Survey Development

Before developing the survey, operational definitions of each of the survey constructs were developed through a review of modern peer-reviewed literature on the topics of female sexuality and marital satisfaction. This ensured that each question was related to our research query, and each question was developed with a rationale based in current scientific literature in order to reduce researcher bias and increase validity.

Survey questions were developed using, where applicable, previously validated studies. For example, questionnaires designed to evaluate the sexual health of women in conservative cultures were utilized. Because these questionnaires were used in non-western countries, questions were minimally edited to make them more accessible in a North American context. Internal reliability checks were inserted throughout the survey by placing a minimum of two questions testing the same construct with alternating polarization on different pages throughout the survey. Multiple reliability and validity checks were built into the survey and were run after responses were collected. Any questions that did not pass said checks were removed from our analyses and are not included in The Great Sex Rescue. In order to reduce bias against or for any particular evangelical resources, we did not name any resources or authors in the survey. The only exception was the belief we measured, “all men struggle with lust; it is every man’s battle.” Despite our desire to remove the last clause of that question, our pilot test determined it was necessary. Any references to books, resources, authors, or organizations being named helpful or harmful came from open-ended questions simply asking which (if any) resources helped or harmed respondents’ marriages. It is worth noting, these questions were not required answers, in order to not force respondents to choose arbitrarily.

Each question was independently reviewed by all members of the research team. Whenever possible, questions used a Likert scale as the response options.

A pilot study was conducted in which the survey was reviewed by outside observers with a variety of marital histories for quality control. Questions were edited or removed from the survey as a result of the pilot study, as is best practice in psychometrics. The survey was distributed beginning on November 1, 2019 and data collection continued until January 5, 2020.

Recruitment of Participants

Recruitment was accomplished in three ways:

1. Self Promotion

First of all, we advertised our survey to the readers of Sheila’s website, tolovehonorandvacuum.com. Different survey links were used for the blog, promotional emails, and social media posts so that differences between participants from each stream could be investigated.

2. Influencers

We reached out to many influencers to spread the word outside of our audience. Influencers who were willing to share the survey were each given a unique collector weblink to the survey.

3. Recruiter Program

Blog readers and other participants were asked if they would serve as recruiters and let other people know about the study. If they agreed to, they were given a unique code so that respondents who were referred by each recruiter could be identified for the purpose of counting.

This three pronged approach ensured that we could identify participants who do not follow the To Love, Honor, and Vacuum blog closely. Because the blog writes frequently about controversial topics in Christianity, we were concerned about biases present in the audience and wanted to ensure that we had a group of respondents who were not already exposed to the ideas espoused on the blog. Due to our three-pronged approach, over half of our respondents were from outside sources, meaning that Sheila’s readers made up a minority of the survey responses. Additionally, the majority of survey respondents currently hold views that are not shared by the authors of the book.

We aimed for a sample large enough to run analyses on less common sexual outcomes, such as vaginismus to the extent that penetration is impossible. Our survey succeeded in this. In building our recruitment protocol, we did not aim for a specifically representative survey, since it was unnecessary for our research question. Our research question was not about prevalence of beliefs or marital and sexual satisfaction outcomes in the general evangelical population, but rather effects of belief on those outcomes. As a result, representativeness was not required for our analyses, merely a large enough discrepancy in beliefs among the respondents, which we received.

To note, many of our frequency statistics you will see below echo peer-reviewed research, suggesting that our survey respondents were not unduly over- or underrepresented in terms of the sexual outcome variables we were addressing in the general population.

Demographics

Our survey was of heterosexual women who had ever been married ages 18 or above. Statistics reported in The Great Sex Rescue include data from women who are professing Christians.

Some additional information about our survey demographics:

*note: frequencies may add to more than 100 due to rounding.

• Age breakdown:
  • 18-19 years old: 0.2%
  • 20-24 years old: 4.1%
  • 25-29 years old: 12.0%
  • 30-39 years old: 37.1%
  • 40-49 years old: 25.3%
  • 50-59 years old: 14.1%
  • 60-69 years old: 5.9%
  • 70 year old or older: 1.1%

• Our survey respondents were highly religious:
  • 51.7% attended church more than once a week.
  • 23.6% attended church once a week.
  • 8.7% attended church a few times a month.
  • 8.1% attended church a few times a year.
  • 7.9% attended church once a year or less.

