Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2009 Jan 19.
Published in final edited form as: Health Psychol. 2007 Nov;26(6):777–786. doi: 10.1037/0278-6133.26.6.777

Willingness to Use Microbicides Varies by Race/Ethnicity, Experience With Prevention Products, and Partner Type

Kathleen M Morrow 1, Joseph L Fava 2, Rochelle K Rosen 2, Anna L Christensen 2, Sara Vargas 2, Candelaria Barroso 2
PMCID: PMC2628547  NIHMSID: NIHMS80709  PMID: 18020851

Abstract

Objective

To investigate women's willingness to use vaginal microbicides to reduce/prevent HIV infection, using measures grounded in the individual, behavioral, and social contexts of sex.

Design

A cross-sectional study that enrolled a sample (N = 531) of 18−55 year old Latina, African-American, and White women in the U.S. between October, 2004, and July, 2005.

Main Outcome Measures

Willingness to use microbicides and individual- and context-related variables (e.g., demographics, relationship status).

Results

Exploratory and confirmatory factor analyses supported a one-dimensional, 8-item scale, with high internal consistency (α = .91). Subgroup analyses within the Latina (n = 166), African-American (n = 193), and White sub-samples (n = 172) also supported a unidimensional scale with strong internal validity and high reliability. Race/ethnicity as a contextual factor, a woman's history of using prevention products, and the nature of the sexual partnership were predictive of willingness to use microbicides (R = .41). That is, women with greater frequencies of condom use, a history of spermicide use, and non-main sexual partners had higher predicted Willingness to Use Microbicides scale scores, while White women had lower predicted scores.

Conclusion

The Willingness to Use Microbicides scale serves as the first psychometrically validated measure of factors related to microbicide acceptability. Developing and implementing psychometrically validated and contextualized microbicide acceptability measures, in an effort to understand microbicide users and circumstances of use, is crucial to both clinical trials and future intervention studies.

Keywords: HIV, microbicide, acceptability, willingness, scale development, race/ethnicity, partner type, prevention products

Globally, approximately half of all new HIV infections occur in women and the proportion of newly infected women is increasing (Ogden, Ogden, Mthembu, & Williamson, 2004; UNAIDS/WHO, 2005). According to the Centers for Disease Control and Prevention [CDC], women represent approximately 30% of new infections in the United States (U.S.) each year, three-quarters of which are transmitted via heterosexual intercourse (Centers for Disease Control and Prevention [CDC], 2005). Among sexually active women, a woman's risk for HIV infection is often related to behaviors beyond her direct control, such as the sex- or drug-related behavior of a sexual partner (Hader, Smith, Moore, & Holmberg, 2001).

Given the rise in the proportion of women becoming infected with HIV annually and the limited ability of women to exercise condom use with their male sexual partners (Beksinska, Rees, McIntyre, & Wilkinson, 2001; Parker, Easton, & Klein, 2000; Sapire, 1995; Smith, 2003; World Health Organization [WHO], 1997; Wingood & DiClemente, 1998), the need for a woman-initiated HIV prevention option is profound. Topical vaginal microbicides offer the best hope for protecting women from HIV/AIDS. The term “microbicide” applies to a range of products that, when applied topically to a woman's vagina prior to sexual intercourse, could prevent or significantly reduce HIV transmission; some may also provide protection against other STDs, and still others may have contraceptive effects (Alliance for Microbicide Development, 2004; Global Campaign for Microbicides, 2006; International Partnership for Microbicides, 2005). Several candidate vaginal microbicides are now undergoing clinical efficacy trials.

In preparation for FDA approval of microbicides, behavioral and social scientists have been exploring psychosocial factors hypothesized to be associated with the acceptability and continued use of these products. It will be critical for successful uptake of the products, and thus their impact on rates of HIV transmission, to understand who would use microbicides and why. For instance, will product use vary by partner type or number, cultural influences on the sexual context, or other factors (Severy, Tolley, Woodsong, & Guest, 2005)? Given a greater understanding of which factors affect microbicide uptake and use, behavioral and social scientists, as well as health care providers and systems, will be better able to develop and target appropriate interventions. The Phoenix Project was designed to gather formative microbicide acceptability data to develop and psychometrically evaluate microbicide acceptability measures.

Several health behavior models, including the theories of reasoned action and planned behavior, posit that behavioral intention, or the “subjective probability” that one will perform a given behavior (Fishbein & Ajzen, 1975), can predict eventual behavior. In one meta-analysis (Sheeran & Orbell, 1998), a medium to strong correlation was found between intentions and condom use, though the degree of relationship has been found to be affected by such factors as gender and partner type (Fishbein et al., 2001; von Haeften, Fishbein, Kaspryzk, & Montano, 2000). Thus, while some studies have found a significant relationship between measures of intention and measures of safer sex behavior (Albarracin, Johnson, Fishbein, & Muellerleile, 2001; Basen-Engquist, 1992; Basen-Engquist & Parcel, 1992; Bryan, Rocheleau, Robbins, & Hutchinson, 2005; Fishbein et al., 2001; Fisher, Fisher, & Rye, 1995; Godin, Gagnon, & Lambert, 2003; Posner, Bull, Ortiz, & Evans, 2004; Roberts & Kennedy, 2006; Sheeran, Abraham, & Orbell, 1999), others have found less evidence for such a relationship (Minnis, Shiboski, & Padian, 2003; Rosenthal, Fernbach, & Moore, 1997).

It is possible that the way the construct is measured may contribute to these less than consistent results. Fishbein and colleagues (1975) discussed the need for specificity in items relating to intentions (i.e., behavior, target object, situation, and time). Within the realm of self-efficacy, Bandura (1985) and Forsyth and Carey (1998) posited the need to develop self-efficacy items that include a specific class of behavior and a specific context or circumstance. Studies of behavioral intentions have used measures with varying levels of specificity (Albarracin et al., 2001; Fishbein, Hennessy, Yzer, & Douglas, 2003), as have measures of self-efficacy (Smith, McGraw, Costa, & McKinlay, 1996).

Gibbons and colleagues (1998) have postulated that distinct constructs of behavioral intention, willingness, and expectation should be considered when predicting behavioral outcomes. While, at face value, items measuring these constructs may appear very similar, Gibbons and colleagues have found them to contribute to the prediction of behavior, both together and independently. These results and the ongoing discussion of how best to contextualize the measurement of such constructs as intention and self-efficacy led the Phoenix Project to contextualize items related to willingness to use microbicide, specifying partner, partner type, and specific episode of sexual behavior, as well as specifying certain microbicide characteristics and accessibility referents. Willingness was chosen because of the still hypothetical nature of a microbicide, in contrast to intention and expectation, which imply an actual plan or goal, in an environment where proof of concept and, therefore, a readily available product, does not currently exist. Thus willingness, or the likelihood of using a microbicide under specific circumstances, best approximated the current status of microbicide development.

