[反思] 20191009 關於研究題材的整理

先前老師要我整理溝通技巧訓練介入措施。依據我過去經驗，我先規劃了統整表格(仿造meta-analysis模式)

然而這樣的表格模式，就如老師所說，僅是文字的剪貼並無實質參考意義。

因此，踏入了不斷修正表格格式的循環中….

坦白說，過程中真的很挫敗，因為任何研究主題都沒有所謂「通用」或「唯一」的理論或概念

因此每閱讀一篇新的文章，就有一個新的理論與概念定義

在缺乏思考力及批判力的狀況下，就覺得每個研究都有自己的獨特點

所以什麼都想要記錄下來，紀錄了一堆的結果，就是反而更抓不到自己要的重點為何

自己亂了慌了也就算了，更拖延了進度….

這兩周開始重新調整步調，先找出具有專家「共識」的理論及概念開始了解

並嘗試透過將理論/概念視覺圖像化，讓自己的資訊更清楚且明確

過程中再慢慢加入自己的想法與批判，去調整原先的理論架構

接下來要做的就是將過程好好做紀錄(寫下)，並與專家好好切磋吧………………

[討論] 20191008 寫作紀錄

***寫作討論修改並非只是將文字做重述，而是要先釐清自己的問題(概念釐清、表達/寫作能力)所在***

本次寫作討論段落[前言]：

Verification of reliability can determine that the change score is a real change in individuals rather than a measurement error. That is, a lack of reliability verification can affect the accuracy of the interpretation of results (i.e., overestimation or underestimation). Test-retest reliability refers to the extent of consistency between two assessments of the same construct administered with the same individuals at two different points in time. If the individuals’ traits to be assessed have stablized between the two assessments, then the measure is reliable. Random measurement error is due to chance factors that affect an observed score for an inevitable reason, that is, the magnitude of the uncertainty around an observed score. Therefore, random measurement error needs to be addressed as this may affect the reliability measure. The amount of random measurement error of a measure can be estimated by calculating the standard measurement error (SEM) and minimal detectable change (MDC). The SEM is a measure of the degree to which measured test scores are spread around a true score. The MDC is the minimal amount of change between repeated assessments that indicate a true change beyond the random measurement error at a certain confidence level (usually 95%).

此為對概念陳述的問題：

[黃底] 這並非信度的「普遍」定義，因此置於前言的段落開頭太過突兀，要記得投稿文章的主要讀者為reviewer!!!  老師協助修改成：”Reliability generally includes test-retest reliability and the amount of random measurement error.”

• Note. 這樣的修改方式更為簡潔，也同時增加該段落概念(test-retest reliablilty and random measurement error)的鋪成/連貫性。

[藍底] 對test-retest reliability的描述無法凸顯其「對研究的意義與價值」，僅為test-retest reliability的方法陳述。

[綠底] 可將此內容移置方法學章節，讓該段落表達之重點更加明確且簡潔。

此為對寫作/英文撰寫的問題：

 

[反思] 20191007 關於研究主題的發想與建構

關於一個新的研究主題發想與建構，宜先了解這個研究主題是否已有相關理論基礎及明確的概念定義，然而一個概念可能依據不同理論背景而有不同定義。

在個人的近期經驗累積下，認為可藉由四大方向去做著手：

1.   找出該研究主題的「專家」做為參考依據：

參考其過去著作(書籍或文章)，以取得在該理論/概念較有「共識」的核心基礎；反之若無，則代表此理論/概念仍備受爭議。

2.   將理論/概念統整，以視覺圖像化方式呈現：

透過圖表方式(視覺化)將過去理論/概念的核心意涵與關係加以呈現。針對統整方式，並非將理論/概念做剪貼而已，應試著用自己的語言陳述對該理論/概念的理解，這才能真正內化成屬於自己的研究核心。

統整過程需掌握以下原則：

A.釐清理論/概念間是否有層次或分類(掌握top-down原則)

B.概念及各概念間需有明確且具體的文字說明

C.盡可能呈現圖文並茂

3.   突顯自己研究獨有的核心與價值：

讓自己立足在先前研究的基礎上，而非僅跟隨他人腳步。過去自己在統整多個研究文章時，多採用標準表格，如填入Citation, Participants, Intervention, Outcome等內容。然而，對於新的研究主題建立而言，我們會有興趣知道或有需要去知道的，應是各研究間所使用的理論/概念、介入模式的內容程序及介入療效大小與時間長短。因此，應掌握以上重點去做整理，對未來研究的基礎才有實質助益。

4.   釐清自己的疑惑與困難：

回應一個問題前，需先找出此問題的「核心」所在。如先自我釐清此問題是對「概念」不瞭解，還是「表達」(文字陳述或英文能力)不明確的問題。真正的釐清問題核心，才能真正有效的解決此問題。

藉由以上的過程記錄，期許自己能夠早日發展並規劃屬於自己的研究生涯。

[討論] 20191003 討論議題

【自覺記憶功能評估工具之回顧與評論】

關於區辨效度，老師建議結果宜採「等級」方式呈現，但有幾點疑惑：

1.    若要將區辨效度分等級，想到的辦法去利用該研究兩組的mean和SD去計算Effect size，但算有客觀數據(如ES)，但其區辨驗證對象不同(有Low memory, MCI, and Dementia三者)，這樣缺乏客觀的可比較性，也就是區辨對象不同時，無法證實A研究(健康/Dementia)就是優於B研究(健康/MCI)，。

