台灣大學生衝動性與問題網路使用之關聯性： 交叉延宕長期追蹤分析

Impulsivity and Problematic Internet Use in Taiwan College Students: A cross-lagged panel analysis

羅孟婷(Meng-Ting Lo)

指導教授 ： 林珊如

國立交通大學/人文社會學院/教育研究所/碩士(2013年)

https://doi.org/10.6842/NCTU.2013.00071

摘要

研究背景與目的
網路的易得性帶給人們許多便利，但過度使用網路則可能對個體身心健康產生負面影響。綜觀過去文獻，心理學家大多根據DSM-IV中(American Psychiatric Association, 1994)物質成癮與病態賭博(為一衝動控制失序症狀)的診斷指標來定義網路成癮(Internet addiction)內涵。然而，「成癮」一詞備受爭議，因而本研究採納問題網路使用(problematic Internet use, PIU)一詞作為研究描述，意為個體無法控制使用網路的衝動，當使用網路時會覺得很愉悅，即使後來過度使用會對自己的身心健康產生負面的影響，也會忽視此影響繼續使用(Shapira, 2000; Shapira et al., 2003)。另外，根據2013年6月發表的DSM-V (American Psychiatric Association, 2013)，問題網路使用(或網路成癮)並未被收錄至任何一個診斷分類中，目前仍需更多研究來探索其在DSM的診斷定位。
根據過去文獻顯示，問題網路使用與衝動性有關，衝動性可能為發展問題網路使用的危險因子(Cao et al., 2007; Lee et al., 2012)，但這些研究大多使用探索性的分析(例如：皮爾森相關及迴歸分析)，無法建立變項之間的時序關係。因此，本研究嘗試運用一種長期研究分析方法-交叉延宕長期追蹤分析(cross-lagged panel analysis)，去探討衝動性與問題網路使用之間的時序關係，並假設衝動性為問題網路使用的前行因素。同時，本研究也將會探討大學生性別、上網時間型態與衝動性之間的關係，以及大學生性別、上網時間型態、背景資訊(如父母關係、壓力、成績等)與問題網路使用之間的關係。

研究樣本與研究工具
本研究樣本為大學生，在兩個時間點(分別在大一下學期與大二上學期測量，測量時間相距六個月)追蹤同一群樣本，在各心理與背景變項上測量兩次。第一個時間點包含827位大學生，第二個時間點包含526位大學生，本研究只取在兩個時間點上皆完成作答的樣本資料進行分析。其中，男生318人，60.6%；女生207人，39.4%。受試者填答背景資訊(包含性別、網路使用經驗、上網時間型態、與父母的關係等)、Problematic Internet Use Scale (PIU scale, Netyouth research team, 2012, 未發表；包含耐受性、退癮症狀、衝動使用、強迫使用、全神貫注及熱切渴望等因素)、Barratt Impulsiveness Scale Version 11 (BIS-11, Barratt et al., 1995)、Beck Depression Inventory-II (BDI-II, Beck, et al., 1996)，以及Unhealthy Lifestyle Scale in Regard to Internet Use (LS)簡短版。

研究結果
藉由執行探索性因素分析、驗證性因素分析，及測量恆等性分析以評估量表的信度與效度。對BIS-11量表進行探索性因素分析，結果發現包含兩個因素，分別為「行動衝動性」與「不善計畫性」，也發現這是一具有信、效度，跨時間及跨性別恆等性的量表。此外，藉由執行驗證性因素分析、測量恆等性分析，以及校標關聯效度分析(問題網路使用與驗證變項之相關: 憂鬱、不健康生活型態)以評估PIU量表之信度與效度，結果發現PIU量表為一具有信、效度，跨時間及跨性別恆等性的量表，且也具同時效度與預測效度。研究主要研究結果說明如下。
一、大學生性別、上網時間型態與衝動性之間的關係：無論在第一或第二個
時間點，男女生在BIS-11分量表上的分數(「行動衝動性」與「不善計畫性」)，皆不具性別差異。且在第一個時間點，凌晨12點到4點使用網路的深夜使用者，其行動衝動性顯著高於非深夜使用者；此外，無論在第一或第二個時間點，深夜使用者與非深夜使用者的不善計畫性分數皆未達顯著差異。
二、大學生性別、上網時間型態、背景資訊與問題網路使用之間的關係：無論在第一或第二個時間點，男女生在PIU分數上皆不具性別差異。此外，無論在第一或第二個時間點，在凌晨12點到4點使用網路的深夜使用者，其PIU分數皆顯著高於非深夜使用者。大學生背景資訊與問題網路使用之間的關係：無論在第一或第二個時間點，PIU分數較高的大學生，在週間與周末皆有較長的網路使用時間，且與父母關係較差，對家庭與人際關係也感到比較多的壓力。此外，在第一個時間點，PIU分數較高的大學生，自評其上學期成績比較低；第二個時間點，PIU分數較高的大學生，對目前的成績表現感到比較困擾，且對目前的學業學習感到比較多的壓力。
三、交叉延宕長期追蹤分析：根據因素間相關分析結果發現，「不善計畫性」與問題網路使用無論在第一或第二個時間點皆無關，因此，本研究將排除此因素不予以進一步分析。而以「行動衝動性」作為交叉延宕模型中的因素。根據交叉延宕分析及模型比較結果發現：同一時間點，行動衝動性與問題網路使用之間的同時相關皆達顯著正相關；在不同時間點，行動衝動性與問題網路使用的自我迴歸係數(半年再測信度)皆達顯著正相關。此外，第一個時間點的行動衝動性顯著預測第二個時間點的問題網路使用；然而，第一個時間點的問題網路使用無法顯著預測第二個時間點的行動衝動性。

