結論與資料摘要
兩個時點的資料皆完整，反向計分後的全量表內部一致性良好：前測 Cronbach’s α = .927，後測 α = .893。四個 SIMS 構面的 α 皆達可接受以上水準，可作為兩個調查時點的動機描述依據。
前測有效作答39
後測有效作答41
各時點遺漏作答0
可配對識別碼35
解讀界線：前、後測分別有 39 與 41 份作答，只有 35 筆可依識別碼配對。跨時點的結果僅是描述性比較，尚未進行配對統計檢定；因此不據此宣稱教學效果。
資料品質與量表設定
資料完整性良好：兩份工作簿皆含 16 題作答欄位；有效作答中的 16 題皆無遺漏值，且各題皆有作答差異。
- 題目方向：使用
答題記錄(完整選項內容) 工作表前置的 1–7 分數判讀 Likert 方向。 - 構面編碼：內在動機（Q1、Q5、Q9、Q13）、認同調節（Q2、Q6、Q10、Q14）、外在調節（Q3、Q7、Q11、Q15）、無動機（Q4、Q8、Q12、Q16）。
- 整體信度：只在全量表 α 的計算中，將 Q4、Q8、Q12、Q16 反向計分；無動機平均數保留原始方向，分數愈高代表無動機愈高。
信度結果
| 分析範圍 | 前測 α | 後測 α | 判讀 |
|---|
| 全量表原始分數（16 題） | .748 | .797 | 混合正向與無動機題，只用於說明反向計分的必要性 |
| 全量表（無動機題反向） | .927 | .893 | 良好至極佳 |
| 內在動機 | .894 | .896 | 良好 |
| 認同調節 | .866 | .907 | 良好至極佳 |
| 外在調節 | .879 | .838 | 良好 |
| 無動機 | .908 | .779 | 前測極佳；後測可接受 |
反向後校正題目－總分相關：前測 .464–.830；後測 .181–.784。後測有較低的題目－總分相關，後續若累積更多樣本，可優先檢視該題在不同情境中的理解情形；本次資料不足以單憑此結果刪題。
構面平均數（7 分量尺）
描述性比較無動機維持原始方向；其餘構面分數愈高，代表該類動機愈高。
| 構面 | 前測平均數 | 後測平均數 | 差異（後－前） | 描述 |
|---|
| 內在動機 | 5.21 | 5.60 | +0.39 | 後測較高 |
| 認同調節 | 5.97 | 6.08 | +0.11 | 後測略高 |
| 外在調節 | 5.28 | 5.35 | +0.07 | 接近 |
| 無動機 | 3.42 | 3.11 | −0.31 | 後測較低 |
效度與使用限制
- 內容對應：題目依 SIMS 四構面編碼分組，可支持量表內容與理論架構的對應說明。
- 建構效度限制：兩個時點的完整樣本數都低於 16 題的五倍門檻（80），不將 KMO、Bartlett、EFA 或 CFA 視為穩定的建構效度證據。
- 比較建議：若要檢驗前後差異，應以 35 筆已配對的識別碼建立配對資料，再預先決定分析方式與報告效果量。
- 效標關聯效度：未提供外部效標或其他量表資料，本報告未檢定此項證據。
Summary and sample
Both time points are complete. After reverse scoring, overall internal consistency is good: Cronbach’s α = .927 at pretest and .893 at posttest. All four SIMS subscales meet at least an acceptable reliability level for describing motivation at each survey time point.
Valid pretest responses39
Valid posttest responses41
Missing item responses0
Matched identifiers35
Interpretation boundary: The two files contain 39 and 41 responses respectively, with only 35 matched identifiers. Cross-time results here are descriptive only; no paired statistical test was run, so this report does not claim a teaching effect.
Data quality and scoring
Good completeness: Both workbooks have 16 response-item columns. There are no missing item responses among valid records, and every item has response variation.
- Score direction: Likert direction was read from the leading 1–7 score in the full-text response-record worksheet.
- Construct mapping: Intrinsic motivation (Q1, Q5, Q9, Q13); identified regulation (Q2, Q6, Q10, Q14); external regulation (Q3, Q7, Q11, Q15); amotivation (Q4, Q8, Q12, Q16).
- Overall reliability: Q4, Q8, Q12, and Q16 were reverse-scored only for the 16-item overall alpha. Amotivation means retain their original direction: a higher score means higher amotivation.
Reliability results
| Scope | Pretest α | Posttest α | Interpretation |
|---|
| Raw overall score (16 items) | .748 | .797 | Mixed positive-direction and amotivation items; shown only to explain reverse scoring |
| Overall score (amotivation items reversed) | .927 | .893 | Good to excellent |
| Intrinsic motivation | .894 | .896 | Good |
| Identified regulation | .866 | .907 | Good to excellent |
| External regulation | .879 | .838 | Good |
| Amotivation | .908 | .779 | Excellent at pretest; acceptable at posttest |
Reverse-scored corrected item-total correlations ranged from .464–.830 at pretest and .181–.784 at posttest. The lower posttest correlation can be reviewed after more data are accumulated; it is not sufficient grounds to delete an item now.
Subscale means (7-point scale)
Descriptive comparisonAmotivation remains in its original direction; higher scores on the other subscales indicate more of that type of motivation.
| Subscale | Pretest mean | Posttest mean | Change (post − pre) | Description |
|---|
| Intrinsic motivation | 5.21 | 5.60 | +0.39 | Higher at posttest |
| Identified regulation | 5.97 | 6.08 | +0.11 | Slightly higher at posttest |
| External regulation | 5.28 | 5.35 | +0.07 | Similar |
| Amotivation | 3.42 | 3.11 | −0.31 | Lower at posttest |
Validity and limits
- Content alignment: Grouping follows the four SIMS constructs, supporting a content-to-theory alignment statement.
- Construct-validity limit: Each complete sample is below five times the 16-item count (80). KMO, Bartlett, EFA, and CFA are not reported as stable construct-validity evidence.
- Comparison recommendation: Use the 35 matched identifiers to create a paired dataset, then pre-specify the statistical test and effect-size reporting before testing pre/post change.
- Criterion validity: No external criterion or additional instrument was supplied, so this evidence was not tested.