 |
 |
 |
 |
 |
 |
 |
|
|||||||
|
|
Predicting Whether or Not People have had Sex with a Partner
Predicting Sex with the Last Date.
Reiss's theory indicates that the person's beliefs about the acceptability of sexual relations with a partner depends on their sexuality standards and the context of the relationship in which the behavior occurs. Thus measures of sexuality standards and attachment to the last date should predict whether or not the person engaged in sex with that date.
As presented in the table below, the group that did not report having sex with their last date scored as less permissive of the Modified Sexuality Standards Scale (MSSS) and reported a lower score for attachment to date. The group who reported having sex with their last date scored as more permissive on the MSSS and reported more attachment to their last date.
| Means of the Predictors for each group. | ||
|---|---|---|
| Had sex with last date? | MSSS Score | Attachment to date |
| No | 15.5 | 3.4 |
| Yes | 22.4 | 4.7 |
The MSSS and attachment to last date were entered into a discriminant
function analysis. The discriminant analysis resulted in one function
that was significant (Wilk's lambda = .73, p < 0001)
and accurately classified 78.6% of cases. This function was defined
by the modified Sexuality Standard Scale score (r = .67) and attachment
to last date (r = .68) and was correlated .52 with group membership
indicating that the discriminant function accounted for 27.04%
of the variability in group membership. The discriminant function
correctly classified 87.1% of the people who did not have sex
with their last date (a 16.3% increase in accurate prediction
above prior probabilities) and 58.1% who did have sex with their
past date (a 28.9% increase in accurate prediction above prior
probabilities). Tau was calculated in order to determine the overall
increase in predictive accuracy over that obtainable by using
prior probabilities of group membership. Tau indicates that using
the discriminant function equation increases predictive accuracy
by 58.4% over prior probabilities in predicting group membership.
| Had Sex with Last Date? | |||
| Actual Group Membership | Predicted Group Membership | ||
| No | Yes | ||
|---|---|---|---|
| No (% accurately classified) Prior Probability |
222 (87.1%) 70.8% |
33 (12.9%) |
255 |
| Yes (% accurately classified) Prior Probability |
44 (41.9%) |
61 (58.1%) 29.2% |
105 |
| Unclassified | 1 | ||
Back to top of page
Do Scores on the Sexual Opinion Survey Increase Predictive Power?
Since I also had scores on the Sexuality Opinion Survey (SOS), I wanted to see if adding the SOS as a predictor in the analysis would increase the accuracy of predicting group membership.
The SOS, MSSS, and attachment to last date were entered into a discriminant function analysis. As presented in the table below, the group that did not report having sex with their last date scored as more erotophobic on the SOS, less permissive of the Modified Sexuality Standards Scale (MSSS) and reported a lower score for attachment to date. The group who reported having sex with their last date scored as more erotophilic on the SOS, more permissive on the MSSS and reported more attachment to their last date.
| Means of the Predictors for each group. | |||
|---|---|---|---|
| Had sex with last date? | SOS | MSSS | Attachment to date |
| No | 50.6 | 15.5 | 3.5 |
| Yes | 61.1 | 22.4 | 4.7 |
The discriminant analysis resulted in one function that was significant
(Wilk's lambda = .73, p < 0001) and accurately classified
77.7% of cases. This function was defined by the modified Sexuality
Standard Scale score (r = .66), attachment to last date (r = .68)
and the SOS (r = .37) and was correlated .52 with group membership
indicating that the discriminant function accounted for 27.04%
of the variability in group membership. The discriminant function
correctly classified 86.6% of the people who did not have sex
with their last date (a 15.6% increase in accurate prediction
above prior probabilities) and 55.8% who did have sex with their
past date (a 26.7% increase in accurate prediction above prior
probabilities). Tau was calculated in order to determine the overall
increase in predictive accuracy over that obtainable by using
prior probabilities of group membership. Tau indicates that using
the discriminant function equation increases predictive accuracy
by 56.8% over prior probabilities in predicting group membership.
| Had Sex with Last Date? | |||
| Actual Group Membership | Predicted Group Membership | ||
|---|---|---|---|
| No | Yes | ||
| No (% accurately classified) Prior Probability | 220 (86.6%) 70.9% |
34 (13.4%) |
254 |
| Yes (% accurately classified) Prior Probability |
46 (44.2%) |
58 (55.8%) 29.0% |
104 |
| Unclassified | 1 | ||
The addition of the Sexual Opinion Survey did not add much to
our ability to predict whether or not people had sex with their
last date over using only scores on the Modified Sexuality Standards
Scale and scores on Attachment to last date. I suspect that is
because the Sexual Opinion Survey contains items that tap various
dimensions of affective reactions toward
sexuality which include attitudes toward: homoeroticism, promiscuity,
pornography and homosexuality. Most of these dimensions are not
directly related to whether a person would have sex with their
past date, however the promiscuity factor comes closest. Since
the Modified Sexuality Standards Scale and Attachment to Last
Date are more closely related conceptually to whether or not the
person had sex with their last date, these measures should do
better in predicting sex with last date.