討論
從交叉延宕長期追蹤分析的結果顯示，行動衝動性為問題網路使用的前行因素，且可能為大學生發展問題網路使用的主要核心因子。最後，將說明本研究在治療與初期預防之應用，並論及本研究限制與未來研究可行方向。

關鍵字

衝動性 ； 問題網路使用 ； 測量恆等性 ； 交叉延宕長期追蹤分析

並列摘要

Background.
Previous research has suggested that “Problematic Internet use (PIU)” is associated with impulse control disorder in DSM-IV (American Psychiatric Association, 1994) and depicts the problematic behaviors as that an individual cannot control the impulse and growing tension of using the internet, feels pleasure when using it as well as subsequently disregards for major role responsibility and has the negative life outcomes (Shapira, 2000; Shapira et al., 2003). According to DSM-V (American Psychiatric Association, 2013), PIU is not be categorized in any section and its diagnostic classification is still in uncertain. Therefore, studies are still needed to explore the diagnostic classification of PIU in the DSM.
Some resarchers further suggest that impulsivity is a risk factor to develop PIU (Cao et al., 2007; Lee et al., 2012). Though there were studies investigated the relationship between problematic Internet use and impulsivity, the evidences were established mainly by exploratory analysis (e.g., Pearson’s correlation and regression analysis) on cross-sectional data in small samples. Limited longitudinal studies on the relationship between impulsivity and PIU have been conducted with samples of college students.

Objective.
Three purposes of this study were, firstly, to describe gender and online time pattern differences on trait impulsivity among college students. Secondly, to describe gender, online time pattern and background issues of problematic Internet use among college students. Finally, using cross-lagged analytic framework, the aim was to identify the temporal order effects hypothesizing the trait of impulsivity as the precedent factor of PIU.

Sample.
In a panel sample of college students, trait impulsivity and PIU were repeatedly collected in two time points (Time 1, n=827; Time 2, n=526). The final sample comprised 526 college students who completed the measures at Time 1 and Time 2 (318 males and 207 females). Over half of the participants (60.6%) were male.

Method.
Participants completed background information, Problematic Internet Use Scale (PIU scale, Netyouth research team, 2012, unpublished) which consisted of six behavioral addiction symptom components (including tolerance, withdrawal, impulsivity, compulsivity, preoccupation, and craving), Barratt Impulsiveness Scale Version 11 (BIS-11, Barratt et al., 1995), Beck Depression Inventory-II (BDI-II, Beck, et al., 1996), and Unhealthy Lifestyle Scale in Regard to Internet Use (LS) which is a short form selected items from Questionnaire of Lifestyle Change in Regard to Internet Use (LC-PIU) (Yeh, et al., 2012) including different types of unhealthy lifestyles, such as physical/social activities and diet/sleep patterns.

Results.
By conducting exploratory, confirmatory analysis, and measurement invariance, the results of measurement evaluation revealed that the measure of impulsivity (including two factors: motor impulsivity and non-planning) was valid, reliable, stable over time, and the score was equivalent for both genders. Moreover, by conducting confirmatory factor analysis, measurement invariance, and correlation analysis between PIU and the validating variables (depression and unhealthy life styles), the results indicated that problematic Internet use scale was valid, reliable, stable over time, and the score was equivalent for both genders; in addition, the concurrent and predictive validity was acceptable.
The major research findings are listed in below. Firstly, the present study did not find any significant gender differences among male and female college students with regard to motor impulsivity and non-planning at time-points 1 and 2. Moreover, in our sample, midnight users (users who used the Internet from 0:00 a.m. to 4:00 a.m.), reported higher motor impulsivity than non-midnight users at time 1 only.
Secondly, the result of the gender differences analysis suggested that no gender differences with regard to problematic Internet use at times 1 and 2. In addition, midnight users reported more PIU than non-midnight users in both time points. The relation between problematic Internet use and personal backgrounds: at time-points 1 and 2, college students who had higher PIU scores, used the Internet longer from Monday to Friday and on the weekend, had poorer relationships with parents, and experienced greater pressure from family and interpersonal relationships; moreover, at time 1, college students who had higher PIU scores had lower grade in the previous semester; at time 2, college students who had higher PIU scores had experienced greater stress because of their current grades, and more pressure to succeed academically.
Finally, because the sub-factor “non-planning” of impulsivity was not correlated with PIU at times 1 and 2, it was excluded for further analysis. The sub-factor “motor impulsivity” was solely used in the cross-lagged model. Using cross-lagged analysis, the present study found significant positive synchronous correlations between motor impulsivity and PIU. The auto-regressive coefficients of motor impulsivity and PIU were also significant, suggesting that measures of motor impulsivity and PIU were stable across one year. Moreover, the results of cross-lagged effect showed that motor impulsivity at time 1 positively predicted PIU at Time 2, while PIU at Time 1 was not predictive to motor impulsivity at Time 2.

Conclusion.
The result that precedent motor impulsivity was a prospective predictor of PIU indicates that motor impulsivity trait is a core risk condition that might pull college students to develop PIU. The implication section offers suggestions for developing treatment and primary prevention of problematic Internet users who has higher motor impulsivity.

並列關鍵字

impulsivity ； problematic Internet use ； measurement invariance ； cross-lagged analysis

參考文獻

黃柏蒼 (2011)。台灣青少年網路成癮之心理病理因子與性別差異:其與網路使

English Reference

Akaike, H. (1987). Factor analysis and AIC. Psychometrika, 52(3), 317-332.

American Psychiatric Association (1994). Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Publishing, Inc.

American Psychiatric Association (2013). Desk Reference to the Diagnostic Criteria From DSM-5. Washington, DC: American Psychiatric Publishing, Inc.