As part of a larger effort to develop a set of microbicide acceptability measures for use in upcoming studies, the Phoenix Project gathered formative microbicide acceptability data to elucidate themes and constructs related to product use, developed quantitative items to measure those constructs, and tested the resulting measures. With respect to the current analysis, a measure of “willingness to use microbicides” was developed and evaluated. The resulting scale was then used in further analyses to explore demographic and other variables related to willingness to use microbicides.

In the case of microbicide use, a woman's willingness to use a product under specific circumstances in the future may predict her actual use. Hence, willingness to use microbicides will ultimately affect the actual uptake and sustained use of a product. Our hypothesized model of acceptability assumes that product-related factors, as well as person-related factors, will be associated with a woman's willingness to use a microbicide.

Method

Design and Procedure

A cross-sectional questionnaire was administered using an audio computer-assisted self-interview (A-CASI) format on laptop computers equipped with headphones. Screening was conducted either in person or over the telephone. If face-to-face, procedures as described for questionnaire administration (below) were utilized: if by telephone, research staff would read each question and potential responses as indicated and enter the potential participant's responses into the database. If eligible, potential participants were provided with a more comprehensive description of the study. Informed consent forms were read to them, with staff providing further explanations and responding to any questions.

Once consent was obtained, staff oriented participants to the computer and initiated a brief practice session for those who wished. Staff positioned themselves in a location where they could oversee the session without infringing on the privacy of the participant, thus being available for questions or any technical computer-related issues. For each item, questions and response spaces or options appeared on the screen simultaneously with an audio recording that read through the question and the response options. Audio read-out could be interrupted by providing a response or could be muted, such that participants who read faster could move through the instrument at a faster pace, if desired. Instruction screens could not be interrupted. All procedures were approved by both primary and site-specific human subjects protection institutional review boards.

Participant Recruitment and Enrollment

Overall, women from four (4) states in the northeast U.S. (Connecticut, Massachusetts, New York, and Rhode Island) participated. Eligibility criteria (self-report) included being 18−55 years old; being Black/African-American (Black), Latina/Hispanic (Latina), or White; having had vaginal sex with at least one male sex partner in the last 12 months; being HIV-negative or of unknown HIV status; and not being pregnant. Women were recruited via public media (e.g., news articles, radio/TV shows) and advertisements (e.g., newspapers, internet, flyers), as well as presentations at community-based organizations (e.g., minority advocacy groups), social service and treatment agencies (e.g., substance abuse treatment, homeless shelters), and word-of-mouth. Nonproportional quota sampling was used to recruit subgroups sufficient to address the overall research questions. Black, Latina, and White women were enrolled because these are the racial/ethnic groups that make up the largest proportions of HIV-infected women in the United States (CDC, 2005). Latina groups in the U.S. are heterogeneous in culture, social characteristics, and other contextual features; thus we limited recruitment to control for Latina subgroup differences by recruiting only those women whose families were originally from Puerto Rico or the Dominican Republic, as women from these groups represent the greatest proportion of Latinas in the northeast (Grieco, 2004; Guzman, 2001; U.S. Census Bureau, 2000). Women were enrolled based on their partner status (i.e., 1 or 2+ vaginal sex partners) to allow for a better understanding of microbicide acceptability constructs by partner status. The sampling strategy was designed to result in the enrollment of approximately one-third Latina, one-third Black, and one-third White women across sites. With respect to partner status, approximately one-third would have had one (1) male vaginal sex partner in the last 12 months, while two-thirds would have had two or more (2+) male vaginal sex partners in the last 12 months. By enrolling one-third single-partnered women and two-thirds multi-partnered women, we hoped to attain relatively equal numbers of main versus casual and other partnership types for the study's primary analyses.

Participants were reimbursed $30 for completing the questionnaire. Where appropriate, participants were provided up to $10 reimbursement for travel or child/elder care required for them to participate.

Randomization of partner

Women who reported having one sexual partner in the last 12 months were asked to label him as either a “main,” “casual,” or “other” partner. A “main” partner was defined as “someone you feel an emotional commitment to, like a boyfriend, fiancé, husband, or your man.” A “casual” partner was defined as “someone you know, such as a friend, but aren't in any sort of committed relationship with,” and an “other” partner was defined as “someone you don't know much about, for example, a client, or maybe someone you have never met before you have sex.” Women who reported having two or more sexual partners in the last 12 months were asked to identify their least frequent, most frequent, most important, and most recent sexual partners and to label them as main, casual, or other partners, as described above. For multi-partnered women, the computer randomly chose one of her identified partners, and for women who reported only one partner, that person was used. The remainder of the questionnaire asked the participant to focus on the one selected partner in order to set a specific context for microbicide acceptability. Items relating to relationship contexts, sexual behavior, and opinions about microbicides were framed within the context of this particular selected partnership.

Measures

The full questionnaire consisted of an initial eligibility screener and questions that assessed the following content areas, in order: demographics; vaginal product use; preferences for specific microbicide product characteristics; contextual self-efficacy for insertion of vaginal products and negotiation of condom and product use; perceived norms of friends and significant others regarding participant's sexual behavior; sexual partner identification and labeling; recent STD and HIV risk perception; relationship context; relationship quality; relationship communication; perceived partner norms regarding sexual behavior; sexual behavior with selected partner; perception of STD and HIV risk with partner; context of participant's last sexual episode with partner; willingness to use microbicides; importance of specific microbicide characteristics; contextual self-efficacy for microbicide use; sexual health, STD history, and HIV testing history; drug use history; sociotropic cognitions; STD knowledge (Jaworski & Carey, 2001); and HIV knowledge (Carey & Schroder, 2002). Items were developed for self-efficacy, norms, partner labeling, risk perception, relationship quality, and sexual behavior through a careful review and modification of measures described in the HIV/STD literature (Abraham, Sheeran, Abrams, & Spears, 1996; Anderson, Wilson, Doll, Jones, & Barker, 1999; Armitage & Conner, 2001; Basen-Engquist, 1992; Buunk, Bakker, Siero, van den Eijnden, & Yzer, 1998; Chan, Cheung, Gray, Ip, & Lee, 2004; Civic, 1999; Forsyth & Carey, 1998; Giles, Liddell, & Bydawell, 2005; Jennings-Dozier, 1999; Rise, Thompson, & Verplanken, 2003; Severy, Robinson, Findley-Klein, & McNulty, 2006; Severy et al., 2005; Sheeran & Abraham, 1994; Sheeran & Orbell, 1998; Smith et al., 1996; Soet, DiIorio, & Dudley, 1998; Soler et al., 2000; St Lawrence et al., 1998), as well as expert reviews of proposed items and ordering of those items. Both section (i.e., topic) and item ordering was decided as a function of the study's desire to contextualize a woman's process of responding to the questionnaire. That is, both the topics discussed and the items asked within each topic reflected a conversational and contextual flow developed as a function of cognitive interviews using draft items and expert review feedback. This flow included the framing of transitions and instructions as well as actual items, such that one question led to the next or, when a particular topic was complete, an appropriate transition was made. Final item order for the Willingness to Use Microbicides Scale is reflected in Table 1.

