Appendix II:

Technical Appendix to Chapter 4

To further evaluate whether patterns of intervention provision were associated with gender and HIV status, a series of logistic analyses was performed on the sample of unduplicated clients (N=1,023 males, N=1,074 females). Specifically, each analysis assessed whether the receipt of a particular type of intervention (yes or no) was associated with (1) gender (male versus female); (2) HIV-positive status (known versus unknown); and (3) the interaction between gender and HIV-positive status. Analyses were performed for each of the main categories of intervention (e.g., where is intervention provided, who provides the service, what items are provided, what type of intervention is being received, what topics are being discussed, what referrals are being made) and are shown in the following tables (T4-1 through T4-11) in this technical appendix.

Each row of the table shows descriptive information about the intervention characteristic (e.g., topics discussed: safer sex) including percentages for (1) males and females; (2) HIV-positive clients and HIV status unknown clients; and (3) the four specific combinations of gender and HIV status. In addition, the three inferential tests associated with the gender effect, the HIV status effect, and the gender by HIV status interaction are provided.

Each table presented in this chapter reflects aspects of the interventions the clients received in the context of these adolescent SPNS projects. Tables should be read one row at a time. Each row of the table shows: (1) the percentage of males and females who received an intervention of a particular type; (2) the percentage of HIV unknown and HIV-positive clients who received an intervention of a particular type; and (3) the percentage of males and females broken down by HIV status (HIV status unknown and HIV-positive).

The individual effects shown in each row of the table are chi-squares associated with the logistic regression model predicting whether or not the particular type of intervention was received from (1) gender; (2) HIV status; and (3) the gender by HIV status interaction. The probability values associated with the effects are extremely affected by sample size; a more appropriate index of determining the size of the effect is the R, which shows the associations between the factor (e.g., gender) and the outcome (e.g., whether or not the intervention was received). The direction of the effect also should be noted by looking back at the specific percentages for their factor.

Table T4-1. What Types of Interventions Did Clients Receive? Analyses correspond with Figure 4-2. Percentage of unduplicated individuals who received at least one session of prevention, treatment, or case management (N=1,023 males, N=1,074 females). Logistic Analyses examine the effects of (1) gender; (2) HIV status; and (3) the interaction between gender and HIV status. Percentages for each service provider, for each effect, are shown, as are the inferential effects associated with each of these tests.

Note: HIV Unk. = HIV status unknown
HIV-Pos. = HIV-positive

Table T4-2. Who Provided the Intervention? Analyses correspond with Figure 4-3. Percentage of unduplicated individuals who received services at least once from various types of providers. (N=1,023 males, N=1,074 females). Logistic Analyses examine the effects of (1) gender; (2) HIV status; and (3) the interaction between gender and HIV status. Percentages for each service provider, for each effect, are shown, as are the inferential effects associated with each of these tests.

Note: HIV Unk. = HIV status unknown
HIV-Pos. = HIV-positive

Table T4-3. What Types of Interventions Did Clients Receive? Analyses correspond with Figure 4-4. Percentage of unduplicated individuals who received at least one session of HIV services. (N=1,023 males, N=1,074 females). Logistic Analyses examine the effects of (1) gender; (2) HIV status; and (3) the interaction between gender and HIV status. Percentages for each service provider, for each effect, are shown, as are the inferential effects associated with each of these tests.

Note: HIV Unk. = HIV status unknown
HIV-Pos. = HIV-positive

Table T4-4. What Types of Interventions Did Clients Receive? Analyses correspond with Figure 4-5. Percentage of unduplicated individuals who received at least one session of medical services. (N=1,023 males, N=1,074 females). Logistic Analyses examine the effects of (1) gender; (2) HIV status; and (3) the interaction between gender and HIV status. Percentages for each service provider, for each effect, are shown, as are the inferential effects associated with each of these tests.

Note: HIV Unk. = HIV status unknown
HIV-Pos. = HIV-positive

Table T4-5. What Types of Psychosocial Services Did Clients Receive at Interventions? Analyses correspond with Figure 4-6. Percentage of unduplicated individuals who received at least one session of psychosocial services. (N=1,023 males, N=1,074 females). Logistic analyses examine the effects of (1) gender; (2) HIV status; and (3) the interaction between gender and HIV status. Percentages for each service provider, for each effect, are shown, as are the inferential effects associated with each of these tests.

Note: HIV Unk. = HIV status unknown
HIV-Pos. = HIV-positive

Table T4-6. What Types of Ancillary Support Services Did Clients Receive at Interventions? Analyses correspond with Figure 4-7. Percentage of unduplicated individuals who received at least one session of ancillary support services. (N=1,023 males, N=1,074 females). Logistic analyses examine the effects of (1) gender; (2) HIV status; and (3) the interaction between gender and HIV status. Percentages for each service provider, for each effect, are shown, as are the inferential effects associated with each of these tests.

Note: HIV Unk. = HIV status unknown
HIV-Pos. = HIV-positive

Table T4-7. What Topics Did Clients Discuss During Interventions? Analyses correspond with Figure 4-8a. Percentage of unduplicated individuals who discussed topics in at least one intervention, Part 1. (N=1,023 males, N=1,074 females). Logistic analyses examine the effects of (1) gender; (2) HIV status; and (3) the interaction between gender and HIV status. Percentages for each service provider, for each effect, are shown, as are the inferential effects associated with each of these tests.

Note: HIV Unk. = HIV status unknown
HIV-Pos. = HIV-positive

Table T4-8. What Topics Were Discussed at Interventions? Analyses correspond with Figure 4-8b. Percentage of unduplicated individuals who discussed topics in at least one intervention, Part 2. (N=1,023 males, N=1,074 females). Logistic Analyses examine the effects of (1) gender; (2) HIV status; and (3) the interaction between gender and HIV status. Percentages for each service provider, for each effect, are shown, as are the inferential effects associated with each of these tests.

Note: HIV Unk. = HIV status unknown
HIV-Pos. = HIV-positive

Table T4-9. What Types of Items Did Clients Receive During Interventions? Analyses correspond with Figure 4-9. Percentage of unduplicated individuals who received items in at least one intervention session (N=1,023 males, N=1,074 females). Logistic analyses examine the effects of (1) gender; (2) HIV status; and (3) the interaction between gender and HIV status. Percentages for each service provider, for each effect, are shown, as are the inferential effects associated with each of these tests.

Note: HIV Unk. = HIV status unknown
HIV-Pos. = HIV-positive

Table T4-10. For What Services Were Clients Referred? Analyses correspond with Figure 4-10. Percentage of referrals made for unduplicated individuals who received at least one intervention session. (N=1,023 males, N=1,074 females). Logistic Analyses examine the effects of (1) gender; (2) HIV status; and (3) the interaction between gender and HIV status. Percentages for each service provider, for each effect, are shown, as are the inferential effects associated with each of these tests.

Back to Table of Contents

© Copyright 1999-2005 by The Measurement Group LLC. All rights reserved.