Data Analysis

# **Data Analysis Framework**

_________________________________________________________________________________________________
> **🎯 Goal:** Develop a rigorous data analysis plan that bridges research questions with specific statistical or thematic tools, ensuring valid and meaningful results.
_________________________________________________________________________________________________

#### **1. Original Prompt Block (English)**

> "Develop a comprehensive data analysis plan for my research on [topic] using [type of data]. Include: 1) recommended statistical or analytical methods appropriate for addressing my research questions, 2) step-by-step analysis procedures, 3) software tools best suited for this analysis with specific features to utilize, 4) approaches for handling missing data and outliers, and 5) visualization techniques to effectively present the findings. My specific research questions are [list research questions]."

#### **2. Analysis & Strategy**

| Phase | Insight |
| :--- | :--- |
| **Analysis** | This prompt tackles the intimidating task of converting raw data into meaning. It bridges the gap between research questions and mathematical/thematic tools. |
| **Review** | The original prompt is solid but dangerous if data types (nominal, ordinal, interval) are unknown. AI might suggest advanced tests (Regression) without checking assumptions (Normality). |
| **Brainstorming** | Create a **"Data Analysis Matrix"** mapping RQs to Variables and Tests. Insist on a **"Pre-processing"** phase (cleaning data) where most mistakes happen. |

#### **3. PRO / CONTRA Analysis**

| **PRO (Advantages)** | **CONTRA (Risks)** |
| :--- | :--- |
| ✅ **Software Guidance:** Helps choose the right tool (SPSS vs. R vs. NVivo) based on complexity. | ⚠️ **Assumption Blindness:** Doesn't explicitly ask to check assumptions (e.g., normality). If violated, results are invalid. |
| ✅ **Visualization:** Asking for specific charts ensures the final report looks professional. | ⚠️ **Qualitative Vague:** For text data, "analysis" is vague. Needs to specify Thematic, Content, or Grounded Theory. |

#### **4. Upgraded Prompt**

**Changes Made:** Introduced a "Mapping Matrix" for clarity, added a specific "Assumptions & Cleaning" section, and forced a distinction between Quantitative and Qualitative logic.

> **Copy & Paste this Prompt:**
>
> "Act as a Senior Data Analyst. Create a rigorous Data Analysis Plan for my study on [topic]. My data is [Quantitative/Qualitative/Mixed].
>
> **Step 1: Data Preparation & Cleaning**
> *   Define how to handle missing data (e.g., Imputation vs. Listwise deletion).
> *   Define how to identify and treat outliers.
> *   *(If Quant):* List specific **Assumptions** (e.g., Normality, Homogeneity) that must be checked before testing.
>
> **Step 2: The Analysis Matrix (Create a Table)**
> *   **Column 1:** Research Question (RQ)
> *   **Column 2:** Variables Involved (IV & DV) & Data Type (Nominal/Ordinal/Scale)
> *   **Column 3:** Specific Statistical Test / Analysis Method (e.g., ANOVA, Pearson Correlation, Inductive Coding)
> *   **Column 4:** Software Command/Procedure (e.g., SPSS: Analyze > Compare Means...)
>
> **Step 3: Visualization & Reporting**
> *   Suggest the best chart/graph type for each RQ (e.g., Boxplot for group comparison).
> *   Provide a template sentence for reporting the result in **APA style**."

#### **5. Hungarian Summary & Best Usage**

| Kategória | Leírás |
| :--- | :--- |
| **Mire a legjobb?**<br>*(Best Usage)* | Ez a prompt az **Eredmények (Results) fejezet** megírásához és magához az elemzéshez elengedhetetlen. Előre megtervezi a matematikai/logikai útvonalat, így biztosan azt mérjük, amit a hipotézisben ígértünk. |
| **Valós példa**<br>*(Real Life Example)* | Egy szociológia hallgató, aki a *"Jövedelmi különbségek hatását vizsgálja az egészségügyi kiadásokra"* (Income vs Health Spending). |
| **A probléma** | Van egy Excel táblája 500 emberrel, de nem tudja, hogy korrelációt vagy regressziót számoljon, és mit kezdjen a hiányzó adatokkal. |
| **A megoldás** | A prompt **"Data Preparation"** része megmondja, hogy a hiányzó adatokat átlaggal pótolja. A **"Matrix"** rész megmutatja, hogy mivel mindkét változó skála típusú, **Lineáris Regressziót** kell használnia, amit egy **Scatter Plot** (pontdiagram) segítségével ábrázoljon. |

_________________________________________________________________________________________________