Table 1.

Component and Factor Pattern Item Loadings

Item PCA MLFA
1. If a microbicide had been available the last time you had sex with him, would you have wanted to use one? .82 .82
2. Would you have used a microbicide when you had sex with this partner the last time? .87 .88
3. Would you have had a microbicide available that last time? .73 .66
4. Would you have remembered to use a microbicide when you had sex with this partner the last time? .75 .70
5. If a microbicide cost about as much as a male condom, that is about 75 cents, would you have used it that last time? .81 .77
6. If a microbicide were [preference: available by prescription or available over-the-countera], would you have used it that LAST TIME? .87 .86
7. If a microbicide were available as a [preferred formulationa] and could be [preference: inserted with an applicator or inserted without an applicatora], would you use it in the future with [randomly chosen partner]? .71 .64
8. If a microbicide were available as a [preferred formulationa] and could be [preference: inserted with an applicator or inserted without an applicatora], would you use it in the future with sexual partners who you feel an emotional commitment to? .67 .59
Open in a new tab

Note. PCA = Principal components analysis; MLFA = Maximum likelihood factor analysis.

a

Participant's response to a previous question was automatically inserted by the A-CASI program.

Sample Characteristics

Potential participants were recruited using a two-step process. Study aims required that the sample be approximately equally divided by race/ethnicity (1/3 Latina, 1/3 Black, and 1/3 White) and comprise a 1:2 ratio of single-partnered and multi-partnered women. As such, a non-proportional quota sampling strategy was employed. Details of recruitment procedures, including the effectiveness of nonproportional quota sampling procedures and community partnerships can be found in Morrow et al. (2007). Nine hundred and twenty-two (922) women were screened for the study. Of these, 17% were ineligible as a function of study eligibility criteria and 16% met study eligibility criteria but were not enrolled because target numbers for their particular race/ethnicity and partner number had been met. The final sample (N = 531) included 166 (31.3%) Latinas, 193 (36.3%) Black women, and 172 (32.4%) White women. Of the Latina cohort, 75.9% (n = 126) reported their family's country of origin as Puerto Rico, while 22.2% (n = 37) reported it as the Dominican Republic, and 1.8% (n = 3) reported that their family came from both Puerto Rico and the Dominican Republic. With respect to partner status, 155 (29.2%) reported having one (1) male vaginal sex partner in the last 12 months, while 376 (70.8%) reported having two or more (2+) male vaginal sex partners in the last 12 months. The sample thus closely approximated the desired proportions targeted by the quota sampling procedures.

The mean age of the sample was 33.8 years (SD = 9.6). Black women were significantly older (M = 35.4, SD = 9.5) than either White (M = 33.0, SD = 10.5) or Latina (M = 32.8, SD = 8.4) women (F(2,528) = 4.235, p = .015). Fifty-eight percent (n = 308) reported never having been married. Over half (54%) of the sample reported a high school education or less, over half (55%) were not employed, and over half (52%) reported an annual household income of less than $15,000.

Measure Development Procedures: Willingness to Use Microbicides Scale

Overview

In order to measure the construct of Willingness to Use Microbicides, a 9-item scale was developed. The complete sample (N = 531) available for the study was divided randomly into 2 sub-samples. Sub-sample 1 (N = 265) was used for exploratory dimensional analyses (EDA) and initial reliability analyses to determine the instrument factor structure and the optimal final set of items. Sub-sample 2 (N = 266) was used for confirmatory factor analysis (CFA) and further instrument refinement. A final CFA was done using the complete sample to obtain more stable estimates of item loadings, model fit, and reliability.

Exploratory Dimensional Analysis (EDA): Procedure

EDA were conducted on the set of nine items developed to measure different aspects of willingness to use microbicides. Of the potential 265 cases that were randomized to Sub-sample 1 for these analyses, two subjects did not answer any of the nine items, and 16 subjects responded to only five of the nine items because of a skip pattern. This resulted in an effective sample size of 247 cases available for EDA after a listwise deletion of items. Preliminary item level analyses were conducted to examine item response distribution, means, standard deviations, skew, and kurtosis in each of the nine items, and the results were judged adequate to proceed with the EDA. The Scree Test (Cattell, 1966), the minimum average partial (MAP) procedure (Velicer, 1976), and an implementation of the parallel analysis (PA) procedure (Horn, 1965; O'Connor, 2000) were next employed in a preliminary step to aid in the determination of the underlying dimensional structure. Both a principal components analysis (PCA) and a maximum likelihood factor analysis (MLFA) were then conducted to examine the suggested recommendations for the underlying dimensional structure.

Exploratory Dimensional Analysis: Results

A one-dimensional solution was judged best for further exploratory analyses based on a convergence of the results of the Scree Test, the MAP, and the PA procedures. An examination of the resulting 9-item loading patterns for both the PCA and the MLFA indicated similar results and supported a one dimensional solution. Eight of the item loadings were rated good (>.55), very good (>.63), or excellent (>.71) in both the PCA and MLFA (Comrey & Lee, 1992). One item with a poor item loading (<.45) was removed from further analyses and a second PCA and MLFA was conducted using only the eight higher loading items. The PCA and MLFA item loadings for the eight retained items from this second analysis are presented in Table 1. The one factor solution accounted for 61% of the variance. The reliability of the resulting 8-item Willingness to Use Microbicides scale, as measured by Cronbach's Coefficient Alpha statistic (Cronbach, 1951), was very high at .91.

Confirmatory Factor Analysis (CFA): Procedure and Results

A CFA using the LISREL 8.7 structural equation modeling program (Jöreskog & Sörbom, 2004) was conducted using the responses of Sub-sample 2 to the eight retained items. Of the potential 266 cases that were randomized to Sub-sample 2 for these analyses, 23 subjects responded to only five of the nine items because of a skip pattern. This resulted in an effective sample size of 243 cases available for the confirmatory analyses after a listwise deletion of items. The initial model (M1) examined specified a one-dimensional solution with all items loading on a single factor and allowing each item error variance to be freely estimated. Item loadings within this model were similar to those found in the exploratory dimensional analyses; however, modification indices given as part of the LISREL output suggested that the overall model fit of M1 could be substantially improved by allowing two error covariances to be freely estimated, and consequently two additional models (M2 and M3) were examined. M2, which allowed the error covariance between items 1 and 2 to be freely estimated, was examined first, and M3, which allowed a second error covariance between items 7 and 8 to be freely estimated, was examined next. Within each model item loadings remained stable, were similar to those found in the EDA, and ranged from .62 to .85 in M1, .64 to .85 in M2, and .60 to .85 in M3. While the overall model based on the Minimum Fit Function Chi-Square statistic did not achieve statistical non-significance, each additional parameter freed did improve overall model fit significantly as judged by the Chi-Square difference test. Several alternative fit indices that are commonly used to evaluate a structural equation model indicated very good fit for M1 with continued improvement for M2 and M3. These indices included the Comparative Fit Index (CFI; Bentler, 1990), the Non-Normed or Tucker-Lewis Index (TLI; Tucker & Lewis, 1973), and the standardized root mean square residual (SRMR; Bentler, 1989). As values range above .95 for the CFI and TLI, in combination with SRMR values of less than .06, model fit is generally considered very good (Hu & Bentler, 1999). Reliability (Cronbach's Coefficient Alpha) for this scale in this sub-sample was also very high and was measured as .91. Complete model fit statistics and indices for the CFA conducted on the sub-sample 2 responses are presented in Table 2a.

