To determine the difference in diversity between participants who have received different types of training about the entity, we compared the number of diagnostic attributes (Diagnostic_Attr) and non-diagnostic attributes of the primary entity (Non- diagnostic_Attr) between the treatment conditions. These comparisons address H1, i.e., there will be no significant difference in the number of diagnostic attributes reported between trained and untrained contributors; and H2, i.e., untrained contributors will report fewer non-diagnostic attributes of a target entity than implicitly trained contributors but more non-diagnostic attributes than explicitly trained contributors. We also compared other non-diagnostic attributes, such as attributes describing the state of the primary entity (Behavior_Attr) and attributes describing the primary entity’s interaction with other entities or its environment (Mutual_Attr), for each image presented to the participants. The results are presented in Tables 3.2 and 3.3.
Table 3.2: ANOVA Results for Primary Entity Attributes*
Variable Group Mean
Untrained Explicit Implicit 𝑭 p-value 𝜼𝟐
Diagnostic_Attr 1.272 1.514 1.440 0.92 0.399 0.001 Non-diagnostic_Attr 𝟐. 𝟎𝟎𝟗 𝟎. 𝟓𝟖𝟎 𝟐. 𝟖𝟎𝟏 𝟔𝟎. 𝟒𝟎𝟓 𝟎. 𝟎𝟎𝟎 𝟎. 𝟎𝟖𝟑 Behavior_Attr 𝟎. 𝟒𝟗𝟏 𝟎. 𝟎𝟒𝟓 𝟎. 𝟎𝟐𝟐 𝟑𝟔. 𝟖𝟐𝟎 𝟎. 𝟎𝟎𝟎 𝟎. 𝟎𝟓𝟐 Mutual_Attr 1.763 0.681 0.725 𝟑𝟐. 𝟖𝟒𝟑 𝟎. 𝟎𝟎𝟎 𝟎. 𝟎𝟒𝟕
*Significant differences are bolded (p=0.05)
Table 3.3: Post-Hoc Test Results for Primary Entity Attributes*
A B mean(A) mean(B) Mean Diff. Std. Err T p-value 𝜼𝟐
Diagnostic_Attr Explicit Implicit 1.514 1.440 0.074 0.183 0.407 0.900 0.000
Explicit Untrained 1.514 1.272 0.242 0.183 1.324 0.383 0.002 Implicit Untrained 1.440 1.272 0.167 0.183 0.917 0.718 0.001 Non- diagnostic_Attr Explicit Implicit 0.580 2.801 -2.221 0.205 -10.844 0.001 0.116 Explicit Untrained 0.580 2.009 -1.429 0.205 -6.975 0.001 0.052 Implicit Untrained 2.801 2.009 0.792 0.205 3.869 0.001 0.016
Behavior_Attr Explicit Implicit 0.045 0.022 0.022 0.062 0.362 0.900 0.000
Explicit Untrained 0.045 0.491 -0.446 0.062 -7.243 0.001 0.055
Implicit Untrained 0.022 0.491 -0.469 0.062 -7.605 0.001 0.061
Mutual_Attr Explicit Implicit 0.681 0.725 -0.045 0.151 -0.295 0.900 0.000
Explicit Untrained 0.681 1.763 -1.083 0.151 -7.162 0.001 0.054
Implicit Untrained 0.725 1.763 -1.038 0.151 -6.867 0.001 0.050
From Table 3.2, the number of diagnostic attributes reported, Diagnostic_Attr, is not significantly different across the groups with F(2,1341) = 0.92, p = 0.399 at a 5% level of significance. We can, therefore, conclude that the number of diagnostic attributes reported is equal across the groups, supporting H1a. However, Non-diagnostic_Attr is significantly different across the three groups with F(2,1341) = 60.405, 𝑝 < 0.0000 at a 5% level of significance. From the post-hoc tests, we observe that all group means are significantly different from each other, with the average Non-diagnostic_Attr for the Implicitly Trained Group being the maximum and that for the Explicitly Trained Group being the minimum, supporting H1b.
