according to Spearman Brown’s formula, 56 adolescents were excluded because they had incom plete data. what does spearman browns formula means?
Searching the Web
Spearman Brown formula reliability psychometrics explanation
The Spearman-Brown Formula - Explained
The Spearman-Brown Prophecy Formula (also called the Spearman-Brown Correction) is a psychometric formula used to estimate the reliability of a test based on its length. It is commonly used in research involving questionnaires, psychological scales, and assessments.
What It Does
It works on a core principle: longer tests are more reliable than shorter ones (assuming similar-quality items). The formula predicts how much the reliability will increase or decrease if you add or remove items from a test.
Most famously, it is used in split-half reliability testing:
- You split a questionnaire into two halves (e.g., odd-numbered vs. even-numbered items)
- You calculate the correlation between participants' scores on each half
- Because a half-test is less reliable than a full test, you apply the Spearman-Brown correction to estimate the full test's reliability
The Formula
r_SB = (n × r) / (1 + (n - 1) × r)
Where:
- r_SB = corrected (predicted) reliability
- r = observed correlation between the two halves
- n = the factor by which the test length is being multiplied (for split-half, n = 2)
For split-half specifically, it simplifies to:
r_SB = (2 × r) / (1 + r)
Why Were 56 Adolescents Excluded?
This is the key clarification: the Spearman-Brown formula itself does not cause participant exclusion. The sentence you read likely means that the researchers applied the Spearman-Brown formula to assess reliability of their scale, and separately, 56 adolescents were excluded because they had incomplete data (i.e., they didn't answer all items).
Reliability formulas like Spearman-Brown require complete data on all items to compute accurate split-half correlations. If participants skipped questions or had missing responses, including them would distort the reliability estimate - so they were removed from the analysis before applying the formula.
In Simple Terms
Think of it like a ruler: if you only have half a ruler, you can still estimate how accurate a full ruler would be - that's essentially what Spearman-Brown does for tests. The 56 adolescents were excluded not because of the formula, but because their incomplete answers would have made the formula's calculation inaccurate.
What is directed acyclic graph?
Searching the Web
directed acyclic graph DAG explained with examples
Directed Acyclic Graph (DAG)
Breaking Down the Name
The term has three parts, each carrying meaning:
| Term | Meaning |
|---|---|
| Directed | Every connection (edge) has a direction - it goes FROM one node TO another (like a one-way street) |
| Acyclic | There are no cycles - you can never follow the arrows and end up back where you started |
| Graph | A structure made of nodes (points) connected by edges (lines/arrows) |
Visual Idea
A --> B --> D
| ^
v |
C ---------
- You can go A → B → D, or A → C → D
- But you can never loop back to A
- This makes it a DAG
A cyclic graph (NOT a DAG) would look like: A → B → C → A (a loop back to the start).
Key Properties
- One-way flow - edges point in a single direction
- No loops - you can never revisit a node by following the arrows
- Topological ordering - you can always arrange the nodes in a sequence where all arrows point "forward" (left to right, top to bottom, etc.)
Where DAGs Are Used
1. Statistics & Causal Research
In research (like the study you were reading), DAGs are used to visually represent cause-and-effect relationships between variables. Researchers draw arrows from causes to effects to map out what influences what - helping decide which variables to include or exclude in a statistical model.
2. Data Pipelines & Workflow Scheduling
Tools like Apache Airflow use DAGs to model task sequences - Task A must complete before Task B, Task B before Task C, and so on, with no circular dependencies.
3. Computer Science / Compilers
DAGs represent expressions and program flow, helping compilers optimize code.
4. Blockchain
Some cryptocurrencies (like IOTA) use DAG structures instead of traditional blockchains.
5. Version Control
Git's commit history is a DAG - each commit points to its parent(s), and you never loop back.
Simple Real-World Analogy
Think of a university course prerequisite system:
- Math 101 → Math 201 → Math 301
- You must take earlier courses before later ones
- You can never "loop back" and have a later course be a prerequisite for an earlier one
That structure is a DAG.
In the Context of Your Research Study
If a paper mentions a DAG, it almost certainly refers to causal DAGs in statistics - a diagram showing assumed causal pathways between variables (e.g., exposure → outcome, with confounders drawn as common causes of both). Researchers use these diagrams to make their assumptions transparent and to guide which variables need to be controlled for in regression models.
Descriptive analyses were performed with estimates of absolute frequencies and percentages. The mean out comes between the groups were compared using the Stu dent’s t-test or Wilcoxon Test and one-way ANOVA or Kruskal–Wallis Test. Regarding the covariates of the study, a directed acy clic graph (DAG) was built from the literature review on the study theme, which shows the complex relationship between the different risk factors (Figs. 1, 2, and 3). DAG is a visual and qualitative tool for selecting confounding variables identified from a theoretical model. The arrow heads inform a path between two variables, and it is possible, through pre-established rules, to identify a min imum set of variables for adjustment. After application of DAG’s rules, the minimum adjustment set of variables for analysis of association between sleep duration and cardi ometabolic outcomes (blood pressure, dyslipidemia and glycemia) were: For the associations between sleep and diabetes, the following variables were selected: age, gen der, adolescent years of study, skin color, economic class, current work, alcohol consumption, smoking, physi cal activity, screen time, food consumption. For sleep and blood pressure, in turn, the variables selected were adjusted for: age, gender, adolescent years of study, skin color, economic class, current work, alcohol consump tion, smoking, physical activity, screen time, food con sumption, depression. In the association between sleep and dyslipidemias, the adjustment variables were: age, gender, adolescent years of study, skin color, economic class, current work, alcohol consumption, smoking, physical activity, screen time, food consumption. To analyze the relationship between sleep duration and cardiometabolic outcomes, a weighted by propensity score was used from the Inverse Probability of Treatment Weighting (IPTW). This study used the propensity score as a way to correct the imbalance in the distribution of predictors (covariate) that occur in observational studies, thus making the groups comparable and avoiding biased estimates of the effect of sleep duration and cardiometa bolic outcomes. The propensity score calculation was performed using the linear regression. The application of linear regression allowed the estimation of the selection probability of ado lescents, conditional on the values of covariates identified in the DAG (Figs. 1, 2 and 3). Checking the common support area and balancing the groups verified adequate propensity score estimation. A balance in the distribution of observed covariates was achieved, obtaining standardized differences between means and proportions less than 0.1 and for variances between 0.18–1.2. The effect of of sleep duration on car diometabolic outcomes was calculated using the beta coefficient with 95% confidence intervals (95% CI) for each cardiometabolic outcome considered in the analysis. No interaction was observed between sleep duration and gender concerning glycemia, blood pressure, and dyslipi demia outcomes evaluated by the gender/sleep duration ratio. Therefore, the sample was not stratified by gender. explain everything about this data analysis in a very understandable way.
This is a research paper studying how sleep duration affects cardiometabolic health (blood pressure, blood sugar, and cholesterol) in adolescents. Let me break down every part of the analysis in plain language.
STEP 1: Descriptive Analysis
"Absolute frequencies and percentages"
Before doing any fancy statistics, the researchers first simply described who was in their study.
- Absolute frequency = the actual count (e.g., "120 adolescents had high blood pressure")
- Percentage = that count as a proportion (e.g., "25% of the sample")
Think of it like a class roster summary: "30 students total - 18 girls (60%), 12 boys (40%)."