Table 2.

Chi-Square, Change in Chi-Square, and Fit Indices for the Different CFA Models in Sub-Sample 2 and the Complete Sample

2a. Sub-sample 2 (n = 243)
Model χ2 df p-value Δχ2 p-value CFI TLI SRMR
M1 144.19 20 <.001 .94 .92 .058
M2 94.31 19 <.001 49.88 <.001 .97 .95 .051
M3
 52.01
 18
 <.001
 42.30
 <.001
 .98
 .98
 .035
2b. Complete sample (N = 490)
M1a 244.05 20 <.001 .95 .93 .057
M2a 154.56 19 <.001 89.49 <.001 .97 .95 .041
M3a 89.32 18 <.001 65.24 <.001 .98 .97 .035
M4 47.17 17 <.001 42.15 <.001 .99 .99 .024
Open in a new tab

Note. χ2 = Minimum Fit Function Chi-Square; Δχ2 = Chi-Square Difference Test value; CFI = Comparative Fit Index; TLI = Tucker-Lewis Index; SRMR = Standardized Root Mean Square Residual.

A final CFA was conducted on the effective complete sample (N = 490) available for analysis. In parallel to the CFA analyses on Sub-sample 2, several models were investigated. In addition to freeing the two previously identified error covariances, the LISREL modification indices suggested that the overall model fit could be further substantially improved by allowing a third error covariance to be freely estimated. In total, four models (M1a, M2a, M3a, and M4) were examined using the data from the complete sample. The last model, M4, allowed a third error covariance between items 3 and 4 to be freely estimated.

In the final CFA analyses, item loadings remained stable, were similar to the previous EDA & CFA analyses, and ranged from .60 to .86 in M1a, .57 to .87 in M2a, .58 to .85 in M3a, and .59 to .86 in M4. While the overall model based on the Minimum Fit Function Chi-Square statistic did not achieve statistical non-significance, each additional parameter freed did improve overall model fit significantly as judged by the Chi-Square difference test. Several alternative indices of model fit (CFI, TLI, & SRMR) also indicated very good fit for M1a with continued improvement for M2a, M3a, and M4. Reliability (Cronbach's Coefficient Alpha) for the scale on the complete sample was also very high and measured .91. Complete model fit statistics and indices for the CFA conducted on the complete sample responses are presented in Table 2b. Table 3 presents and compares the item loadings for model M3 from Sub-sample 2 with those from model M4 from the complete sample.

Table 3.

CFA Item Loadings for Model M3 From Sub-Sample 2 and Model M4 From the Complete Sample

Item M3a M4b
1. If a microbicide had been available the last time you had sex with him, would you have wanted to use one? .74 .76
2. Would you have used a microbicide when you had sex with this partner the last time? .81 .82
3. Would you have had a microbicide available that last time? .72 .67
4. Would you have remembered to use a microbicide when you had sex with this partner the last time? .68 .68
5. If a microbicide cost about as much as a male condom, that is about 75 cents, would you have used it that last time? .85 .82
6. If a microbicide were [preference: available by prescription or available over-the-counterc], would you have used it that LAST TIME? .83 .86
7. If a microbicide were available as a [preferred formulationc] and could be [preference: inserted with an applicator or inserted without an applicatorc], would you use it in the future with [randomly chosen partner]? .69 .66
8. If a microbicide were available as a [preferred formulationc] and could be [preference: inserted with an applicator or inserted without an applicatorc], would you use it in the future with sexual partners who you feel an emotional commitment to? .60 .59
Open in a new tab
a

Based on final CFA model with 2 freely estimated error covariances on subsample 2

b

Based on final CFA model with 3 freely estimated error covariances on whole sample

c

Participant's response to a previous question was automatically inserted by the A-CASI program.

Reliability and Confirmatory Factor Analyses Within Subgroups

In order to investigate whether the overall sample analyses might mask possible racial/ethnic subgroup differences in reliability or factor structure we conducted within group reliability and CFA analyses. Reliability (Cronbach's Coefficient Alpha) for the 8-item Willingness scale remained very high within each of the subgroups and measured .91 for each of the African-American (n = 178) and White (n = 159) subgroups, and measured .89 for the Latina (n = 153) subgroup. We then conducted a CFA that fit the final model M4 with three error covariances freed within each subgroup. The one dimensional structure was supported within each subgroup and item loadings remained stable with all, except one item loading in the Latina subgroup, rated good to excellent (Comrey & Lee, 1992), and ranging from .65 to .84 in the African-American subgroup, .55 to .89 in the White subgroup, and .48 to .84 in the Latina subgroup. Overall model fit based on the Minimum Fit Function Chi-Square statistic achieved statistical non-significance in the Latina, χ2(17, N = 153) = 24.34, p = .11, and White, χ2(17, N = 159) = 21.36, p = .21, subgroups, but was significant in the African-American, χ2(17, N = 178) = 41.52, p < .001, subgroup. The alternative indices of overall model fit (CFI, TLI, & SRMR) all indicated very good fit for the model in each subgroup with the CFI = .98, .99 & .99, the TLI = .97, .99 and .99, and the SRMR = .038, .023, and .039 in the respective African-American, White and Latina subgroups.

Results

Validity Analyses

Overview

After completion of the CFA, concurrent validity analyses were conducted. Twenty-five theoretically relevant variables were examined for their relationship to the resultant 8-item Willingness to Use Microbicides Scale. This scale used a 5-point Likert scale ranging from 1 (definitely not) to 5 (definitely), and had a potential score range of 8−40. Participant scores ranged from 9−40, with an overall mean of 31.5 (SD = 6.9) and a median value of 32. After the initial validity analyses, a Bonferroni adjustment was applied to create a reduced set of significant independent variables that was entered into a standard multiple regression analysis (MRA). The purpose of the MRA was to determine a parsimonious set of theoretically important predictor variables that would be useful for future measurement and intervention studies with microbicides.

