有性活動青少女避孕行為因果模式的建立

A Causal Model of Contraceptive Behavior among Sexually Active Female Adolescents

王瑞霞(Ruey-Hsia Wang)

指導教授 ： 王秀紅

高雄醫學大學/護理學院/護理學系研究所/博士(2005年)

摘要

摘要 ( Abstract )
本研究目的為建立有性活動青少女避孕行為因果模式，主要以社會學習理論（social learning theory）為架構，納入過去行為經驗、社會情境（social situation）、自我效能（self-efficacy）、決定衡量（decisional balance）及權力（power）等概念形成本研究的理論架構，探討避孕經驗、社會情境、避孕自我效能、避孕決定衡量及權力，對台灣有性活動青少女避孕行為的直接及間接作用，並建構一個因果模式。
本研究為橫斷式研究，以最近三個月有性行為，且年滿15至19歲之高職青少女為研究對象。以方便取樣，選取高雄市三所願意參與研究的公私立高職學校，依照各校所提供一至三年級班級數及青少女總人數，依比率採隨機的班級集束抽樣，並在研究進行中，以能抽樣到較多最近三個月有性行為青少女為原則，調整由各校各年級抽出的參與班級。對抽中班級中的青少女，以無記名自填式問卷調查收集資料，並篩選其中最近三個月有性行為青少女的資料進行分析。問卷內容包括人口學資料、與異性朋友交往情形、避孕經驗及避孕方法、個人內在權力、關係權力、性行為頻率、避孕行為、避孕決定衡量、避孕自我效能、自覺同儕避孕行為、自覺重要他人支持及自覺男朋友支持等量表。量表擬定後，先經由專家效度、內部一致性及二週再測信度的檢定。統計方法除包括率、百分比、平均值、標準差及皮爾森積差相關外，並以驗證性因素分析（confirmatory factor analysis）檢定各量表的建構效度及刪除不具效度之題目，最後再以EQS 6.0版軟體，進行結構方程模式的檢定，以了解收集的資料是否支持假設的「有性活動青少女避孕行為因果模式」。
本研究共抽樣102個班級，有2,325位青少女完成問卷填寫，其中426位最近三個月有性行為，佔所有填寫問卷青少女的18.32%，扣除回答不完整者，共得有效問卷404份，有效問卷率為94.84%。由結構方程模式的檢定，顯示本研究原始假設的「有性活動青少女避孕行為因果模式」，未能獲得實際資料的支持。 經由理論修飾獲得一個資料支持且簡效的模式，此模式雖然χ2值為82.082 ( df =30 ), p <.01，但GFI=.960, SRMR=.039, RMSEA=.066, NFI=.949, NNFI=.950, CFI=.967，顯示此模式是適當的。由模式中可知，「避孕經驗」對「避孕自我效能」( β=.114, p <.01) 及「避孕行為」 ( β=.220, p <.01) 皆有顯著的直接效果，但「避孕經驗」對「避孕行為」( β=.036, p >.01) 的間接效果並不顯著。「社會情境」對「避孕自我效能」 ( β=.869, p <.01 ) 及「避孕決定衡量」 ( β=.913, p <.01) 皆有顯著的直接效果，而「社會情境」對「避孕行為」 為顯著的間接效果 ( β=.595, p <.01)。「避孕自我效能」( β=.317, p <.01）及「避孕決定衡量」( β=.350, p <.01) 對「避孕行為」皆有顯著的直接效果，模式中所有的變項能解釋55.3%避孕行為的變異量。

本研究建議未來仍須以不同的樣本檢測，以了解本研究「有性活動青少女避孕行為因果模式」在其它樣本的適用性；而建構不同的有性活動青少女避孕行為因果模式，也是未來應努力的方向。本研究結果，將有助於擬定青少女避孕行為介入計畫的參考，並有助於建構青少女避孕行為的知識。

關鍵字

結構方程模式 ； 青少女 ； 避孕行為

並列摘要

The purpose of this study was to model contraceptive behavior of sexually active female adolescents. Our hypothesized theory was adapted from a Social Learning Theory, which combines the concept of past experience of contraception, social situation, self-efficacy for contraception, decisional balance for contraception, power and contraceptive behavior.
We conducted a cross-sectional survey on female adolescents who were 15 to 19 years old and had sexual intercourse in the past 3 months. By convenient sampling, three vocational high schools in Kaohsiung City were selected. A total of 102 classes of 10th to 12th grade girls were then selected by cluster sampling. Among 2,325 female adolescents who completed the questionnaire, 426 (18.32%) had sexual intercourse in the past 3 months. Excluding the questionnaires with incomplete answer, a total of 404 female adolescents, with response rate was 94.84%, were comprised in our study samples.
The scales in the questionnaire included demographic data, dating with heterosexual friends, frequency of sexual behavior, past experience of contraception and contraceptive method used, self-efficacy for contraception scale, decisional balance for contraception scale, perception of support from significant others scale, perception of peers’ contraceptive behavior scale, perception of support from sexual partners scale, personal power scale, relationship power scale and contraceptive behavior scale. All scales were examined for content validity and for test-retest reliability on 11 samples. Internal consistency was examined on 404 samples. Confirmatory factor analysis was used to test the validity of each scale. Some invalid items were deleted according to the results.
The hypothesized causal model was tested by structural equation modeling using EQS 6.0 software. Maximum Likelihood method was used to test the fit of the model. The original model was refined and yield a most parsimonious model suggested that: past experience of contraception had significant direct effect on self-efficacy for contraception (β=.114, p <.01) and contraceptive behavior (β=.220, p <.01), however; past experience of contraception had not significant indirect effect on contraceptive behavior ( β=.036, p >.01). Social situation had significant direct effect on self-efficacy for contraception (β=.869, p <.01) and on decisional balance for contraception (β=.913, p <.01). Social situation had significant indirect effect on contraceptive behavior (β=.595, p <.01). Self-efficacy for contraception (β=.317, p <.01) and decisional balance for contraception (β=.350, p <.01) had significant direct effect on contraceptive behavior. The goodness-of-fit indices indicated the model best fit and the construct included in the model accounted for 55.3% variance of contraceptive behavior.
The findings of this study provide a reference of intervention program for sexually active female adolescents and are useful to construct the knowledge of contraceptive behavior in the female adolescent population. Further replicate studies in different samples need to be conducted to validate our model.

並列關鍵字

structural equation modeling ； contraceptive behavior ； female adolescents

參考文獻

內政部 (無日期)•內政部統計通報•2003年 5月13日取自http://www.moi.gov.tw/w3/stat/home.asp

王瑞琪 (1992)•台北市高職三年級學生避孕行為及其相關因素探討•未發表的碩士論文，台北：師大衛生教育研究所。

王瑞霞、王秀紅、許秀月 (2003)•青少女避孕行為因果模式建立及激能教育介入效果之探討•國科會專題研究成果報告 (NSC91-2314-B-037-301)。

王瑞霞、王秀紅、許敏桃 (1999)•女性青少年懷孕及避孕行為因素的探討•國科會專題研究成果報告 (NSC88-2314-B037-005)。

王瑞霞、王秀紅、許敏桃 (2002)•青少女母親產後避孕行為及影響因素之探討•國科會專題研究成果報告 (NSC89-2314-B-037-168)。