STEP 2: Comparing Groups - Which Test Was Used and Why
The researchers compared average outcomes between groups (e.g., short sleepers vs. normal sleepers). They used two pairs of tests, depending on the data:
Pair 1: Comparing TWO groups
| Test | When Used |
|---|---|
| Student's t-test | When data is normally distributed (bell-shaped curve) |
| Wilcoxon Test | When data is NOT normally distributed (skewed) |
Pair 2: Comparing THREE or MORE groups
| Test | When Used |
|---|---|
| One-way ANOVA | When data is normally distributed |
| Kruskal-Wallis Test | When data is NOT normally distributed |
Simple analogy: Imagine comparing test scores across groups. If the scores are spread evenly (normal), use t-test/ANOVA. If most scores are bunched at one end (skewed), use Wilcoxon/Kruskal-Wallis. Both pairs do the same job - just suited to different data shapes.
STEP 3: The DAG (Directed Acyclic Graph) - Choosing the Right Variables
This is where the researchers got smart about confounding - a major problem in observational studies.
What is Confounding?
Imagine you find that "adolescents who sleep less have higher blood pressure." But wait - maybe those same adolescents also:
- Exercise less
- Eat more junk food
- Come from lower-income families
These factors could ALSO raise blood pressure. So is it really the sleep causing the problem, or is it those other factors? Those "other factors" are called confounders.
What the DAG Does
The researchers drew a diagram (the DAG) mapping out ALL the known relationships between variables, based on existing literature. Arrows show causal pathways:
Physical Activity --> Blood Pressure
^
|
Sleep Duration --> Blood Pressure
|
v
Food Consumption --> Blood Pressure
By following pre-established rules on the DAG, they identified the minimum set of variables they needed to control for - not too few (which causes bias), not too many (which causes other statistical problems).
What Variables Were Selected for Each Outcome?
| Outcome | Variables Adjusted For |
|---|---|
| Sleep & Diabetes (Glycemia) | Age, gender, years of study, skin color, economic class, work status, alcohol, smoking, physical activity, screen time, food consumption |
| Sleep & Blood Pressure | All of the above + depression |
| Sleep & Dyslipidemia (Cholesterol) | Same as diabetes (without depression) |
Note that depression was added for blood pressure but not the others - because the DAG showed depression has a specific causal pathway to blood pressure that doesn't apply the same way to the other outcomes.
STEP 4: Propensity Score & IPTW - Making Groups Comparable
This is the most sophisticated part of the analysis.
The Core Problem in Observational Studies
In a randomized trial, participants are randomly assigned to groups (short sleep vs. normal sleep), so the groups are balanced. But this is an observational study - the researchers didn't assign sleep durations. Adolescents naturally sleep different amounts, and the short-sleepers might be systematically different from normal-sleepers in many ways (poorer, more stressed, less active, etc.).
This imbalance means a direct comparison would give biased results.
What is a Propensity Score?
A propensity score is the probability that a participant ends up in a particular group (e.g., "short sleeper"), given all their background characteristics (the covariates identified by the DAG).
Think of it like this: Two adolescents both have a 70% chance of being short sleepers based on their profile (age, income, physical activity, etc.). Even though one sleeps short and the other doesn't, their background is essentially the same - so comparing them is fair.
The propensity score was calculated using linear regression - a mathematical equation that takes all the covariates and outputs a single probability score for each adolescent.
What is IPTW (Inverse Probability of Treatment Weighting)?
Once you have the propensity scores, IPTW uses them as weights to rebalance the groups mathematically.
- An adolescent who is a short sleeper but had a LOW probability of being one gets a high weight (they're unusual and represent many similar people who didn't become short sleepers)
- An adolescent who is a short sleeper and had a HIGH probability of being one gets a low weight (they're typical, less informative)
Analogy: Imagine a survey where rich people are over-represented. You'd "down-weight" rich respondents and "up-weight" poor respondents to make the sample mirror the real population. IPTW does the same thing for the groups being compared.
The result: a pseudo-population where short sleepers and normal sleepers look similar in terms of all background variables - just like a randomized trial.
STEP 5: Checking the Propensity Score Worked
"Common Support Area"
This checks that both groups (short sleepers and normal sleepers) have overlapping propensity score ranges. If one group has scores of 0.1-0.4 and the other has 0.6-0.9, there's no overlap and comparison is invalid. Overlap was confirmed here - good sign.
"Standardized Differences < 0.1"
After applying IPTW, they checked whether the covariates were now balanced between groups. A standardized difference below 0.1 (10%) is the accepted threshold for good balance. They achieved this - meaning the groups were now comparable.
"Variance Ratio 0.18-1.2"
This checks that the spread (variance) of variables is similar across groups after weighting. Values close to 1.0 are ideal. Their range of 0.18-1.2 was acceptable.
STEP 6: Measuring the Effect - Beta Coefficient & 95% CI
Once the groups were balanced, they calculated:
- Beta coefficient (β): How much does each cardiometabolic outcome change for each unit change in sleep duration? For example, "each additional hour of sleep is associated with a β = -2.3 mmHg drop in systolic blood pressure."
- 95% Confidence Interval (95% CI): The range within which the true effect likely falls. If the CI does NOT cross zero, the result is statistically significant.
STEP 7: No Gender Interaction Found
They tested whether the effect of sleep on outcomes differed between boys and girls (called an interaction or effect modification). They checked this by looking at the "gender/sleep duration ratio."
No significant interaction was found - meaning sleep affected blood pressure, blood sugar, and cholesterol similarly in both sexes. Therefore, they analyzed boys and girls together (no need to split the sample).
The Big Picture - Summary Flowchart
Describe the sample (frequencies & %)
↓
Compare groups (t-test / ANOVA or non-parametric equivalents)
↓
Build DAG → identify confounders to control for
↓
Calculate propensity scores (linear regression)
↓
Apply IPTW → balance the groups
↓
Check balance (standardized differences < 0.1)
↓
Estimate effect (beta coefficient + 95% CI)
↓
Test for gender interaction → none found → analyze combined
The entire analysis was designed to answer one question as cleanly as possible: Does how long adolescents sleep actually affect their cardiometabolic health - or is the apparent relationship just due to other background differences between groups? The propensity score approach was their way of simulating a fair comparison despite not having a randomized experiment.