2.    針對等級的劃分，老師先前給的建議是「看數據(mean值)，差異大代表良好，若僅統計顯著，那代表可接受。」。但以下四篇研究所提供數據不一致(部分研究沒有提供mean值)，故各評估工具間無法統一用「等級」(良好/可接受)去比較，且缺乏可引用作為判斷準則之參考文獻。

3.    目前想到的解決辦法就是，有顯著則呈現”acceptable”，若不顯著則呈現”unacceptable”，不知道是否適切? (修改成黃色背景欄位)

各評估工具之區辨效度統整表格如下：

評估工具	EMQ-20 items	EMQ-11 items	MMQ-Original	SMCQ-Original
區辨效度（原）	Normal (n=102) / Low memory (n=10) p<0.001	Normal (n=766) / MCI (n=78) p<0.001	Normal (n=20) / Low memory (n=12) p=0.003	Normal (n=1464) / Dementia (n=187) p<0.001
區辨效度（修）	Normal / Low memory: acceptable	Normal / MCI: acceptable	Normal / Low memory: acceptable	Normal / Dementia: acceptable
該文章區辨效度驗證結果的相關資訊	Ÿ Scores on the 20-item EMQ were lower in the Normal Memory (M=44.6±16.7) than in the Low Performance group (M=75.1±24.4) [t(28)=3.71, p<0.001].	Ÿ FII (U = 17,175; p < 0.001), EF (U = 17,651; p < 0.001) and PM (U = 19,015; p < 0.001) were found to differentiate between both groups.	Ÿ Scores on MMQ-Contentment were lower in the low-memory group (M = 30.7) than the normal-memory group (M = 43.1), t(30) = 3.10, p = .004. Ÿ Similarly, scores on MMQ-Ability were lower in the low memory group (M = 37.3) than the normal-memory group (M = 46.2), t(30) = 3.19, p = .003.	Ÿ The SMCQ-T, SMCQ-G and SMCQ-E scores were all significantly higher in the dementia patients than in the elderly subjects without dementia after adjustment for the influence of age, sex, education and depressive symptoms (F = 55.4, d.f. = 1,645, p < 0.001), indicating that the SMCQ had discriminant properties for dementia ( table 3 ).

[紀錄] 201901002 寫作紀錄

此為上周五(09/27)寫作討論段落：

The SEM of the 3 MMQ scales ranged from 4.8 to 6.6, which are similar¹ to that in a previous study (SEM = 3.5 to 4.4). These findings suggest² that the scores of the MMQ are affected by random measurement errors and thus, the scores of the MMQ appear to fluctuate through the levels of individuals' memory function remained unchanged.³ The random measurement error of the MMQ⁴ may result from the interval between two assessments. Specifically, because the 3 months was relatively large for the examination of test-retest reliability of self-perceived memory function (e.g., 4 weeks-interval), which may yield individual variability in their memory function and thus, lead to larger SEM⁵ values. Therefore,⁶ the MMQ may be unstable for repeated assessments over 3 months.

¹ 兩研究SEM數據沒有overlap的地方，不適合用similar，且未說明time interval差異，無法呼應可能原因的解釋。

² findings suggest在語意上是指「本研究發現…」，但本來每個測驗都會受到隨機測量誤差的影響，非本研究才證實這點。因此使用詞彙不恰當。

³ 於討論段落中不需要在重複說明SEM之定義（綠底）。

⁴ 使用The random measurement error of the MMQ 為該句子之主詞不易讓讀者理解，宜改寫成the SEM of the MMQ。

⁵ 句子中突然出現test-retest reliability，這不易讓不瞭解心理計量特性的讀者理解。

⁶  larger SEM直接寫出此結果為高估或低估，句子會更為明確。

⁷ 結語句宜呼應段落主旨(SEM)之概念，目前寫法比較像是再測信度的概念。

-------------------------------------------------------------------------------------------------------------------------------------------

-------------------------------------------------------------------------------------------------------------------------------------------

修正後段落：

The SEM of the 3 MMQ scales (4.8 to 6.6 at 3 months) in our study was slightly greater than that of a previous study (SEM= 3.5 to 4.4 at 1 month). These findings indicate that the scores of the MMQ are affected by random measurement errors and thus, the scores of the MMQ appear to fluctuate through the levels of individuals' memory function remained unchanged. A possible reason for the SEM of the MMQ is caused by the time interval between two assessments (i.e., 3 months) was relatively long compared to previous studies (i.e., 1 month). Generally, the longer is the time gap, the greater is the chance that the two assessments may change between two assessments (due to random error) and thus, lead to the overestimation of the SEM values. Therefore, our results revealed that the MMQ may be unstable for two assessments over 3 months due to random measurement errors.

[回應上述第1點] 修正為slightly greater並加上兩研究的time interval，以呼應後續可能原因之解釋。

[回應上述第2點] 修正為findings indicate。

[回應上述第3點] 詳見[問題2]

[回應上述第4點] 修正主詞為the SEM of the MMQ

[回應上述第5點] 修正原先語句之表達方式(黃底)，避免模糊字詞如原先的test-retest reliability

[回應上述第6點] 修正overestimation of the SEM

[回應上述第7點] 修正結語句，待與老師澄清是否有呼應SEM之概念。

[問題1] 若考量樣本依賴性的問題，本研究結果(SMC老人)適合去和先前研究(健康老人)比較嗎??

[問題2] 若討論段落中不需要在重複說明SEM之定義，那僅在前言中提到SEM概念 ”The SEM is a measure of the degree to which measured test scores are spread around a true score.”，對讀者而言會不會不易了解?? 還是應該放在哪個章節段落說明較為適切?