Follow-Up: Theonormativity in Psych Studies

My post describing a seriously flawed psych study in which my daughter participated attracted a bit of attention a while back. I sent my criticisms of the methodology to the researchers; that was what people read here.

Well, they wrote back. The text of their email to me is blockquoted italics; my responses are interpolated. Judge for yourself how well they address the issues I raised:

Dear Warren,

Thank you both for participating in our studies and for sharing your concerns about this study with us. We share your concerns about the line of questions that you discussed in your message; if we can impose on you a bit further, we would appreciate your help in addressing them.

Me: Sure.

First a bit of background. The purpose of this study is to investigate how children feel about other children who do or do not share their beliefs and behaviors. Thus, the data from your daughter that we analyze concern her choices between two children, one of whom shares a belief (but not a behavior) she expressed earlier in the study, and the other of whom shares a behavior (but not a belief) with her. In order to interpret these data, however, we need to know which target child’s belief and behavior each child in the study shares. The preliminary questions that are the focus of your concern were designed to provide this information.

Me: I understand that; thank you for reiterating the aim of your research. Was your rationale for including religious questions simply that they provided an example of a “belief.”? See below for more on this line of questioning.

We worked very hard to ask these questions in a way that would not presuppose one answer or another. It turns out to be particularly difficult to do this, however, when the beliefs concern the existence of one or more entities. “Do you or do you not believe that the tooth fairy puts money under your pillow” is an ambiguous question: one interpretation presupposes the tooth fairy’s existence (does she put money or candy under your pillow?); the other interpretation questions her existence (does she, or your parents, put money under your pillow?). We attempt to focus children on the second interpretation by using intonation, and contrasting belief claims, that emphasize the entity itself rather than the action or property that is attributed to it.

Me: Well, to begin with, one approach is to simply leave out questions based a belief that the child does not share. If my daughter doesn’t believe in “god,” then jump ahead to the next larger belief category, and omit the questions that presuppose a belief in “god.” If what you’re really after is “how children feel about other children who do or do not share their beliefs and behaviors,” then you can do that easily, without asking an avowed unbeliever whether she thinks god “hears the songs people sing to him.”

If the person says, “Yes, I believe there is only one god,” then ask the pertinent sub-questions. If not, leave them out.

This approach would resolve “poor heuristic structure” criticism. See below for more details.


“Here is Timmy. Timmy believes in god, but he doesn’t sleep on a pillow.”

Unfortunately, we have found no framing that children are able to understand that specifies the first interpretation decisively (Obviously, we avoided the unambiguous framing that *does* presuppose the entity’s existence, such as “Do you believe that the Tooth Fairy does or does not put money…”).

Me:Unfortunately, “Do you or do you not believe that god hears the songs people sing to him?” is pretty unambiguous. Most grownups, looking at that question, would be hard put to read it as one which “weights” the nonexistence of a putative deity equally with his/her/their/its existence. It’s probably asking too much for a six-year-old to do that. Furthermore, the question is also “weighted” in favor of a male version of the deity (although I notice that your gender default for the tooth fairy is female).

Asking for a “yes” or “no” answer to this question provides you with an extremely coarse taxonomic grid for the analysis of subject responses. One child might answer “no,” reasoning that “god” is female; another might answer “no,” reasoning that god is male but deaf; another might reason that god doesn’t exist, etc., etc., etc.

If your heuristics had a more robust internal structure, you could handle this more easily.


Question 6: “DYODYNB (do you or do you not believe) there is such a thing as a god?” (notice that this phrasing admits to polytheistic options).

IF “YES,” go to 6-A. IF “NO,” go to 7.

6-A: “DYODYNB that there is more than one god? (this is better IMO than “only one” as the default)

IF “YES,” go to 6-A-1. IF “NO,” go to 6-B

6-A-1: DYODYNB that there are three or more gods?

etc., etc., etc.

6-B: DYODYNB that there is only one god?

IF “YES,” go to 6-B-1. IF “NO,” ask subject to explain.

6-B-1: DYODYNB that “god” is a boy?

IF “YES,” go to 6-B-2. IF “NO,” go to 6-B-3.

etc., etc., etc.

6-C: DYODYNB think that when people sing to a god, that he/she/it hears those songs?

7: DYODYNB think that brushing your teeth is a good idea?


If god-related questions turn out to be too much trouble to ask, why ask them? The minds of children are full of beliefs:

DYODYNB that when you step on a crack, you’ll break your mother’s back?
DYODYNB that Santa Claus brings you presents at Christmas?
DYODYNB that Harry Potter is real?
DYODYNB that your dog understands the words you say?
DYODYNB that wishing really hard for something will make it happen?
DYODYNB that there are little people inside the TV?
DYODYNB that your father/mother knows everything in the world?

Conversely, if you consider the god-related questions to be at the core of your study, why not frame your research appropriately? If what you’re really after is plumbing the depths or scaling the heights of children’s theological orientations, and how they influence the perception of others — then it’s time to change your description.


Theonormativity, like other forms of default assumptions about other people, is very tricky to recognize as long as it’s an assumption that you share. Disbelief is no protection; many atheists are theonormative in the sense that they conceive of a specific God which serves as a mental placeholder for a larger “god-concept.” A “Christian atheist” would be one who thinks of Yahweh and the personae of the Christian myths as normative objects of rejection — a category distinguished from “pure” atheism, which simply eschews theonormativity in any sense (recognizing it in others isn’t the same thing, of course; a clinical psychologist needn’t believe in the corporeal existence of a schizophrenic’s hallucinations to discuss them rationally).

In Western European culture, some form of Abrahamism provides the theonormative framework for subsequent schisms based on soteriological, eschatological or teleological minutiae. In India, Hindu theonormativity is the usual default (“DYODYNB in Brahma, Vishnu and Shiva?”).

As you are conducting your experiments within the context of Western European culture, it behooves you to examine your own default assumptions with extreme care, and to work meticulously to eliminate their influence on your research.

It is interesting that this conversation is now addressing some of the same questions you’re asking about children.

If you can suggest a better phrasing for these questions, we would be very grateful to learn it. And thanks again for sharing your concerns with us.


They responded:

Dear Warren,

Thank you for these ideas. We appreciate you taking the time to share them with us and will consider them as we design future studies.


Leave a Reply

Your email address will not be published. Required fields are marked *