• Our survey respondents were majority evangelical.
  • 64.6% currently identified as evangelical.
  • 12.9% no longer identify as evangelical, but used to.
  • 22.5% do not identify as evangelical and never have.

• The vast majority of respondents were raised in a home with at least one believing parent:
  • 70.9% were raised in a home where both parents were Christian.
  • 13.7% were raised in a home where one parent was a Christian.
  • 15.3% were raised in a home with non-Christian parents.

Reported belief in evangelical teachings

Our survey’s function was to test the effects of evangelical teachings on women’s sexual and marital satisfaction outcomes. To test this, we needed to ensure that our respondents held differing beliefs so that we could contrast groups. Our survey respondents reported a wide range of beliefs, which allowed us to do just that.

We found the following prevalence belief and exposure rates of these common evangelical teachings on marriage and sex (for information on the outcomes of these beliefs, see The Great Sex Rescue):

Boys will want to push girls’ sexual boundaries.

• Before marriage: 88% of respondents believed this before they were married, and 62% of respondents report having been taught this message by church, Christian media, and/or family before marriage.
• After marriage: 81% of respondents reported believing this currently, and 40% reported being taught this message by their church and/or Christian media today.

All men struggle with lust; it is every man’s battle.

• Before marriage: 79% of respondents agreed with this before marriage, 63% reported being taught this by the church, Christian media and/or family before marriage.
• After marriage: 62% of respondents reported believing this currently, and 52% reported being taught this by their church, Christian media and/or family today.

If you have sex before you are married, that means you have less of yourself to give to your future spouse because a part of you will always belong to someone else.

• Before marriage: 76% of respondents agreed with this before marriage, 76% reported being taught this by the church, Christian media and/or family before marriage.
• After marriage: 57% of respondents reported believing this currently, and 63% reported being taught this by their church, Christian media and/or family today.

The only biblical reason for divorce is an affair.

• Before marriage: 56% of respondents agreed with this before marriage, 63% reported being taught this by the church, Christian media and/or family before marriage.
• After marriage: 25% of respondents reported believing this currently, and 35% reported being taught this by their church, Christian media and/or family today.

It is a man’s God-given role to provide for his family and a woman’s God-given role to stay home raising her children.

• Before marriage: 53% of respondents agreed with this before marriage, 57% reported being taught this by the church, Christian media and/or family before marriage.
• After marriage: 43% of respondents reported believing this currently, and 33% reported being taught this by their church, Christian media and/or family today.

In a family with children, a working mom and a stay at home dad is as good as a stay at home mom and a working dad.

• Before marriage: 47% of respondents agreed with this before marriage, 13% reported being taught this by the church, Christian media and/or family before marriage.
• After marriage: 62% of respondents reported believing this currently, and 13% reported being taught this by their church, Christian media and/or family today.

Women should have frequent sex with their husbands to keep them from watching porn.

• Before marriage: 40% of respondents agreed with this before marriage, 34% reported being taught this by the church, Christian media and/or family before marriage.
• After marriage: 18% of respondents reported believing this currently, and 17% reported being taught this by their church, Christian media and/or family today.

A wife is obligated to give her husband sex when he wants it.

• Before marriage: 39% of respondents agreed with this before marriage, 43% reported being taught this by the church, Christian media and/or family before marriage.
• After marriage: 21% of respondents reported believing this currently, and 23% reported being taught this by their church, Christian media and/or family today.

overall frequency of sexual satisfaction outcomes

The following tables detail some of the primary sexual variables analyzed in relation to evangelical teachings. You will find the overall frequency data here:

Orgasm frequency in current marriages

Almost always or always	48.7%
Slightly more than half of the time	18.7%
Half of the time	9.5%
Less than half of the time	11.0%
Almost never or never	12.1%

Orgasm frequency in previous marriages that ended in divorce

Almost always or always	20.3%
Slightly more than half of the time	17.1%
Half of the time	13.2%
Less than half of the time	19.8%
Almost never or never	29.7%

Orgasm frequency in previous marriages that ended in death

Almost always or always	33.9%
Slightly more than half of the time	22.8%
Half of the time	14.4%
Less than half of the time	13.9%
Almost never or never	15.0%

