0
$\begingroup$

I am experiencing a problem with ez that is similar to the one described in this thread: Error in `ezANOVA` with balanced dataset with no missing data

Specifically, I am getting the error message

Error in ezANOVA_main(data = data, dv = dv, wid = wid, within = within,  :  One or more cells is missing data. Try using ezDesign() to check your data.

while I am fairly confident my dataset actually does not contain any missing cells.

My dataframe looks something like this:

      id                     szenario measurement  default   PI  congruency    ego avoidability
694   170     Con_Inac_1_SAP_Orphanage           0 inaction high   congruent    ego    avoidable
10396 170   Con_Inac_4_SII_Cliffhanger           0 inaction  low   congruent    ego   inevitable
15247 170       Incon_Inac_6_OAI_Nobel           0 inaction  low incongruent no_ego    avoidable
19405 170  Incon_Inac_7_OII_Euthanasia           1 inaction  low incongruent no_ego   inevitable
26335 170         Incon_Ac_2_SAI_Ebola           1   action  low incongruent    ego    avoidable
27721 170        Incon_Ac_3_SIP_Rescue           1   action high incongruent    ego   inevitable
33958 170      Con_Ac_5_OAP_Transplant           0   action high   congruent no_ego    avoidable
40888 170             Con_Ac_8_OIP_Spy           0   action high   congruent no_ego   inevitable
4160  171       Incon_Inac_2_SAI_Ebola           1 inaction  low incongruent    ego    avoidable
8318  171      Incon_Inac_3_SII_Rescue           1 inaction  low incongruent    ego   inevitable
14555 171    Con_Inac_5_OAI_Transplant           0 inaction  low   congruent no_ego    avoidable
18713 171           Con_Inac_8_OIP_Spy           0 inaction high   congruent no_ego   inevitable
22871 171       Con_Ac_1_SAP_Orphanage           0   action high   congruent    ego    avoidable
29801 171     Con_Ac_4_SIP_Cliffhanger           1   action high   congruent    ego   inevitable
34652 171         Incon_Ac_6_OAP_Nobel           0   action high incongruent no_ego    avoidable
41582 171    Incon_Ac_7_OII_Euthanasia           1   action  low incongruent no_ego   inevitable
1389  175       Incon_Inac_2_SAP_Ebola           0 inaction high incongruent    ego    avoidable
5547  175      Incon_Inac_3_SIP_Rescue           0 inaction high incongruent    ego   inevitable
11784 175    Con_Inac_5_OAP_Transplant           0 inaction high   congruent no_ego    avoidable
21486 175           Con_Inac_8_OII_Spy           0 inaction  low   congruent no_ego   inevitable
25644 175       Con_Ac_1_SAI_Orphanage           0   action  low   congruent    ego    avoidable
32574 175     Con_Ac_4_SII_Cliffhanger           0   action  low   congruent    ego   inevitable
37425 175         Incon_Ac_6_OAI_Nobel           0   action  low incongruent no_ego    avoidable
38811 175    Incon_Ac_7_OIP_Euthanasia           0   action high incongruent no_ego   inevitable
4162  177       Incon_Inac_2_SAI_Ebola           1 inaction  low incongruent    ego    avoidable
8320  177      Incon_Inac_3_SII_Rescue           0 inaction  low incongruent    ego   inevitable
14557 177    Con_Inac_5_OAI_Transplant           0 inaction  low   congruent no_ego    avoidable
18715 177           Con_Inac_8_OIP_Spy           1 inaction high   congruent no_ego   inevitable
22873 177       Con_Ac_1_SAP_Orphanage           0   action high   congruent    ego    avoidable
29803 177     Con_Ac_4_SIP_Cliffhanger           0   action high   congruent    ego   inevitable
34654 177         Incon_Ac_6_OAP_Nobel           1   action high incongruent no_ego    avoidable
41584 177    Incon_Ac_7_OII_Euthanasia           1   action  low incongruent no_ego   inevitable`

When running ANOVAs, it works out for most of them, meaning none of the one factor ANOVAs produce an error, and almost none of the two-factor ANOVAs do. However, for the following two-factor ANOVA I get the error message described above:

default_PI_model <- ezANOVA(data = df6, dv = .(measurement), wid = .(id), within = .(default, PI), within_full = .(default,congruency,ego,PI,avoidability), detailed = TRUE, type = 3)

--->

Error in ezANOVA_main(data = data, dv = dv, wid = wid, within = within,  :  One or more cells is missing data. Try using ezDesign() to check your data.

I see that the design for the default x PI interaction is less balanced than for the other factor combinations, but the two-level interaction should still be computable, I think.

Does anybody have an idea on this? Thanks in advance.

$\endgroup$
3
  • 1
    $\begingroup$ How does this question differ from the one you cite at the outset? It looks identical to me. If it's not, could you please demonstrate that it is different? $\endgroup$
    – whuber
    Feb 28, 2019 at 15:00
  • $\begingroup$ The question is indeed the same, but when I posted in the other thread I was asked by a moderator to submit it as a new question, so that is what I did. So yes, I may have the same problem, but so far no solution for it. Maybe I'm also not getting the sense behind using 'temp = as.data.frame(table(df6id))′ as proposed there. But for me this just returns 1 for all participants and thus doesn't seem to diagnose anything. $\endgroup$
    – user239239
    Mar 1, 2019 at 9:17
  • $\begingroup$ I apologize for the misunderstanding. The moderator's intention was for you to post anything that might be different, new, or in addition to the existing question. This protocol is perhaps the most basic element of every SE site: we aim not to have duplicate (and therefore variable and confusing) threads. $\endgroup$
    – whuber
    Mar 1, 2019 at 13:50

0

Browse other questions tagged or ask your own question.