Create a visual summary document explaining this statistical method
Asking for Preferences
Reading File
Reading File
Running Command
ls -la /home/daytona/workspace/ && mkdir -p /home/daytona/workspace/stats-summary
Writing File
~/stats-summary/create_pptx.js
const pptxgen = require("pptxgenjs");
const pres = new pptxgen();
pres.layout = "LAYOUT_16x9";
pres.title = "Statistical Methods – Sleep & Cardiometabolic Health";
// ─── COLOR PALETTE ───────────────────────────────────────────────────────────
const C = {
navy: "0D1F3C", // dominant dark
teal: "1B7F8A", // primary accent
mint: "2EC4B6", // highlight
amber: "F4A300", // warm accent
light: "F0F4F8", // light bg
white: "FFFFFF",
gray: "6B7A8D",
lgray: "D6DDE8",
red: "E84855",
};
// ─── HELPERS ─────────────────────────────────────────────────────────────────
function headerBar(slide, title, subtitle) {
// dark header band
slide.addShape(pres.ShapeType.rect, { x: 0, y: 0, w: 10, h: 1.15, fill: { color: C.navy } });
slide.addShape(pres.ShapeType.rect, { x: 0, y: 1.15, w: 10, h: 0.06, fill: { color: C.mint } });
slide.addText(title, { x: 0.35, y: 0.08, w: 7.5, h: 0.65, fontSize: 22, bold: true, color: C.white, fontFace: "Calibri", margin: 0 });
if (subtitle) slide.addText(subtitle, { x: 0.35, y: 0.72, w: 8, h: 0.38, fontSize: 11, color: C.mint, fontFace: "Calibri", margin: 0 });
}
function stepBox(slide, x, y, w, h, num, title, body, bg) {
slide.addShape(pres.ShapeType.roundRect, { x, y, w, h, fill: { color: bg || C.light }, line: { color: C.lgray, width: 1 }, rectRadius: 0.08 });
// number circle
slide.addShape(pres.ShapeType.ellipse, { x: x + 0.1, y: y + 0.1, w: 0.42, h: 0.42, fill: { color: C.teal } });
slide.addText(String(num), { x: x + 0.1, y: y + 0.1, w: 0.42, h: 0.42, fontSize: 13, bold: true, color: C.white, align: "center", valign: "middle", fontFace: "Calibri", margin: 0 });
slide.addText(title, { x: x + 0.6, y: y + 0.1, w: w - 0.7, h: 0.4, fontSize: 12, bold: true, color: C.navy, fontFace: "Calibri", margin: 0 });
slide.addText(body, { x: x + 0.15, y: y + 0.55, w: w - 0.25, h: h - 0.65, fontSize: 10, color: C.navy, fontFace: "Calibri", margin: 0, valign: "top" });
}
function arrowRight(slide, x, y) {
slide.addShape(pres.ShapeType.rightArrow, { x, y, w: 0.4, h: 0.3, fill: { color: C.teal } });
}
function arrowDown(slide, x, y) {
slide.addShape(pres.ShapeType.downArrow, { x, y, w: 0.3, h: 0.35, fill: { color: C.teal } });
}
// ═══════════════════════════════════════════════════════════════════════════════
// SLIDE 1 – TITLE
// ═══════════════════════════════════════════════════════════════════════════════
{
const s = pres.addSlide();
s.addShape(pres.ShapeType.rect, { x: 0, y: 0, w: 10, h: 5.625, fill: { color: C.navy } });
// accent stripe
s.addShape(pres.ShapeType.rect, { x: 0, y: 2.7, w: 10, h: 0.07, fill: { color: C.mint } });
s.addShape(pres.ShapeType.rect, { x: 0, y: 2.77, w: 10, h: 0.03, fill: { color: C.amber } });
s.addText("Statistical Methods", { x: 0.6, y: 0.7, w: 8.8, h: 0.8, fontSize: 36, bold: true, color: C.white, fontFace: "Calibri", align: "center" });
s.addText("Sleep Duration & Cardiometabolic Outcomes in Adolescents", { x: 0.6, y: 1.55, w: 8.8, h: 0.6, fontSize: 18, color: C.mint, fontFace: "Calibri", align: "center" });
s.addText("A Visual Step-by-Step Summary", { x: 0.6, y: 2.2, w: 8.8, h: 0.4, fontSize: 13, color: C.lgray, fontFace: "Calibri", align: "center", italic: true });
// steps overview boxes
const steps = ["1. Describe", "2. Compare", "3. DAG", "4. Propensity Score", "5. IPTW", "6. Results"];
const bw = 1.42, bx0 = 0.3;
steps.forEach((t, i) => {
s.addShape(pres.ShapeType.roundRect, { x: bx0 + i * (bw + 0.07), y: 3.0, w: bw, h: 0.65, fill: { color: C.teal }, line: { color: C.mint, width: 1 }, rectRadius: 0.07 });
s.addText(t, { x: bx0 + i * (bw + 0.07), y: 3.0, w: bw, h: 0.65, fontSize: 10, bold: true, color: C.white, align: "center", valign: "middle", fontFace: "Calibri", margin: 0 });
});
s.addText("Study Goal: Does how long adolescents sleep affect their blood pressure, blood sugar, and cholesterol?", {
x: 0.6, y: 3.85, w: 8.8, h: 0.55, fontSize: 12, color: C.amber, fontFace: "Calibri", align: "center", italic: true, bold: true
});
s.addText("Each slide breaks down one analytical step — what it is, why it was done, and what it tells us.", {
x: 0.6, y: 4.45, w: 8.8, h: 0.4, fontSize: 10, color: C.lgray, fontFace: "Calibri", align: "center"
});
}
// ═══════════════════════════════════════════════════════════════════════════════
// SLIDE 2 – STEP 1: DESCRIPTIVE ANALYSIS
// ═══════════════════════════════════════════════════════════════════════════════
{
const s = pres.addSlide();
s.addShape(pres.ShapeType.rect, { x: 0, y: 0, w: 10, h: 5.625, fill: { color: C.light } });
headerBar(s, "STEP 1 — Descriptive Analysis", "Getting to know the sample before any testing");
// left explanation
s.addShape(pres.ShapeType.roundRect, { x: 0.25, y: 1.4, w: 4.5, h: 3.8, fill: { color: C.white }, line: { color: C.lgray, width: 1 }, rectRadius: 0.1 });
s.addText("What is it?", { x: 0.4, y: 1.5, w: 4.2, h: 0.35, fontSize: 13, bold: true, color: C.teal, fontFace: "Calibri", margin: 0 });
s.addText([
{ text: "Before any complex statistics, researchers simply described who was in the study using:\n\n", options: {} },
{ text: "• Absolute Frequency:", options: { bold: true } },
{ text: " The raw count of people in each category\n (e.g., 120 had high blood pressure)\n\n", options: {} },
{ text: "• Percentage:", options: { bold: true } },
{ text: " That count as a proportion of the whole\n (e.g., 25% of the sample)\n\n", options: {} },
{ text: "Think of it like a class roll-call summary:\n\"30 students — 18 girls (60%), 12 boys (40%)\"", options: { italic: true, color: C.gray } },
], { x: 0.4, y: 1.9, w: 4.2, h: 3.1, fontSize: 10.5, color: C.navy, fontFace: "Calibri", valign: "top" });
// right: visual icons
s.addShape(pres.ShapeType.rect, { x: 5.1, y: 1.4, w: 4.6, h: 1.55, fill: { color: C.teal }, line: { color: C.teal }, rectRadius: 0.08 });
