1. A proper survey of religious attitudes would include a section where the respondent could provide a definition for 'God' (perhaps by filling out boxes representing specific properties, e.g., 'created the universe,' 'omniscient,' 'has a tendency to smite dem dere slovenly heathens'). I'm confident that a survey that bothered to do this would reveal that there are so many combinations of properties being ascribed to the word (i.e., so many different definitions) that it simply isn't meaningful -- or otherwise is positively misleading -- to group together everyone who affirms belief in some definition of the word or other.
For example, some not insignificant class of Democrats in the survey are affirming belief in a New Age entity: the universe as a whole, an eternal Karmic spirit, or some other such non-sense. What's the purpose of grouping such persons in with Christians just because they both use the same word for the object of their belief ('God')?
2. The actually interesting result is that almost a quarter of respondents affirm no belief in God. I've never seen such a high proportion, which -- unless I'm simply ill-informed -- means either that the design of the survey didn't utilize a genuinely random sample, or that the trend toward atheism has accelerated significantly.
3. To answer the OP's question, yes, at least 40% of Democrats are stupid, but not because they think they have upheld orders from God (as they define it). While there's a negative correlation between religiosity and intelligence, probably the causality mostly* runs as follows: 'already stupid --> exhibit religiosity'.
*I say 'mostly' because I believe that the causality also runs the other way, to some degree: once the beliefs (for whatever reason) set up camp in a person's mind, their presence can apparently lead to a higher prevalence of cognitive biases that hamper, or perhaps even disable, that person's ability to think -- about certain subjects, anyway.