The number of attributes reported that describe the primary entity’s behaviour Behavior_Attr is also significantly different across the groups. The post-hoc test results suggest that the group means for the Explicitly Trained Group, and the Implicitly Trained Group are significantly lower than the average for the Untrained Group. However, the Explicitly Trained Group and the Implicitly Trained Group are not significantly different. The number of mutual attributes is also significantly different across the groups. From the post-hoc tests, we again observe that the group means of the Explicitly Trained Group and the Implicitly Trained Group are significantly lower than that of the Untrained Group. However, there is no significant difference between the Explicitly Trained Group and the Implicitly Trained Group.
For the secondary entities, we analyze the variables Secondary_Ent, Secondary_Ent_Attr, Secondary_Ent_Behavior, and Secondary_Ent_Mutual. Further, since we define information diversity as the number of unique attributes reported about an entity, information diversity is the sum of all the attributes reported for each image, given as:
Information Diversity = Diagnostic_Attr + Non-diagnostic_Attr + Behavior_Attr + Mutual_Attr + Secondary_Ent_Attr + Secondary_Ent_Mutual + Secondary_Ent_Behavior + Diagnostic_Attr_Variance + Non-Diagnostic_Attr_Variance
We present the results of our analyses in Tables 3.4 and 3.5.
Table 3.4: ANOVA Results for Information Diversity*
Variable Group Mean
Untrained Explicit Implicit 𝐹 p-
value 𝜂 2 Secondary_Ent_Attr 1.730 0.246 1.261 43.953 0.000 0.062 Secondary_Ent_Behavior 0.491 0.056 0.257 13.193 0.000 0.019 Secondary_Ent_Mutual 1.105 0.826 0.547 8.640 0.000 0.013 Information Diversity 8.984 4.070 7.433 85.967 0.000 0.114
Table 3.5: Post-hoc Results for Information Diversity*
*Significant differences are bolded (p=0.05)
Table 3.4 shows that Secondary_Ent_Attr is significantly different across the groups. The post hoc test results in Table 3.5 suggest that all group means are significantly different from one another, with the Untrained Group being the maximum and the Explicitly Trained Group being the minimum. Secondary_Ent_Behavior is also significantly different across the three groups. The post hoc tests show that the average Secondary_Ent_Behavior for Untrained Group is the maximum and is significantly higher than that of the Explicitly Trained Group, which is the minimum. However, there is no statistically significant difference between the number of Secondary_Ent_Behavior reported by the Untrained
A B mean(A) mean(B) Mean Diff. Std. Err T p-value 𝜼𝟐
Secondary_Ent_Attr
Explicit Implicit 0.246 1.261 -1.016 0.162 -6.274 0.001 0.042
Explicit Untrained 0.246 1.730 -1.484 0.162 -9.170 0.001 0.086
Implicit Untrained 1.261 1.730 -0.469 0.162 -2.896 0.011 0.009
Secondary_Ent_Behavior Explicit Implicit 0.056 0.257 -0.201 0.085 -2.368 0.047 0.006
Explicit Untrained 0.056 0.491 -0.435 0.085 -5.131 0.001 0.029
Implicit Untrained 0.257 0.491 -0.234 0.085 -2.763 0.016 0.008
Secondary_Ent_Mutual Explicit Implicit 0.826 0.547 0.279 0.134 2.078 0.094 0.005
Explicit Untrained 0.826 1.105 -0.279 0.134 -2.078 0.094 0.005
Implicit Untrained 0.547 1.105 -0.558 0.134 -4.157 0.001 0.019
Information Diversity Explicit Implicit 4.070 7.433 -3.363 0.383 -8.777 0.001 0.079
Explicit Untrained 4.070 8.984 -4.914 0.383 -12.825 0.001 0.155
Group and the Implicitly Trained Group. Secondary_Ent_Mutual is also significantly different across the groups. The post hoc tests show that the Untrained Group’s mean for Secondary_Ent_Mutual is highest and significantly higher than that for the Explicitly Trained Group and the Implicitly Trained Group. However, there is no significant difference between the Explicitly Trained Group and the Implicitly Trained Group for Secondary_Ent_Mutual.
Furthermore, as shown in Table 3.4 information diversity is significantly different across the groups with 𝐹(2,81) = 85.967, 𝑝 = 0.000. Post hoc test results in Table 3.5 shows that the group mean for the Untrained Group is significantly higher than the Explicitly Trained Group and the Implicitly Trained Group, while the group mean for the Implicitly Trained Group is significantly greater than the mean for the Explicitly Trained Group.