2. ✍️ Macros and the Art of Memory

2. ✍️ Macros and the Art of Memory#

🌀 What is a Macro?#

A macro is not just a variable—it’s a name-bound spell, a container of intent. It stores strings, numbers, lists, or expressions for later invocation.

local age 25
global study "Heart Disease Research"

When referenced, a macro doesn’t execute—it whispers its contents into the void.

  • age ⇒ 25

  • study ⇒ "Heart Disease Research"

Think: Macros = reusable echoes of thought.

❤️ Why Are Macros Sacred?#

1. 🔁 Code Reusability#

Write once, invoke forever.

local varlist "age gender income education"
summarize `varlist'
regress outcome `varlist'

2. 🕰️ Dynamic Temporal Logic#

Macros adapt to the present moment.

local today = date("`c(current_date)'", "DMY")
local filename "analysis_`today'.dta"
save "`filename'"

3. 🛡️ Error Prevention#

Macros protect you from typos and repetition.

global datapath "/Users/researcher/data"
use "$datapath/study1.dta"

4. ⚖️ Conditional Elegance#

Even logic can be laced with grace.

local n = _N
if `n' < 100 {
    display "Warning: Small sample size (`n')"
}

🧠 How to Use Macros with Discipline#

Local vs Global: Scope is Philosophy#

  • Local (local name value) → Context-bound, ephemeral

  • Global (global name value) → Environment-wide, persistent

local controls "age gender income"
local method "robust"
regress outcome treatment `controls', `method'

global project_dir "/Users/researcher/cardiology_study"

Use local unless you must go global. Treat globals like salt—not sugar.

🎯 Macro Style: Clean Code is Clean Thought#

// ✅ Good
local baseline_controls "age gender race education"
local model1_vars "treatment `baseline_controls'"
local model2_vars "`model1_vars' comorbidities"

// ❌ Bad
local x "age gender race education"

Rule of thumb: If future-you can’t read it, rewrite it now.

🧩 Components of a Macro#

1. Name — The handle#

2. Value — The contents#

3. Scope — Local or global#

4. Type — String, number, list, expression#

local filename "data_analysis"
local sample_size 1000
local mean_age = (25 + 30 + 35)/3
local covariates age gender income
local n_obs = _N

📌 Referencing Rules#

// Local
`macroname'

// Global
$macroname or ${macroname}

// Example
local city "Baltimore"
display "Welcome to `city'"     // Welcome to Baltimore

global state "Maryland"
display "State: $state"         // State: Maryland

🔍 Macro Origins: Where They Come From#

1. ✍️ Manual Definition#

local author "Dr. Smith"
global alpha 0.05

2. 📜 System Constants (c() class)#

local today "`c(current_date)'"
local version "`c(stata_version)'"

3. 📦 Command Results (r() class)#

summarize age
local mean_age = r(mean)

4. 📊 Estimation Results (e() class)#

regress income education
local rsq = e(r2)

5. 🧮 Computed On-the-Fly#

count if missing(income)
local missing_pct = (r(N)/_N) * 100

🧭 The Triumvirate: return, creturn, and ereturn#

🔁 return list — The Moment’s Echo#

“What just happened?”

tab gender, chi2
return list

Think: Transient truth—r() holds the results of the last command.

🧘 creturn list — The Introspective Core#

“Who am I? Where am I?”

creturn list

Think: System-level constants—c() is the self-knowledge of Stata.

📜 ereturn list — The Archive of Estimation#

“What did the model discover?”

regress income education
ereturn list

Think: e() is the historical ledger of your estimation journey.

🗂️ Triumvirate in Motion#

regress income education      // e()
summarize income              // r()
display "`c(username)'"      // c()

local mean = r(mean)
local rsq = e(r2)
local user = "`c(username)'"

🧠 Mnemonic: r = Recent c = Constants e = Estimation

Hierarchy of Memory:

  • r() dies with the next command

  • e() persists until the next regression

  • c() is eternal (until reboot)

🗺️ Macro Quick Reference — The Portable Codex#

// Define
local name "value"
global name "value"

// Use
`localname'
$globalname

// Capture Results
local mean = r(mean)
local user = "`c(username)'"
local rsq = e(r2)

// Inspect
macro list

// Reset
macro drop _all

🧘 Final Word#

“A macro is a mnemonic mirror—a reflection of the logic you’ll return to tomorrow. So make it readable. Make it sing.”

Let this guide be both a keyboard map and a philosophical compass as you wander deeper into the forest of Stata. The language may be rigid, but your use of it doesn’t have to be.

Contents