s.addText("ABSOLUTE FREQUENCY", { x: 5.2, y: 1.48, w: 4.4, h: 0.35, fontSize: 11, bold: true, color: C.white, fontFace: "Calibri", charSpacing: 2 });
s.addText("\"How many?\" → 120 adolescents had high BP", { x: 5.2, y: 1.85, w: 4.4, h: 0.35, fontSize: 11, color: C.mint, fontFace: "Calibri" });
s.addText("Raw count of people per group or category", { x: 5.2, y: 2.2, w: 4.4, h: 0.35, fontSize: 10, color: C.lgray, fontFace: "Calibri", italic: true });
s.addShape(pres.ShapeType.rect, { x: 5.1, y: 3.1, w: 4.6, h: 1.55, fill: { color: C.navy }, line: { color: C.navy }, rectRadius: 0.08 });
s.addText("PERCENTAGE (%)", { x: 5.2, y: 3.18, w: 4.4, h: 0.35, fontSize: 11, bold: true, color: C.white, fontFace: "Calibri", charSpacing: 2 });
s.addText("\"What proportion?\" → 25% of the sample", { x: 5.2, y: 3.55, w: 4.4, h: 0.35, fontSize: 11, color: C.amber, fontFace: "Calibri" });
s.addText("Puts the count in context relative to the whole group", { x: 5.2, y: 3.9, w: 4.4, h: 0.35, fontSize: 10, color: C.lgray, fontFace: "Calibri", italic: true });
s.addText("PURPOSE: Paint a clear picture of the sample's demographics before testing begins", {
x: 0.25, y: 5.1, w: 9.5, h: 0.38, fontSize: 10, color: C.gray, fontFace: "Calibri", italic: true, align: "center"
});
}
// ═══════════════════════════════════════════════════════════════════════════════
// SLIDE 3 – STEP 2: COMPARING GROUPS
// ═══════════════════════════════════════════════════════════════════════════════
{
const s = pres.addSlide();
s.addShape(pres.ShapeType.rect, { x: 0, y: 0, w: 10, h: 5.625, fill: { color: C.light } });
headerBar(s, "STEP 2 — Comparing Group Means", "Choosing the right statistical test based on data distribution");
// decision question
s.addShape(pres.ShapeType.roundRect, { x: 3.2, y: 1.35, w: 3.6, h: 0.65, fill: { color: C.amber }, rectRadius: 0.08 });
s.addText("Is the data normally distributed?", { x: 3.2, y: 1.35, w: 3.6, h: 0.65, fontSize: 12, bold: true, color: C.navy, align: "center", valign: "middle", fontFace: "Calibri" });
// YES branch
s.addShape(pres.ShapeType.rect, { x: 0.15, y: 1.35, w: 0.7, h: 0.65, fill: { color: C.mint }, line: { color: C.mint } });
s.addText("YES", { x: 0.15, y: 1.35, w: 0.7, h: 0.65, fontSize: 13, bold: true, color: C.white, align: "center", valign: "middle", fontFace: "Calibri" });
// NO branch
s.addShape(pres.ShapeType.rect, { x: 9.15, y: 1.35, w: 0.7, h: 0.65, fill: { color: C.red }, line: { color: C.red } });
s.addText("NO", { x: 9.15, y: 1.35, w: 0.7, h: 0.65, fontSize: 13, bold: true, color: C.white, align: "center", valign: "middle", fontFace: "Calibri" });
// lines connecting
s.addShape(pres.ShapeType.line, { x: 0.85, y: 1.67, w: 2.35, h: 0, line: { color: C.mint, width: 2, dashType: "solid" } });
s.addShape(pres.ShapeType.line, { x: 6.8, y: 1.67, w: 2.35, h: 0, line: { color: C.red, width: 2, dashType: "solid" } });
// Parametric (YES) side
s.addShape(pres.ShapeType.roundRect, { x: 0.15, y: 2.2, w: 4.4, h: 1.3, fill: { color: C.mint }, rectRadius: 0.08 });
s.addText("PARAMETRIC TESTS", { x: 0.25, y: 2.26, w: 4.2, h: 0.35, fontSize: 11, bold: true, color: C.white, fontFace: "Calibri", charSpacing: 2 });
s.addText([
{ text: "2 groups: ", options: {} }, { text: "Student's t-test", options: { bold: true } }, { text: "\n", options: {} },
{ text: "3+ groups: ", options: {} }, { text: "One-Way ANOVA", options: { bold: true } }
], { x: 0.25, y: 2.62, w: 4.2, h: 0.75, fontSize: 11, color: C.white, fontFace: "Calibri" });
// Non-Parametric (NO) side
s.addShape(pres.ShapeType.roundRect, { x: 5.45, y: 2.2, w: 4.4, h: 1.3, fill: { color: C.red }, rectRadius: 0.08 });
s.addText("NON-PARAMETRIC TESTS", { x: 5.55, y: 2.26, w: 4.2, h: 0.35, fontSize: 11, bold: true, color: C.white, fontFace: "Calibri", charSpacing: 2 });
s.addText([
{ text: "2 groups: ", options: {} }, { text: "Wilcoxon Test", options: { bold: true } }, { text: "\n", options: {} },
{ text: "3+ groups: ", options: {} }, { text: "Kruskal-Wallis Test", options: { bold: true } }
], { x: 5.55, y: 2.62, w: 4.2, h: 0.75, fontSize: 11, color: C.white, fontFace: "Calibri" });
// analogy box
s.addShape(pres.ShapeType.roundRect, { x: 0.15, y: 3.7, w: 9.7, h: 1.6, fill: { color: C.white }, line: { color: C.lgray, width: 1 }, rectRadius: 0.08 });
s.addText("Why two sets of tests?", { x: 0.35, y: 3.78, w: 4, h: 0.35, fontSize: 12, bold: true, color: C.teal, fontFace: "Calibri", margin: 0 });
s.addText("Parametric tests assume a bell-shaped (normal) distribution. When data is skewed (bunched to one side), those assumptions break down and non-parametric alternatives are used instead. Both pairs ask the same question — \"Are the group means different?\" — but are suited to different data shapes. It's like choosing the right tool for the job.", {
x: 0.35, y: 4.15, w: 9.35, h: 0.95, fontSize: 10.5, color: C.navy, fontFace: "Calibri", valign: "top"
});
}
// ═══════════════════════════════════════════════════════════════════════════════
// SLIDE 4 – STEP 3: DAG
// ═══════════════════════════════════════════════════════════════════════════════
{
const s = pres.addSlide();
s.addShape(pres.ShapeType.rect, { x: 0, y: 0, w: 10, h: 5.625, fill: { color: C.light } });
headerBar(s, "STEP 3 — Directed Acyclic Graph (DAG)", "A causal map to identify which variables need to be controlled for");
// left: what & why
s.addShape(pres.ShapeType.roundRect, { x: 0.2, y: 1.35, w: 4.2, h: 3.9, fill: { color: C.white }, line: { color: C.lgray, width: 1 }, rectRadius: 0.1 });
s.addText("What is a DAG?", { x: 0.35, y: 1.45, w: 3.9, h: 0.35, fontSize: 13, bold: true, color: C.teal, fontFace: "Calibri", margin: 0 });
s.addText([
{ text: "A DAG is a diagram where:\n\n", options: {} },
{ text: "• Nodes", options: { bold: true } },
{ text: " = variables (sleep, age, diet…)\n\n", options: {} },
{ text: "• Arrows", options: { bold: true } },
{ text: " = assumed causal direction\n\n", options: {} },
{ text: "• No loops", options: { bold: true } },
{ text: " — arrows never circle back\n\n", options: {} },
{ text: "By following DAG rules, researchers identify the ", options: {} },
{ text: "minimum set of variables", options: { bold: true, color: C.teal } },