Concurrent Validity: Results

Of the 25 variables entered, eleven were not statistically significant (p > .05); five variables were nominally significant (p < .05), but not significant (p = .002 or less) after applying a Bonferroni adjustment; and nine variables were significant (p ≤ .002) after applying a Bonferroni adjustment. The eleven variables not statistically significant included employment status, marital status, number of times moved in the past year, incarceration history, STD history, participant's perception of her own HIV risk, sexual coercion, participant's belief that partner was under the influence of drugs or alcohol during their last sexual encounter, preference for specific form (e.g., gel, suppository, etc) of microbicide available, preference for insertion of microbicide with an applicator or not, and preference for access to microbicide (i.e., by prescription or over-the-counter). The five variables nominally significant before Bonferroni adjustment included history of HIV testing (p = .019), history of vaginal medication use (p = .020), history of douching (p = .027), history of sexual lubricant use (p = .032), and use of pregnancy prevention method in past 3 months (p = .029). The nine variables significant after Bonferroni adjustment included education, F(1, 488) = 12.48, p < .001, household income, F(2, 465) = 7.59, p = .001, race/ethnicity, F(2, 487) = 13.30, p < .001, number of vaginal sex partners, F(1, 488) = 10.03, p = .002, partner type, F(1, 488) = 12.48, p < .001, if the participant had lived with the sexual partner in past 12 months, F(1, 488) = 19.80, p < .001, participant's perception of her partner's behaviors that put him at risk for HIV, F(1, 483) = 17.80, p < .001, frequency of condom use, F(1, 488) = 36.73, p < .001, and history of spermicide use, F(1, 488) = 12.65, p < .001. Table 4 presents the means, standard deviations, and effect sizes of the variables that were significant after the Bonferroni adjustment.

Table 4.

Means, Standard Deviations, Effect Sizes and Post Hoc Differences for Variables Entered Into the Initial Multiple Regression Analysis

Variable N Mean SD Omega-squared
Household income
    Low (<15K)a 263 32.3 6.5 .03
    Medium (15K–36K)a 148 31.2 6.9
    High (36K+)b 57 28.5 7.4
Race/Ethnicity
    Latinaa 153 32.5 6.6 .05
    Blacka 178 32.5 6.3
    Whiteb 159 29.2 7.3
Number of vaginal sex partners
    One 133 29.8 7.2 .02
    Two or more 357 32.1 6.7
Lived with sexual partner past 12 months
    Yes 202 30.1 7.5 .02
    No 288 32.4 6.3
Frequency of condom use
    0 to <50% time 245 29.6 7.3 .07
    50% time or more 245 33.3 6.0
Partner type
    Main 271 30.2 7.1 .04
    Casual/Other 219 33.0 6.3
Partner risk for HIV
    No/Unlikely 224 30.1 7.1 .03
    Yes/Likely 261 32.7 6.4
Education
    High school diploma or less 268 32.4 7.1 .02
    Some college or more 222 30.3 6.5
History of spermicide use
    Yes 143 33.2 6.1 .02
    No 347 30.7 7.1
Open in a new tab

Note. Omega-squared is a population-adjusted measure of systematic variance accounted for. Variable subscripts (a, b) that differ indicate significant differences (p < .05) between levels of a grouping variable with more than 2 levels based on the Tukey HSD post hoc test.

Multiple Regression Analysis

The set of nine Bonferroni-adjusted significant variables was entered into an initial MRA with the score on the Willingness to Use Microbicides Scale as the dependent measure. The nine variables represented grouping variables used in the previous concurrent validity analyses, and two of these variables (race/ethnicity & household income) had three levels, which resulted in eleven variables entered into the MRA using dummy coding methods (Cohen & Cohen, 1983). The resulting MRA with all variables in the model resulted in a Multiple R of .45, F(11, 452) = 10.65, p < .001, a multiple R2 of .21, and an adjusted R2 of .19. Of the variables entered into this MRA, three of the variables (frequency of condom use, White versus other race/ethnicities, and history of spermicide use) contributed significantly to prediction of the Willingness score, while a 4th variable (partner type) was marginally significant. These four variables were entered into a final MRA, and Table 5 presents a summary of this analysis displaying the correlations among the variables, the unstandardized regression coefficients (B) and intercept, the standardized regression coefficients (β), the semipartial correlations (sri2), R2, and adjusted R2. This analysis resulted in a Multiple R of .41, F(4, 485) = 23.94, p < .001, a multiple R2 of .17, and an adjusted R2 of .16. Cohen's f2 statistic (Cohen, 1988), an alternate measure of effect size, was equal to .20, which, according to Cohen, would be considered slightly larger than a medium effect size for a multiple regression analysis. An examination of the unstandardized regression coefficients (B) indicates that women with greater frequency of condom use, a history of spermicide use, and those with casual or other sexual partners have higher predicted scores on the Willingness to Use Microbicides Scale, while White women have their predicted scores reduced.

Table 5.

Standard Multiple Regression of Predictors of Willingness to Use Microbicides With the Correlations Among the Variables, the Unstandardized Regression Coefficients, the Standardized Regression Coefficients, the Semipartial Correlations, R, R2, and Adjusted R2

Variables Willingness (DV) Condom use White vs. other Partner type B β sri2 (unique)
Condom use .27 3.23** .234 .054
White vs. Other −.23 −.02 −3.21** −.218 .048
Partner type .20 .10 −.02 2.27** .163 .026
Spermicide use .16 .07 −.01 .05 2.01* .133 .017
Intercept 29.28
    R = .406
    R2 = .165a
    Adjusted R2 = .158
Open in a new tab
a

Unique variability = .145 (Σsri2); shared variability = .020 (R2 – Σsri2).

*

p < .01.

**

p < .001.

Discussion

Results suggest that person-related (i.e., race/ethnicity), contextual (i.e., partnership status), and product-related (i.e., prevention product history) variables are all associated with willingness to use microbicides, as hypothesized. While the reliability, concurrent validity, and multiple regression analyses suggest that the Willingness to Use Microbicides Scale may be a useful tool for future microbicide research, it should be noted that these results represent the first attempt to measure this construct within microbicide acceptability. Correlated error covariances within the CFAs suggest that there may be further dimensions underlying a possible hierarchical “willingness to use microbicides” construct that should be explored in future studies. For instance, an examination of the specific item content reveals three potential dimensions: willingness within a specific and real sexual context (items 1−2), various aspects of accessibility (items 3−6), and a less contextualized and future-oriented potential willingness (items 7−8). In spite of the imbedded contextualization of preferred formulation and preferred application mode, the less contextualized future-oriented potential willingness items (7 & 8) had the lowest item loadings on the scale. Whether this is a function of a hypothetical future or the lack of partnership context, or both, requires further evaluation.

Once microbicides are approved for use in the U.S. and globally, it will be important for health care providers, health care systems, and researchers to understand who will be likely to use microbicides and under what circumstances. That is, it will be important to understand and account for both the person-related and context-related aspects of a woman's decision to use a microbicide. According to results presented here, race/ethnicity, previous experience with other prevention products, and a woman's perception of whether her partner is “main,” “casual,” or “other” are key variables requiring attention.