Frequency of Sexual Pain

Sex has never been painful	27.3%
Only a little discomfort that changing positions fixed	40.5%
I have experienced pain that took the fun out of sex	25.5%
Penetration was impossible due to pain	6.8%

Libido Differences (Currently)

The wife has the higher libido	22.0%
The husband has the higher libido	58.2%
The husband and wife have the same libido	19.8%

Results

Correlations between common evangelical beliefs about sex and marital and sexual satisfaction outcomes were identified using the confidence interval method for odds ratios. Other statistical tests, such as the test for proportions and the chi-square test for independence were used when appropriate. For more on our research, check out The Great Sex Rescue and check back soon to see when our research is published in peer reviewed journals

FAQs

Here are some commonly asked questions about our research:

Are you concerned that your respondents all thought the same?

No, we are not, because our respondents had all sorts of different beliefs. In fact, if our respondents had all had the same viewpoint we wouldn’t have been able to do a comparison study! Claiming that our over 20,000 respondents all had the same viewpoint shows a fundamental misunderstanding of the nature of odds ratios. We were able to make our conclusions only because we had a diverse set of beliefs among the survey respondents. If the respondents all had the same point of view, we would not have been able to make comparisons. For instance, in order to judge how the belief “all men struggle with lust; it’s every man’s battle” impacted women, we needed both women who believed it and women who didn’t.

How did you make sure you didn’t ask leading questions about the beliefs you measured?

We did this in multiple ways. We measured three variables with our survey: first, marital satisfaction. To reduce bias due to leading questions, we used previously validated survey questions whenever possible and we didn’t ask about beliefs at all in this section. Second, sexual satisfaction. We used the same protocols here, and we were simply measuring various aspects of female sexual satisfaction, nothing about beliefs or introspective accounts of what has helped or harmed them in this area. Just do you orgasm, are you aroused, are you satisfied with how emotionally close you are with your spouse after sex, those kinds of things.

Finally, we asked about beliefs. We did not mention any evangelical teachings until after we had already gathered all the information we would need for marital and sexual satisfaction so that the respondents could not alter their answers in response to the beliefs. When we asked about beliefs, we did not ask for any subjective interpretation of beliefs, any introspection on how the beliefs harmed them, or even which specific resources had taught them these beliefs before. We simply measured (a) if (and where) they were taught it earlier in life, (b) if they believed it earlier in life, (c) if (and where) they are being taught it today, and (d) if they believe it today.

We did not say, “do you think this belief harmed you?”, nor did we say, “we are going to give you a set of harmful beliefs to see what you think.” Instead, we listed beliefs, some healthy and some unhealthy, all mixed up with each other so that there was no indication of whether we thought they were healthy or not, and asked if respondents believed them. This is one of the many ways we limited bias.

Do you think your sample had a lot of people who are already angry at evangelical marriage advice?

We recruited people to the survey with a wide variety of social media messages, in order to attract people with a wide variety of views (we needed as wide variety as possible in order to make comparisons). At one point we recruited people by asking, “have you ever been bothered by Christian marriage teaching?” We did this particular strategy because once we had about 10,000 respondents, we saw that our pool was overwhelmingly conservative, and we needed more progressive or liberal voices in order to make comparisons, so we tried using messages that would recruit those people. Even so, the largest group of people to take our survey are complementarian, and do agree with some of the harmful teachings we identified. Additionally, as shown above, even if we did have all angry respondents, there really was no way for them to target or present biased opinions against any particular teaching or author in the quantitative survey data, which is the only data included in our current analyses and the book.

But weren’t these all just your readers?

We had over 230 unique links that were used to send people to our survey, so we can measure where people came from. We were thrilled to see that over half of our respondents were not from Sheila’s links at all. Additionally, many people shared Sheila’s social media link (including some of the authors that are now accusing us of only recruiting our own people). Therefore, even the people who used Sheila’s link did not all come from Sheila’s list.
However, once again, even if they did, that would not make a difference to our survey results, since our main findings were not about frequency, but were instead comparing two different groups based on whether or not they believed a teaching. Whether they came from Sheila’s list or not is irrelevant; the issue was whether or not they believed each teaching. It just so happens that we also managed to conduct a survey where over half of the respondents did not come from our list, even though it was not required for our research.