{ text: " to adjust for — not too few (bias), not too many (over-control).", options: {} },
], { x: 0.35, y: 1.85, w: 3.9, h: 3.2, fontSize: 10.5, color: C.navy, fontFace: "Calibri", valign: "top" });
// right: mini DAG diagram
s.addShape(pres.ShapeType.roundRect, { x: 4.65, y: 1.35, w: 5.1, h: 2.6, fill: { color: C.navy }, rectRadius: 0.1 });
s.addText("Example Causal Pathway", { x: 4.75, y: 1.42, w: 4.9, h: 0.35, fontSize: 11, bold: true, color: C.mint, fontFace: "Calibri" });
// nodes
const nodes = [
{ label: "Physical\nActivity", x: 4.85, y: 1.95, w: 1.2, h: 0.6, col: C.teal },
{ label: "Sleep\nDuration", x: 6.4, y: 2.4, w: 1.2, h: 0.6, col: C.amber },
{ label: "Blood\nPressure", x: 7.95, y: 1.95, w: 1.2, h: 0.6, col: C.red },
{ label: "Diet /\nFood", x: 4.85, y: 2.85, w: 1.2, h: 0.6, col: C.teal },
];
nodes.forEach(n => {
s.addShape(pres.ShapeType.roundRect, { x: n.x, y: n.y, w: n.w, h: n.h, fill: { color: n.col }, rectRadius: 0.07 });
s.addText(n.label, { x: n.x, y: n.y, w: n.w, h: n.h, fontSize: 9, bold: true, color: C.white, align: "center", valign: "middle", fontFace: "Calibri", margin: 0 });
});
// arrows (lines)
s.addShape(pres.ShapeType.line, { x: 6.05, y: 2.25, w: 1.9, h: -0.05, line: { color: C.mint, width: 1.5 } });
s.addShape(pres.ShapeType.line, { x: 6.05, y: 3.15, w: 1.9, h: -0.95, line: { color: C.mint, width: 1.5 } });
s.addShape(pres.ShapeType.line, { x: 5.45, y: 1.95, w: 0.95, h: 0.45, line: { color: C.lgray, width: 1, dashType: "dash" } });
// adjustment sets table
s.addShape(pres.ShapeType.roundRect, { x: 4.65, y: 4.1, w: 5.1, h: 1.2, fill: { color: C.white }, line: { color: C.lgray, width: 1 }, rectRadius: 0.08 });
s.addText("Minimum Adjustment Sets Identified:", { x: 4.8, y: 4.16, w: 4.8, h: 0.3, fontSize: 10, bold: true, color: C.navy, fontFace: "Calibri", margin: 0 });
s.addText([
{ text: "Sleep & Diabetes / Dyslipidemia: ", options: { bold: true } },
{ text: "age, gender, education, skin color, income, work, alcohol, smoking, physical activity, screen time, diet\n", options: {} },
{ text: "Sleep & Blood Pressure: ", options: { bold: true } },
{ text: "All above + depression", options: { color: C.red } },
], { x: 4.8, y: 4.5, w: 4.85, h: 0.72, fontSize: 9.5, color: C.navy, fontFace: "Calibri", valign: "top" });
}
// ═══════════════════════════════════════════════════════════════════════════════
// SLIDE 5 – STEP 4: PROPENSITY SCORE
// ═══════════════════════════════════════════════════════════════════════════════
{
const s = pres.addSlide();
s.addShape(pres.ShapeType.rect, { x: 0, y: 0, w: 10, h: 5.625, fill: { color: C.light } });
headerBar(s, "STEP 4 — Propensity Score", "Estimating each participant's probability of being a short sleeper");
// problem box
s.addShape(pres.ShapeType.roundRect, { x: 0.2, y: 1.35, w: 9.6, h: 1.1, fill: { color: C.red }, rectRadius: 0.08 });
s.addText("THE PROBLEM IN OBSERVATIONAL STUDIES", { x: 0.35, y: 1.4, w: 5, h: 0.3, fontSize: 10, bold: true, color: C.white, fontFace: "Calibri", charSpacing: 1 });
s.addText("Unlike clinical trials, researchers didn't randomly assign sleep durations. Short sleepers may naturally differ from normal sleepers in many ways (poorer, more stressed, less active) — creating an unfair, biased comparison.", {
x: 0.35, y: 1.72, w: 9.3, h: 0.55, fontSize: 10.5, color: C.white, fontFace: "Calibri"
});
// solution
s.addText("THE SOLUTION: Propensity Score", { x: 0.25, y: 2.6, w: 9.5, h: 0.35, fontSize: 13, bold: true, color: C.teal, fontFace: "Calibri" });
// three boxes
const boxes = [
{ x: 0.2, label: "1. Collect Covariates", body: "Gather all background variables identified by the DAG (age, income, physical activity, diet, etc.)" },
{ x: 3.45, label: "2. Run Linear Regression", body: "Feed all covariates into a linear regression model to calculate each person's probability of being a short sleeper" },
{ x: 6.7, label: "3. Propensity Score", body: "One number per participant: the estimated probability (0–1) of being a short sleeper given their profile" },
];
boxes.forEach(b => {
s.addShape(pres.ShapeType.roundRect, { x: b.x, y: 3.05, w: 3.1, h: 1.8, fill: { color: C.white }, line: { color: C.teal, width: 1.5 }, rectRadius: 0.08 });
s.addText(b.label, { x: b.x + 0.12, y: 3.12, w: 2.86, h: 0.4, fontSize: 11, bold: true, color: C.teal, fontFace: "Calibri", margin: 0 });
s.addText(b.body, { x: b.x + 0.12, y: 3.55, w: 2.86, h: 1.2, fontSize: 10, color: C.navy, fontFace: "Calibri", valign: "top" });
});
// arrows between boxes
arrowRight(s, 3.3, 3.85);
arrowRight(s, 6.55, 3.85);
s.addText("Analogy: Two students with identical backgrounds — same grades, habits, and family income. One studies 8h, the other 4h. The propensity score makes these two comparable, because their profiles are virtually the same.", {
x: 0.2, y: 5.08, w: 9.6, h: 0.42, fontSize: 10, color: C.gray, fontFace: "Calibri", italic: true, align: "center"
});
}
// ═══════════════════════════════════════════════════════════════════════════════
// SLIDE 6 – STEP 5: IPTW
// ═══════════════════════════════════════════════════════════════════════════════
{
const s = pres.addSlide();
s.addShape(pres.ShapeType.rect, { x: 0, y: 0, w: 10, h: 5.625, fill: { color: C.light } });
headerBar(s, "STEP 5 — IPTW: Inverse Probability of Treatment Weighting", "Re-balancing groups to simulate a randomized trial");
s.addText("What is IPTW?", { x: 0.25, y: 1.38, w: 9.5, h: 0.38, fontSize: 14, bold: true, color: C.navy, fontFace: "Calibri" });
s.addText("Once propensity scores are calculated, IPTW uses them as weights to mathematically balance the groups — making short sleepers and normal sleepers similar in background characteristics.", {
x: 0.25, y: 1.78, w: 9.5, h: 0.5, fontSize: 11, color: C.navy, fontFace: "Calibri"
});
// weighting diagram
const wboxes = [
{ x: 0.2, bg: C.teal, title: "SHORT SLEEPER\nHigh Probability", body: "Expected to be short sleeper\n→ LOW weight\n(common, less informative)", w_label: "Weight: LOW", wc: C.amber },
