do not see how I can decide this.

I am sorry if this seems basic to you,

Comment

Jacob Joris

Join Date: Jul 2018

Posts: 12
#5

29 Oct 2019, 07:59

Hello all,

I am still struggling with how to interpret my results. Can someone help me?
Only: Not at all important#Not very confident is significant, what does this mean regarding moderation? Thank you in advance
Comment

Announcement