We very much appreciate the scientific discussion caused by the article. In this statement, we want to respond to constructive comments (only some of the commentary was scientific and thus appropriate for further discussion). In the following, we emphasize what can and what cannot be deduced from the study (see also the Discussion section of the article and the comment by the Academic Editor http://goo.gl/WVcNuW).

1) We cannot draw conclusions about conservatives in general. The sample was not representative of the US. As is the rule in psychological studies, relations between constructs were tested.

2) Our study does not rule out the possibility that liberalism might also be related to judging BS statements as profound when BS statements are flavored with liberal content.

3) Based on previous research, we tested the prediction that political conservatism and favorable views of Republican candidates for US president are positively related to judging BS statements as profound. Assuming a linear relation on a theoretical level, this was tested using a linear model. Results supported the central prediction. As a follow-up, we applied additional analyses to test the prediction. Robust regression analyses corroborated the findings reported in the article and contradict other suggestions (see http://goo.gl/w8ujjZ). Additionally, the mean correlation of Republican candidates with the BSR scale was significantly higher than the mean correlation of Democratic candidates with the BSR scale, especially when controlling for mundane statements (given that the BS scale and mundane statements are positively correlated).

4) Looking at the raw data and using locally-weighted regression lines (see http://goo.gl/w8ujjZ) might lead to the (false) impression that applying a linear model is erroneously specified. However, this potential impression is driven by only 5 out of 196 participants. Confidence intervals are (too) huge for these individuals. Excluding these individuals from data analyses does not change our results in any regard. Modelling a curvilinear relation while controlling for mundane statement does not speak in favor of a curvilinear relation.

In sum, we are confident that the results and our conclusions are valid.

The Authors


The above comment is authored by the authors of the article, Stefan Pfattheicher and Simon Schindler, who agreed to PLOS ONE Editorial Staff posting this comment on their behalf in response to the comments "Not worthy of PLoS - see Ashley Landrum's critique" (http://goo.gl/xqtLAh), "Bull5hit!" (http://goo.gl/AaUGry), and "Not finding result" (http://goo.gl/0ZR4eH).