{ x: 3.45, bg: C.navy, title: "SHORT SLEEPER\nLow Probability", body: "NOT expected to be short sleeper, but is\n→ HIGH weight\n(unusual, highly informative)", w_label: "Weight: HIGH", wc: C.mint },
{ x: 6.7, bg: C.red, title: "RESULT", body: "Groups become statistically equivalent in background — like a randomised trial", w_label: "Balanced Groups ✓", wc: C.white },
];
wboxes.forEach(b => {
s.addShape(pres.ShapeType.roundRect, { x: b.x, y: 2.45, w: 3.1, h: 2.0, fill: { color: b.bg }, rectRadius: 0.08 });
s.addText(b.title, { x: b.x + 0.1, y: 2.52, w: 2.9, h: 0.55, fontSize: 10, bold: true, color: C.white, fontFace: "Calibri", align: "center", valign: "middle" });
s.addText(b.body, { x: b.x + 0.1, y: 3.1, w: 2.9, h: 0.9, fontSize: 10, color: C.lgray, fontFace: "Calibri", align: "center" });
s.addText(b.w_label, { x: b.x + 0.1, y: 4.05, w: 2.9, h: 0.3, fontSize: 11, bold: true, color: b.wc, fontFace: "Calibri", align: "center" });
});
arrowRight(s, 3.3, 3.35);
arrowRight(s, 6.55, 3.35);
// balance check
s.addShape(pres.ShapeType.roundRect, { x: 0.2, y: 4.56, w: 9.6, h: 0.88, fill: { color: C.white }, line: { color: C.mint, width: 1.5 }, rectRadius: 0.08 });
s.addText([
{ text: "Balance Check Passed: ", options: { bold: true, color: C.teal } },
{ text: "Standardized differences < 0.1 ", options: {} },
{ text: "✓ ", options: { bold: true, color: C.mint } },
{ text: " | Variance ratios 0.18–1.2 ", options: {} },
{ text: "✓", options: { bold: true, color: C.mint } },
], { x: 0.35, y: 4.62, w: 9.3, h: 0.38, fontSize: 11, color: C.navy, fontFace: "Calibri", align: "center" });
s.addText("Both metrics confirm that groups are now equivalent in background characteristics — the weighting worked.", {
x: 0.35, y: 5.0, w: 9.3, h: 0.3, fontSize: 9.5, color: C.gray, fontFace: "Calibri", align: "center", italic: true
});
}
// ═══════════════════════════════════════════════════════════════════════════════
// SLIDE 7 – STEP 6: RESULTS & GENDER INTERACTION
// ═══════════════════════════════════════════════════════════════════════════════
{
const s = pres.addSlide();
s.addShape(pres.ShapeType.rect, { x: 0, y: 0, w: 10, h: 5.625, fill: { color: C.light } });
headerBar(s, "STEP 6 — Effect Estimation & Gender Interaction Test", "Measuring the impact of sleep and checking if gender changes the story");
// beta coefficient box
s.addShape(pres.ShapeType.roundRect, { x: 0.2, y: 1.38, w: 4.6, h: 2.5, fill: { color: C.navy }, rectRadius: 0.1 });
s.addText("Beta Coefficient (β)", { x: 0.35, y: 1.48, w: 4.3, h: 0.38, fontSize: 13, bold: true, color: C.mint, fontFace: "Calibri", margin: 0 });
s.addText([
{ text: "Measures how much a cardiometabolic outcome changes for each unit change in sleep duration, after balancing.\n\n", options: {} },
{ text: "Example: ", options: { bold: true, color: C.amber } },
{ text: "β = −2.3 mmHg means each extra hour of sleep is associated with a 2.3 mmHg drop in systolic blood pressure.\n\n", options: {} },
{ text: "If β = 0, there is no association.", options: { italic: true, color: C.lgray } },
], { x: 0.35, y: 1.9, w: 4.3, h: 1.85, fontSize: 10.5, color: C.white, fontFace: "Calibri", valign: "top" });
// 95% CI box
s.addShape(pres.ShapeType.roundRect, { x: 5.2, y: 1.38, w: 4.6, h: 2.5, fill: { color: C.teal }, rectRadius: 0.1 });
s.addText("95% Confidence Interval (CI)", { x: 5.35, y: 1.48, w: 4.3, h: 0.38, fontSize: 13, bold: true, color: C.white, fontFace: "Calibri", margin: 0 });
s.addText([
{ text: "The range where the true effect likely falls, 95% of the time.\n\n", options: {} },
{ text: "Significant: ", options: { bold: true, color: C.amber } },
{ text: "CI does NOT include zero\n e.g., β = −2.3 [−4.1, −0.5] ✓\n\n", options: {} },
{ text: "Not Significant: ", options: { bold: true, color: C.lgray } },
{ text: "CI crosses zero\n e.g., β = −0.5 [−2.1, 1.1] ✗", options: {} },
], { x: 5.35, y: 1.9, w: 4.3, h: 1.85, fontSize: 10.5, color: C.white, fontFace: "Calibri", valign: "top" });
// gender interaction
s.addShape(pres.ShapeType.roundRect, { x: 0.2, y: 4.05, w: 9.6, h: 1.35, fill: { color: C.white }, line: { color: C.amber, width: 1.5 }, rectRadius: 0.08 });
s.addText("Gender Interaction Test — Result: NO Interaction Found", { x: 0.35, y: 4.1, w: 6, h: 0.38, fontSize: 12, bold: true, color: C.amber, fontFace: "Calibri", margin: 0 });
s.addText("Researchers tested whether the effect of sleep on outcomes differed between boys and girls (gender × sleep interaction). No significant difference was found — sleep affected blood pressure, blood sugar, and cholesterol similarly in both sexes. Therefore, the sample was NOT split by gender, and boys and girls were analysed together.", {
x: 0.35, y: 4.5, w: 9.3, h: 0.78, fontSize: 10.5, color: C.navy, fontFace: "Calibri"
});
}
// ═══════════════════════════════════════════════════════════════════════════════
// SLIDE 8 – FULL PIPELINE SUMMARY
// ═══════════════════════════════════════════════════════════════════════════════
{
const s = pres.addSlide();
s.addShape(pres.ShapeType.rect, { x: 0, y: 0, w: 10, h: 5.625, fill: { color: C.navy } });
s.addShape(pres.ShapeType.rect, { x: 0, y: 1.1, w: 10, h: 0.05, fill: { color: C.mint } });
s.addText("THE FULL ANALYTICAL PIPELINE", { x: 0.3, y: 0.12, w: 9.4, h: 0.55, fontSize: 20, bold: true, color: C.white, fontFace: "Calibri", align: "center", charSpacing: 3 });
s.addText("Sleep Duration → Cardiometabolic Outcomes in Adolescents", { x: 0.3, y: 0.65, w: 9.4, h: 0.35, fontSize: 11, color: C.mint, fontFace: "Calibri", align: "center", italic: true });
const pipeline = [
{ n: "1", title: "DESCRIBE", sub: "Frequencies & %", col: C.teal },
{ n: "2", title: "COMPARE", sub: "t-test / ANOVA\nWilcoxon / Kruskal-Wallis", col: "#1B7F8A" },
{ n: "3", title: "DAG", sub: "Select confounders\nvia causal map", col: "#156C76" },
{ n: "4", title: "PROPENSITY\nSCORE", sub: "Linear regression\nper participant", col: "#0F5A62" },
{ n: "5", title: "IPTW", sub: "Balance groups\nwith weights", col: "#094750" },
{ n: "6", title: "RESULTS", sub: "β coefficient\n+ 95% CI", col: C.amber },
];
pipeline.forEach((p, i) => {
const bx = 0.22 + i * 1.62;
s.addShape(pres.ShapeType.roundRect, { x: bx, y: 1.35, w: 1.45, h: 3.5, fill: { color: p.col }, rectRadius: 0.1 });