Race/ethnicity may reflect a combination of other variables, such as social class, health beliefs and practices, access to health care and prevention, or other cultural morés that impact sexual or health-related behaviors (Brawley & Freeman, 1999; Kaplan & Bennett, 2003; O'Malley, Le, Glaser, Shema, & West, 2003). While we did not control for all of these variables in this analysis, we hypothesize some of the differences that predict willingness to use microbicides likely have their basis in these or other unmeasured variables rather than in any specific biologically-ascribed racial/ethnic factor. Likewise, how much experience with prevention products and whether that experience functions to medicate, prevent disease, and/or prevent conception also may ultimately have important predictive implications. Additionally, how a woman defines or labels her sexual partnership appears to have an impact on her decision-making process with regard to protecting herself from HIV and other STDs. Indeed, much of the condom use literature reflects this same dynamic of the perceived need for less protection in established “main” relationships and a greater need for protection in more “casual” or “other” (e.g., one-night-stand, commercial sex) relationships (Bankole, Darroch, & Singh, 1999; Chan et al., 2004; Polacsek, Celentano, O'Campo, & Santelli, 1999; Rosengard et al., 2005).

The significant predictors from the regression analysis might also be studied in the future as possible moderator or mediator variables in the context of any specific microbicide uptake interventions when they are available. For example, partner type and race/ethnicity could potentially act as moderators depending on the intervention logic, and prevention product experience could be influenced by an intervention and potentially affect microbicide use as a mediator or could be a moderator of willingness to try a microbicide for the first time. Also, there were five additional variables (see Table 4) initially entered into the regression equation that did not enter the final equation. These variables were highly significant with respect to willingness and should also be considered for study with respect to their usefulness as moderators or mediators. In addition, there were five variables that were nominally significant prior to Bonferroni adjustments. Three of these seem to focus on vaginal product use, and two on prevention of either disease or pregnancy. While nominally significant in this sample and with each variable accounting for a small amount of the variance (1%) on the Willingness to Use Microbicides Scale, future studies might consider retaining them for re-analysis in other sexual contexts or populations.

Because microbicides are not currently available for use, measuring “intention to use microbicides” as a construct fell beyond the purview of this study. Once a microbicide is available, it will be incumbent upon the microbicide acceptability field to also develop measures of intention to use microbicides, as Gibbons and colleagues (1998) have found that both “willingness” and “intention” independently predict actual behavior.

The Willingness to Use Microbicides Scale serves as a solid step toward understanding and measuring a woman's decision-making process for protection against STDs and HIV. Future research should allow for the further refinement of the scale. In particular, we hope to clarify whether there are other unrepresented dimensions within a hierarchical construct of willingness to use microbicides that should be measured, or whether we can create a briefer instrument without sacrificing breadth of construct. In addition, while the use of a microbicide will be in itself contextualized by relationship, societal sexual norms, socioeconomic status, and other domains, responding to a scale within a larger assessment instrument is also by necessity contextualized as a function of all the questions that come before it. The Phoenix Project paid particular attention to this contextualization in the ordering of questions and other scales. It will be important for future research to do the same to capitalize on the person-related and sexual context that likely shapes responses.

While a strength of the current study is that the present sample was derived from several sites, results may not be generalizable to women outside the northeast United States, and specifically to women from other racial/ethnic groups and in other countries. This sample contains women from lower socioeconomic strata and with limited education, likely a result of recruitment strategies that relied heavily on community-based organizations and word of mouth (Morrow et al., 2007). It will be imperative for researchers in other sociogeographic arenas to utilize the Willingness to Use Microbicides Scale with a clear understanding that language and sociopolitical culture will need to be accounted for and the scale re-evaluated. Further, it will be important to consider procedural issues so that needs for privacy and ability to negotiate such methods as A-CASI are considered.