Back to top of page
Predicting Whether or not they have had sex with the person for whom they took the Love Test Experience Version
How well do sexuality standards alone predict whether they report having or not having had sex with the person for whom they took the Experience Version of The Love Test? To answer this question, I selected unmarried participatants who had met at least one time in person.
| Percent of The Love Test Sample who Report that They Have or Have Not had Sex with the Person for Whom They Took the Love Test Experience Version by Sexuality Standards Group. | |||
|---|---|---|---|
| Had sex? | Sexuality Standards Group | ||
| < Love | Love | > Love | |
| No | 15% | 30% | 80% |
| Yes | 85% | 70% | 20% |
| Total Percent | 100% | 100% | 100% |
r = .35, Chi square (2) = 864.6, p < .001
There are striking differences between the sexuality standards groupings in terms of whether they report having had sex with their romantic interest. 85% of people who say they only need liking or physical attraction prior to sex report having had sex with their romantic interest, 70% of people who say they require love prior to the onset of sexual intercourse report having had sex with their romantic interest whereas only 20% of people who say they need engagement or marriage prior to sex report having had sex with their romantic interest.
Reiss's theory indicates that the person's beliefs about the acceptability of sexual relations with a partner depends on their sexuality standards and the context of the relationship in which the behavior occurs. Thus a measure of sexuality standards, self reports of love for the person, estimates of how likely they are to be together in five years and self reports of lust for partner should predict whether or not the person engaged in sex with that date.
The measure of sexuality standards, self reports of love for the person, estimates of how likely they are to be together in five years and self reports of lust for partner were entered into a discriminant function analysis. As presented in the table below, the group that reported having sex with their romantic interest scored as more permissive of the Sexuality Standards Item, reported more love for their partner, reported a higher liklihood of being together in 5 years and reported higher levels of lust for their partner as compared to the group who reports not having had sex with their romantic interest.
| Means of the predictors entered into the discriminant analysis | ||||
|---|---|---|---|---|
| Had sex with last date? | Sexuality Standard (0-5) | Love for partner (0-10) | Likelihood of being together in 5 years (1-5) | Lust for partner (0-10) |
| No | 2.4 | 8.4 | 2.8 | 8.0 |
| Yes | 3.5 | 8.7 | 3.2 | 8.7 |
The discriminant analysis resulted in one function that was significant
(Wilk's lambda = .70, p < 0001) and accurately classified
77.5% of cases. This function was defined by the modified Sexuality
Standard Scale score (r = .89), lust for partner (r
= .26), likelihood of being together in 5 years (r = .30)
and love for partner (r = .12) and was correlated .55 with
group membership indicating that the discriminant function accounted
for 30.25% of the variability in group membership. The discriminant
function correctly classified 58.3% of the people who did not
have sex with their last date (a 17.7% increase in accurate prediction
above prior probabilities) and correctly classified 90.6% of the
people who did have sex with their past date (a 31.2% increase
in accurate prediction above prior probabilities). Tau was calculated
in order to determine the overall increase in predictive accuracy
over that obtainable by using prior probabilities of group membership.
Tau indicates that using the discriminant function equation increases
predictive accuracy by 61.5% over prior probabilities in predicting
group membership.
| Had Sex with Partner? | |||
|---|---|---|---|
| Actual Group Membership | Predicted Group Membership | n | |
| No | Yes | ||
| No
(% accurately classified) Prior Probability |
932
(58.3%) 40.6% |
666
(41.7%) |
1598 |
| Yes
(% accurately classified) Prior Probability |
220
(9.4%) |
2118
(90.6%) 59.4% |
2338 |
Back to top of page
The three sets of results produced convergent findings
Using three different samples that used different measures of sexuality
standards and attachment to partner, I obtained convergent findings
for the question of whether sexuality standards and attachment
to partner could predict whether or not someone had sex with a
partner. In all three analyses, I was able to accurately classify
almost 80% of participants in terms of whether or not they report
having had sex with a partner. The highest correlation with the
discriminant function in all three analyses was the measure of sexuality
standards. However, for the Texas sample, measures of attachment
to the romantic partner showed a stronger relationship to the
discriminant function than they did in the Love Test sample. The
addition of the measures of commitment and lust for partner significantly
added to our ability to predict group membership in the Love Test
sample where the addition of the Sexual Opinion Survey in the
analysis for the Texas sample did not add to our ability to predict
whether or not the person had sex with the partner. It's interesting
to note that the prediction equation achieved the most accurate
prediction for the group who report not having sex with their
last date in the Texas sample, but achieved the most accurate
prediction for the group who reports having had sex with their
partner in The Love Test sample.
Back to top of page
|
|
|
|
|
|
|