s.addShape(pres.ShapeType.ellipse, { x: bx + 0.5, y: 1.48, w: 0.45, h: 0.45, fill: { color: C.white } });
s.addText(p.n, { x: bx + 0.5, y: 1.48, w: 0.45, h: 0.45, fontSize: 14, bold: true, color: p.col, align: "center", valign: "middle", fontFace: "Calibri", margin: 0 });
s.addText(p.title, { x: bx + 0.05, y: 2.05, w: 1.35, h: 0.7, fontSize: 10, bold: true, color: C.white, align: "center", valign: "middle", fontFace: "Calibri", margin: 0 });
s.addText(p.sub, { x: bx + 0.05, y: 2.8, w: 1.35, h: 1.8, fontSize: 9, color: C.lgray, align: "center", valign: "top", fontFace: "Calibri", margin: 0 });
});
s.addText("GOAL: Determine whether sleep duration causally affects cardiometabolic health — controlling for confounders and balancing groups to avoid biased conclusions.", {
x: 0.25, y: 5.05, w: 9.5, h: 0.48, fontSize: 10.5, color: C.lgray, fontFace: "Calibri", align: "center", italic: true
});
}
// ─── WRITE FILE ──────────────────────────────────────────────────────────────
pres.writeFile({ fileName: "/home/daytona/workspace/stats-summary/Statistical_Methods_Sleep_Study.pptx" })
.then(() => console.log("DONE: /home/daytona/workspace/stats-summary/Statistical_Methods_Sleep_Study.pptx"))
.catch(e => { console.error(e); process.exit(1); });
Running Command
node /home/daytona/workspace/stats-summary/create_pptx.js
Your visual summary presentation is ready! Here's what's inside:
Statistical Methods Sleep Study
Presentation · PPTX
The deck contains 8 slides:
| Slide | Content |
|---|---|
| 1 - Title | Overview of the study goal and all 6 steps at a glance |
| 2 - Step 1 | Descriptive Analysis - frequencies and percentages explained |
| 3 - Step 2 | Group Comparison tests - visual decision tree (parametric vs. non-parametric) |
| 4 - Step 3 | DAG - causal map diagram with minimum adjustment sets listed |
| 5 - Step 4 | Propensity Score - the problem + 3-step solution flow |
| 6 - Step 5 | IPTW - weighting logic with balance check results |
| 7 - Step 6 | Beta coefficient, 95% CI, and gender interaction test |
| 8 - Summary | Full pipeline visual from start to finish |
The design uses a dark navy/teal/amber color palette with labeled boxes, arrows, and analogy callouts to keep everything readable and visually engaging.
can you not create a single slide?
Writing File
~/stats-summary/create_single_slide.js
const pptxgen = require("pptxgenjs");
const pres = new pptxgen();
pres.layout = "LAYOUT_WIDE"; // 13.3" x 7.5" — more room for content
const C = {
navy: "0D1F3C",
teal: "1B7F8A",
mint: "2EC4B6",
amber: "F4A300",
light: "F0F4F8",
white: "FFFFFF",
gray: "6B7A8D",
lgray: "D6DDE8",
red: "C0392B",
};
const s = pres.addSlide();
// ── BACKGROUND ────────────────────────────────────────────────────────────────
s.addShape(pres.ShapeType.rect, { x: 0, y: 0, w: 13.3, h: 7.5, fill: { color: C.navy } });
// ── TOP HEADER ────────────────────────────────────────────────────────────────
s.addShape(pres.ShapeType.rect, { x: 0, y: 0, w: 13.3, h: 0.9, fill: { color: C.teal } });
s.addShape(pres.ShapeType.rect, { x: 0, y: 0.9, w: 13.3, h: 0.06, fill: { color: C.mint } });
s.addText("Statistical Methods: Sleep Duration & Cardiometabolic Outcomes in Adolescents", {
x: 0.2, y: 0.08, w: 9.5, h: 0.72, fontSize: 18, bold: true, color: C.white, fontFace: "Calibri", valign: "middle", margin: 0
});
s.addText("Study Goal: Does sleep duration affect blood pressure, blood sugar & cholesterol?", {
x: 9.8, y: 0.08, w: 3.3, h: 0.72, fontSize: 9.5, color: C.mint, fontFace: "Calibri", italic: true, valign: "middle", align: "right", margin: 0
});
// ── HELPER: step card ─────────────────────────────────────────────────────────
function card(x, y, w, h, num, title, titleColor, bg, lines) {
s.addShape(pres.ShapeType.roundRect, { x, y, w, h, fill: { color: bg }, rectRadius: 0.08 });
// number pill
s.addShape(pres.ShapeType.roundRect, { x: x + 0.1, y: y + 0.1, w: 0.38, h: 0.32, fill: { color: C.navy }, rectRadius: 0.05 });
s.addText(num, { x: x + 0.1, y: y + 0.1, w: 0.38, h: 0.32, fontSize: 10, bold: true, color: C.mint, align: "center", valign: "middle", fontFace: "Calibri", margin: 0 });
// title
s.addText(title, { x: x + 0.55, y: y + 0.1, w: w - 0.65, h: 0.32, fontSize: 11, bold: true, color: titleColor, fontFace: "Calibri", margin: 0, valign: "middle" });
// body lines
s.addText(lines, { x: x + 0.12, y: y + 0.48, w: w - 0.22, h: h - 0.58, fontSize: 9.5, color: C.navy, fontFace: "Calibri", valign: "top", margin: 0 });
}
// ── HELPER: connector arrow ───────────────────────────────────────────────────
function arrowR(x, y) {
s.addShape(pres.ShapeType.rightArrow, { x, y, w: 0.28, h: 0.22, fill: { color: C.mint } });
}
function arrowD(x, y) {
s.addShape(pres.ShapeType.downArrow, { x, y, w: 0.22, h: 0.28, fill: { color: C.amber } });
}
// ═══════════════════════════════════════════════════════════════
// ROW 1 — Steps 1, 2, 3 (top row)
// ═══════════════════════════════════════════════════════════════
const ROW1_Y = 1.08;
const ROW2_Y = 4.2;
const CW = 3.8; // card width
const CH = 2.85; // card height
const GAP = 0.37;
const X1 = 0.18, X2 = X1 + CW + GAP + 0.28, X3 = X2 + CW + GAP + 0.28;
// STEP 1 – DESCRIPTIVE
card(X1, ROW1_Y, CW, CH, "01", "Descriptive Analysis", C.amber, C.light, [
{ text: "Purpose: ", options: { bold: true } },
{ text: "Summarise who is in the sample before any testing.\n\n", options: {} },
{ text: "Absolute Frequency", options: { bold: true } },
{ text: " — raw count per category\n e.g. 120 adolescents had high blood pressure\n\n", options: {} },
{ text: "Percentage (%)", options: { bold: true } },
{ text: " — count as proportion of whole\n e.g. 25% of the sample\n\n", options: {} },
{ text: "Gives a clear demographic picture of the study population.", options: { italic: true, color: C.gray } },
]);
arrowR(X1 + CW + 0.04, ROW1_Y + CH / 2 - 0.11);
// STEP 2 – GROUP COMPARISON
card(X2, ROW1_Y, CW, CH, "02", "Group Comparison Tests", C.amber, C.light, [
{ text: "Compares mean outcomes between sleep groups.\n\n", options: {} },