References

  1. Abraham CS, Sheeran P, Abrams D, Spears R. Health beliefs and teenage condom use: A prospective study. Psychology and Health. 1996;11(5):641–655. doi: 10.1080/08870449608404994. [DOI] [PubMed] [Google Scholar]
  2. Albarracin D, Johnson BT, Fishbein M, Muellerleile PA. Theories of reasoned action and planned behavior as models of condom use: A meta-analysis. Psychological Bulletin. 2001;127(1):142–161. doi: 10.1037/0033-2909.127.1.142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Alliance for Microbicide Development [February 6, 2006];Alliance for microbicide development. 2004 from http://www.microbicide.org.
  4. Anderson JE, Wilson R, Doll L, Jones TS, Barker P. Condom use and HIV risk behaviors among U.S. adults: Data from a national survey. Family Planning Perspectives. 1999;31(1):24–28. [PubMed] [Google Scholar]
  5. Armitage CJ, Conner M. Efficacy of the Theory of Planned Behaviour: A meta-analytic review. British Journal of Social Psychology. 2001;40(Pt 4):471–499. doi: 10.1348/014466601164939. [DOI] [PubMed] [Google Scholar]
  6. Bandura A. Social foundations of thought and action: A social cognitive theory. Prentice-Hall; Englewood Cliffs, NJ: 1985. [Google Scholar]
  7. Bankole A, Darroch JE, Singh S. Determinants of trends in condom use in the United States, 1988−1995. Family Planning Perspectives. 1999;31(6):264–271. [PubMed] [Google Scholar]
  8. Basen-Engquist K. Psychosocial predictors of “safer sex” behaviors in young adults. AIDS Education and Prevention. 1992;4(2):120–134. [PubMed] [Google Scholar]
  9. Basen-Engquist K, Parcel GS. Attitudes, norms, and self-efficacy: A model of adolescents’ HIV-related sexual risk behavior. Health Education Quarterly. 1992;19(2):263–277. doi: 10.1177/109019819201900209. [DOI] [PubMed] [Google Scholar]
  10. Beksinska ME, Rees VH, McIntyre JA, Wilkinson D. Acceptability of the female condom in different groups of women in South Africa: A multicentred study to inform the national female condom introductory strategy. South African Medical Journal. 2001;91(8):672–678. [PubMed] [Google Scholar]
  11. Bentler PM. EQS: A structural equations program. BMDP Statistical Software; Los Angeles, CA: 1989. [Google Scholar]
  12. Bentler PM. Comparative fit indexes in structural models. Psychological Bulletin. 1990;107(2):238–246. doi: 10.1037/0033-2909.107.2.238. [DOI] [PubMed] [Google Scholar]
  13. Brawley OW, Freeman HP. Race and outcomes: Is this the end of the beginning for minority health research? Journal of the National Cancer Institute. 1999;91(22):1908–1909. doi: 10.1093/jnci/91.22.1908. [DOI] [PubMed] [Google Scholar]
  14. Bryan A, Rocheleau CA, Robbins RN, Hutchinson KE. Condom use among high-risk adolescents: Testing the influence of alcohol use on the relationship of cognitive correlates of behavior. Health Psychology. 2005;24(2):133–142. doi: 10.1037/0278-6133.24.2.133. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Buunk BP, Bakker AB, Siero FW, van den Eijnden RJ, Yzer MC. Predictors of AIDS-preventive behavioral intentions among adult heterosexuals at risk for HIV-infection: Extending current models and measures. AIDS Education and Prevention. 1998;10(2):149–172. [PubMed] [Google Scholar]
  16. Carey MP, Schroder KE. Development and psychometric evaluation of the brief HIV Knowledge Questionnaire. AIDS Education and Prevention. 2002;14(2):172–182. doi: 10.1521/aeap.14.2.172.23902. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Cattell RB. The scree test for the number of factors. Multivariate Behavioral Research. 1966;1:245–276. doi: 10.1207/s15327906mbr0102_10. [DOI] [PubMed] [Google Scholar]
  18. Centers for Disease Control and Prevention [February 6, 2006];HIV/AIDS surveillance report, 2004 (Vol. 16) 2005 from http://www.cdc.gov/hiv/topics/surveillance/resources/reports/index.htm.
  19. Chan DK, Cheung SF, Gray A, Ip A, Lee B. Identifying the psychosocial correlates of condom use by female sex workers in Hong Kong. AIDS Care. 2004;16(4):530–539. doi: 10.1080/09540120410001683439. [DOI] [PubMed] [Google Scholar]
  20. Civic D. The association between characteristics of dating relationships and condom use among heterosexual young adults. AIDS Education and Prevention. 1999;11(4):343–352. [PubMed] [Google Scholar]
  21. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Erlbaum; Hillsdale, NJ: 1988. [Google Scholar]
  22. Cohen J, Cohen P. Applied multiple regression/correlation analysis for the behavioral sciences. Erlbaum; Hillsdale, NJ: 1983. [Google Scholar]
  23. Comrey AL, Lee HB. A first course in factor analysis. 2nd ed. Erlbaum; Hillsdale, NJ: 1992. [Google Scholar]
  24. Cronbach LJ. Coefficient alpha and the internal structure of tests. Psychometrika. 1951;16:297–334. [Google Scholar]
  25. Fishbein M, Ajzen I. [October 25, 2005];Belief, attitude, intention, and behavior: An introduction to theory and research. 1975 from http://www.people.umass.edu/aizen/f&a1975.html.
  26. Fishbein M, Hennessy M, Kamb M, Bolan GA, Hoxworth T, Iatesta M, et al. Using intervention theory to model factors influencing behavior change. Project RESPECT. Evaluation and the Health Professions. 2001;24(4):363–384. doi: 10.1177/01632780122034966. [DOI] [PubMed] [Google Scholar]
  27. Fishbein M, Hennessy M, Yzer M, Douglas J. Can we explain why some people do and some people do not act on their intentions? Psychology, Health and Medicine. 2003;8(1):3–18. doi: 10.1080/1354850021000059223. [DOI] [PubMed] [Google Scholar]
  28. Fisher WA, Fisher JD, Rye BJ. Understanding and promoting AIDS-preventive behavior: Insights from the theory of reasoned action. Health Psychology. 1995;14(3):255–264. doi: 10.1037//0278-6133.14.3.255. [DOI] [PubMed] [Google Scholar]
  29. Forsyth AD, Carey MP. Measuring self-efficacy in the context of HIV risk reduction: Research challenges and recommendations. Health Psychology. 1998;17(6):559–568. doi: 10.1037//0278-6133.17.6.559. [DOI] [PubMed] [Google Scholar]
  30. Gibbons FX, Gerrard M, Blanton H, Russell DW. Reasoned action and social reaction: Willingness and intention as independent predictors of health risk. Journal of Personality and Social Psychology. 1998;74(5):1164–1180. doi: 10.1037//0022-3514.74.5.1164. [DOI] [PubMed] [Google Scholar]
  31. Giles M, Liddell C, Bydawell M. Condom use in African adolescents: The role of individual and group factors. AIDS Care. 2005;17(6):729–739. doi: 10.1080/09540120500038181. [DOI] [PubMed] [Google Scholar]
  32. Global Campaign for Microbicides [February 6, 2006];About microbicides. 2006 January 31; from http://www.global-campaign .org/about_microbicides.htm.
  33. Godin G, Gagnon H, Lambert LD. Factors associated with maintenance of regular condom use among single heterosexual adults: A longitudinal study. Canadian Journal of Public Health. 2003;94(4):287–291. doi: 10.1007/BF03403607. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Grieco EM. [February 6, 2006];The Dominican population in the United States: Growth and distribution. 2004 Sept; from http://www.migrationpolicy.org/pubs.
  35. Guzman B. [February 6, 2006];The Hispanic Population (Issue Brief C2KBR/01−3) 2001 May; from http://www.census.gov/prod/2001pubs/
  36. Hader SL, Smith DK, Moore JS, Holmberg SD. HIV infection in women in the United States: Status at the Millennium. Journal of the American Medical Association. 2001;285(9):1186–1192. doi: 10.1001/jama.285.9.1186. [DOI] [PubMed] [Google Scholar]
  37. Horn JL. A rationale and test for the number of factors in factor analysis. Psychometrika. 1965;30:179–185. doi: 10.1007/BF02289447. [DOI] [PubMed] [Google Scholar]
  38. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling. 1999;6(1):1–55. [Google Scholar]