{ text: "Normal data", options: { bold: true, color: C.teal } },
{ text: " (bell-shaped):\n", options: {} },
{ text: " 2 groups → Student's t-test\n 3+ groups → One-Way ANOVA\n\n", options: {} },
{ text: "Skewed data", options: { bold: true, color: C.red } },
{ text: " (non-normal):\n", options: {} },
{ text: " 2 groups → Wilcoxon Test\n 3+ groups → Kruskal-Wallis Test\n\n", options: {} },
{ text: "Both sets ask the same question — just suited to different data shapes.", options: { italic: true, color: C.gray } },
]);
arrowR(X2 + CW + 0.04, ROW1_Y + CH / 2 - 0.11);
// STEP 3 – DAG
card(X3, ROW1_Y, CW, CH, "03", "Directed Acyclic Graph (DAG)", C.amber, C.light, [
{ text: "A causal diagram built from literature to identify confounders.\n\n", options: {} },
{ text: "Nodes", options: { bold: true } },
{ text: " = variables | ", options: {} },
{ text: "Arrows", options: { bold: true } },
{ text: " = causal direction | ", options: {} },
{ text: "No loops\n\n", options: { bold: true } },
{ text: "Minimum adjustment sets selected:\n", options: {} },
{ text: " Sleep & Diabetes/Dyslipidemia:", options: { bold: true } },
{ text: " age, gender, education, skin color, income, work, alcohol, smoking, physical activity, screen time, diet\n", options: {} },
{ text: " Sleep & Blood Pressure:", options: { bold: true } },
{ text: " all above + ", options: {} },
{ text: "depression", options: { bold: true, color: C.red } },
]);
// ═══════════════════════════════════════════════════════════════
// DOWN ARROW between rows
// ═══════════════════════════════════════════════════════════════
arrowD(X3 + CW / 2 - 0.11, ROW1_Y + CH + 0.05);
// ═══════════════════════════════════════════════════════════════
// ROW 2 — Steps 4, 5, 6 (bottom row, reversed right-to-left flow)
// ═══════════════════════════════════════════════════════════════
const CH2 = 2.85;
// STEP 6 – RESULTS (right bottom — flow ends here)
card(X3, ROW2_Y, CW, CH2, "06", "Effect Estimation & Gender Test", C.mint, "1B3A5C", [
{ text: "Beta Coefficient (β):", options: { bold: true, color: C.mint } },
{ text: " Change in outcome per unit change in sleep\n e.g. β = −2.3 mmHg drop per extra hour of sleep\n\n", options: { color: C.lgray } },
{ text: "95% Confidence Interval:", options: { bold: true, color: C.mint } },
{ text: " Significant if CI does NOT cross zero\n ✓ [−4.1, −0.5] vs ✗ [−2.1, 1.1]\n\n", options: { color: C.lgray } },
{ text: "Gender Interaction Test:", options: { bold: true, color: C.amber } },
{ text: " No interaction found → sleep affects boys & girls equally → sample NOT split by gender.", options: { color: C.lgray } },
]);
// left arrow pointing TO step 6 (from step 5)
arrowR(X2 + CW + 0.04, ROW2_Y + CH2 / 2 - 0.11);
// STEP 5 – IPTW
card(X2, ROW2_Y, CW, CH2, "05", "IPTW — Inverse Probability Weighting", C.mint, "1B3A5C", [
{ text: "Uses propensity scores as weights to re-balance groups — simulating a randomised trial.\n\n", options: { color: C.lgray } },
{ text: "Short sleeper, expected → ", options: { color: C.lgray } },
{ text: "LOW weight\n", options: { bold: true, color: C.amber } },
{ text: "Short sleeper, NOT expected → ", options: { color: C.lgray } },
{ text: "HIGH weight\n\n", options: { bold: true, color: C.mint } },
{ text: "Balance verified:\n", options: { bold: true, color: C.mint } },
{ text: " Standardized differences < 0.1 ✓\n Variance ratios 0.18–1.2 ✓\n\n", options: { color: C.lgray } },
{ text: "Groups are now statistically equivalent in background.", options: { italic: true, color: C.gray } },
]);
arrowR(X1 + CW + 0.04, ROW2_Y + CH2 / 2 - 0.11);
// STEP 4 – PROPENSITY SCORE
card(X1, ROW2_Y, CW, CH2, "04", "Propensity Score (Linear Regression)", C.mint, "1B3A5C", [
{ text: "Problem:", options: { bold: true, color: C.red } },
{ text: " Observational study — groups are NOT randomly assigned, so short sleepers may differ systematically from normal sleepers.\n\n", options: { color: C.lgray } },
{ text: "Solution:", options: { bold: true, color: C.mint } },
{ text: " Calculate each participant's probability of being a short sleeper, given their covariates (from the DAG).\n\n", options: { color: C.lgray } },
{ text: "Method:", options: { bold: true, color: C.amber } },
{ text: " Linear regression → one propensity score per person → used as weight in IPTW.", options: { color: C.lgray } },
]);
// ── FLOW LABEL ────────────────────────────────────────────────────────────────
s.addText("→ STEP 4 → STEP 5 → STEP 6", {
x: X1, y: ROW2_Y - 0.3, w: X3 + CW - X1, h: 0.26,
fontSize: 9, color: C.mint, fontFace: "Calibri", align: "center", italic: true
});
s.addText("STEP 1 → STEP 2 → STEP 3 ↓", {
x: X1, y: ROW1_Y - 0.22, w: X3 + CW - X1, h: 0.2,
fontSize: 9, color: C.amber, fontFace: "Calibri", align: "center", italic: true
});
// ── BOTTOM FOOTER ─────────────────────────────────────────────────────────────
s.addShape(pres.ShapeType.rect, { x: 0, y: 7.32, w: 13.3, h: 0.18, fill: { color: C.teal } });
s.addText("Descriptive → Group Tests → DAG (Confounder Selection) → Propensity Score → IPTW Balancing → β Coefficient + 95% CI", {
x: 0.2, y: 7.33, w: 12.9, h: 0.15, fontSize: 8, color: C.white, fontFace: "Calibri", align: "center"
});
// ── WRITE ─────────────────────────────────────────────────────────────────────
pres.writeFile({ fileName: "/home/daytona/workspace/stats-summary/Statistical_Methods_Single_Slide.pptx" })
.then(() => console.log("DONE"))
.catch(e => { console.error(e); process.exit(1); });
Running Command
node /home/daytona/workspace/stats-summary/create_single_slide.js
Here it is - everything on one single slide:
Statistical Methods Single Slide
Presentation · PPTX
The layout is a 2-row grid flowing like a Z-path:
[ Step 1: Describe ] → [ Step 2: Compare ] → [ Step 3: DAG ]
↓
[ Step 4: Propensity ] → [ Step 5: IPTW ] → [ Step 6: Results ]
All 6 steps are shown as cards on a wide (13.3" × 7.5") slide, with a header title bar, color-coded sections (light cards for Steps 1-3, dark cards for Steps 4-6), connecting arrows, and a summary footer strip at the bottom.
This is a shared conversation. Sign in to Orris to start your own chat.