  39. International Partnership for Microbicides [February 6, 2006];About microbicides. 2005 from http://www.ipm-microbicides.org/microbicides/english/index.htm.
  40. Jaworski BC, Carey MP. Effects of a brief, theory-based STD-prevention program for female college students. Journal of Adolescent Health. 2001;29(6):417–425. doi: 10.1016/s1054-139x(01)00271-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Jennings-Dozier K. Predicting intentions to obtain a Pap smear among African American and Latina women: Testing the theory of planned behavior. Nursing Research. 1999;48(4):198–205. doi: 10.1097/00006199-199907000-00002. [DOI] [PubMed] [Google Scholar]
  42. Jöreskog K, Sörbom D. LISREL 8.7 for Windows. Scientific Software International, Inc.; Lincolnwood IL: 2004. [Google Scholar]
  43. Kaplan JB, Bennett T. Use of race and ethnicity in biomedical publication. Journal of the American Medical Association. 2003;289(20):2709–2716. doi: 10.1001/jama.289.20.2709. [DOI] [PubMed] [Google Scholar]
  44. Minnis AM, Shiboski SC, Padian NS. Barrier contraceptive method acceptability and choice are not reliable indicators of use. Sexually Transmitted Diseases. 2003;30(7):556–561. doi: 10.1097/00007435-200307000-00005. [DOI] [PubMed] [Google Scholar]
  45. Morrow KM, Vargas S, Rosen RK, Christensen AL, Salomon L, Shulman L, et al. The utility of non-proportional quota sampling for recruiting at-risk women for microbicide research. AIDS and Behavior. 2007;II:586–595. doi: 10.1007/s10461-007-9213-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. O'Connor BP. SPSS and SAS programs for determining the number of components using parallel analysis and Velicer's MAP test. Behavior Research Methods, Instruments, and Computers. 2000;32(3):396–402. doi: 10.3758/bf03200807. [DOI] [PubMed] [Google Scholar]
  47. Ogden L, Ogden J, Mthembu P, Williamson N. [July 11, 2005];Impact of HIV on women internationally. 2004 November; doi: 10.3201/eid1011.040624_01. from http://www.cdc.gov/ncidod/EID/vol10no11/04-0624_01.htm. [DOI] [PMC free article] [PubMed]
  48. O'Malley CD, Le GM, Glaser SL, Shema SJ, West DW. Socioeconomic status and breast carcinoma survival in four racial/ethnic groups: A population-based study. Cancer. 2003;97(5):1303–1311. doi: 10.1002/cncr.11160. [DOI] [PubMed] [Google Scholar]
  49. Parker RG, Easton D, Klein CH. Structural barriers and facilitators in HIV prevention: A review of international research. AIDS. 2000;14(Suppl 1):S22–S32. doi: 10.1097/00002030-200006001-00004. [DOI] [PubMed] [Google Scholar]
  50. Polacsek M, Celentano DD, O'Campo P, Santelli J. Correlates of condom use stage of change: Implications for intervention. AIDS Education and Prevention. 1999;11(1):38–52. [PubMed] [Google Scholar]
  51. Posner SF, Bull SS, Ortiz C, Evans T. Factors associated with condom use among young Denver inner city women. Preventive Medicine. 2004;39(6):1227–1233. doi: 10.1016/j.ypmed.2004.04.037. [DOI] [PubMed] [Google Scholar]
  52. Rise J, Thompson M, Verplanken B. Measuring implementation intentions in the context of the theory of planned behavior. Scandinavian Journal of Psychology. 2003;44(2):87–95. doi: 10.1111/1467-9450.00325. [DOI] [PubMed] [Google Scholar]
  53. Roberts ST, Kennedy BL. Why are young college women not using condoms? Their perceived risk, drug use, and developmental vulnerability may provide important clues to sexual risk. Archives of Psychiatric Nursing. 2006;20(1):32–40. doi: 10.1016/j.apnu.2005.08.008. [DOI] [PubMed] [Google Scholar]
  54. Rosengard C, Clarke JG, DaSilva K, Hebert M, Rose J, Stein MD. Correlates of partner-specific condom use intentions among incarcerated women in Rhode Island. Perspectives on Sexual and Reproductive Health. 2005;37(1):32–38. doi: 10.1363/psrh.37.32.05. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Rosenthal D, Fernbach M, Moore S. The singles scene: Safe sex practices and attitudes among at-risk heterosexual adults. Psychology and Health. 1997;12(2):171–182. doi: 10.1080/08870449708407398. [DOI] [PubMed] [Google Scholar]
  56. Sapire KE. The female condom (Femidom): A study of user acceptability. South African Medical Journal. 1995;85(Suppl 10):1081–1084. [PubMed] [Google Scholar]
  57. Severy LJ, Robinson J, Findley-Klein C, McNulty J. Acceptability of a home monitor used to aid in conception: Psychosocial factors and couple dynamics. Contraception. 2006;73(1):65–71. doi: 10.1016/j.contraception.2005.07.008. [DOI] [PubMed] [Google Scholar]
  58. Severy LJ, Tolley E, Woodsong C, Guest G. A framework for examining the sustained acceptability of microbicides. AIDS and Behavior. 2005;9(1):121–131. doi: 10.1007/s10461-005-1687-y. [DOI] [PubMed] [Google Scholar]
  59. Sheeran P, Abraham C. Measurement of condom use in 72 studies of HIV-preventive behaviour: A critical review. Patient Education and Counseling. 1994;24(3):199–216. doi: 10.1016/0738-3991(94)90065-5. [DOI] [PubMed] [Google Scholar]
  60. Sheeran P, Abraham C, Orbell S. Psychosocial correlates of heterosexual condom use: A meta-analysis. Psychological Bulletin. 1999;125(1):90–132. doi: 10.1037/0033-2909.125.1.90. [DOI] [PubMed] [Google Scholar]
  61. Sheeran P, Orbell S. Do intentions predict condom use? Meta-analysis and examination of six moderator variables. British Journal of Social Psychology. 1998;37(Pt 2):231–250. doi: 10.1111/j.2044-8309.1998.tb01167.x. [DOI] [PubMed] [Google Scholar]
  62. Smith KW, McGraw SA, Costa LA, McKinlay JB. A self-efficacy scale for HIV risk behaviors: Development and evaluation. AIDS Education and Prevention. 1996;8(2):97–105. [PubMed] [Google Scholar]
  63. Smith LA. Partner influence on noncondom use: Gender and ethnic differences. Journal of Sex Research. 2003;40(4):346–350. doi: 10.1080/00224490209552200. [DOI] [PubMed] [Google Scholar]
  64. Soet JE, DiIorio C, Dudley WN. Women's self-reported condom use: Intra and interpersonal factors. Women and Health. 1998;27(4):19–32. doi: 10.1300/J013v27n04_02. [DOI] [PubMed] [Google Scholar]
  65. Soler H, Quadagno D, Sly DF, Riehman KS, Eberstein IW, Harrison DF. Relationship dynamics, ethnicity and condom use among low-income women. Family Planning Perspectives. 2000;32(2):82, 88, 101. [PubMed] [Google Scholar]
  66. St Lawrence JS, Eldridge GD, Reitman D, Little CE, Shelby MC, Brasfield TL. Factors influencing condom use among African American women: Implications for risk reduction interventions. American Journal of Community Psychology. 1998;26(1):7–28. doi: 10.1023/a:1021877906707. [DOI] [PubMed] [Google Scholar]
  67. Tucker LR, Lewis C. A reliability coefficient for maximum likelihood factor analysis. Psychometrika. 1973;38:1–10. [Google Scholar]
  68. U.S. Census Bureau [July 9, 2004];Census 2000 summary file 1 (SF 1) 100-Percent Data. 2000 April 1; from http://factfinder.census.gov/servlet/DatasetMainPageServlet?_lang=en.
  69. UNAIDS/WHO [February 6, 2006];AIDS epidemic update: Special report on HIV prevention. 2005 December; from http://www.unaids.org/epi/2005.
  70. Velicer WF. Determining the number of components from the matrix of partial correlations. Psychometrika. 1976;41:321–327. [Google Scholar]
  71. von Haeften I, Fishbein M, Kaspryzk D, Montano D. Acting on one's intentions: Variations in condom use intentions and behaviours as a function of type of partner, gender, ethnicity and risk. Psychology, Health and Medicine. 2000;5(2):163–171. doi: 10.1080/713690182. [DOI] [PubMed] [Google Scholar]
  72. World Health Organization [April, 2005];The female condom: A review. 1997 from http://www.femalehealth.com/download/
  73. Wingood GM, DiClemente RJ. Partner influences and gender-related factors associated with noncondom use among young adult African American women. American Journal of Community Psychology. 1998;26(1):29–51. doi: 10.1023/a:1021830023545. [DOI] [PubMed] [Google Scholar]

RESOURCES