Data collection — primary vs secondary, tally charts, frequency tables

Primary vs Secondary Data

[!definition] Primary Data

Primary data is information you collect yourself, firsthand, for a specific purpose.

• Examples: Surveying your classmates about favorite subjects, measuring plant heights in your garden, recording daily temperatures with your own thermometer
• WHY primary? You control quality, know exactly how it was gathered, and it answers YOUR specific question
• HOW to collect: Surveys, experiments, observations, measurements

[!definition] Secondary Data

Secondary data is information already collected by someone else, which you use for your analysis.

• Examples: Census data from government websites, rainfall records from meteorological departments, historical test scores from school archives
• WHY secondary? Saves time and resources; gives access to large-scale data you couldn't collect alone
• HOW to access: Published reports, databases, research papers, websites

Tally Charts

[!definition] Tally Chart

A tally chart uses marks (tallies) to count observations as you collect them in real-time. Each tally represents one occurrence. Every fifth tally crosses the previous four: |||| making groups of 5 for easy counting.

WHY Tally Charts?

1. Real-time recording: Mark tallies instantly as events happen (no need to remember)
2. Error reduction: Grouped in 5s → fewer counting mistakes
3. Visual clustering: Quick to see which category has more marks

HOW to Create a Tally Chart

Step-by-step derivation from the need:

1. Identify categories: What are you counting? (colors of cars, survey responses, dice rolls)
2. Draw two columns: Category | Tally
3. As each observation occurs, add ONE tally mark in the corresponding row
4. Every 5th mark crosses the previous 4: |||| (this is the key convention)
5. Count tallies at the end: Each |||| = 5, plus remaining marks

Frequency Tables

[!definition] Frequency Table

A frequency table summarizes data by listing each category and its frequency (count). It's the final cleaned version of a tally chart, replacing tallies with numbers.

Structure:

WHY frequencies? Numbers are faster to analyze than tallies. You can now calculate percentages, compare categories quantitatively, and spot the mode (most common value).

Deriving the Frequency Table from a Tally Chart

Given: A tally chart with grouped marks
Goal: Convert to numeric frequency

Formula (trivial but explicit): $\text{Frequency} = (\text{Number of complete groups of 5}) \times 5 + \text{Remaining tallies}$

Extending Frequency Tables: Additional Columns

Real-world frequency tables often include derived columns:

Cumulative Frequency

Cumulative frequency is the running total of frequencies.

Formula: $\text{Cumulative Freq}_i = \sum_{k=1}^{i} f_k$

WHY? Answers questions like "How many students scored up to 70?" or "What's the median position?"

Relative Frequency (Proportion)

Relative frequency converts counts to proportions or percentages.

Formula: $\text{Relative Freq} = \frac{f_i}{\sum f} \quad \text{or as percentage:} \quad \frac{f_i}{\sum f} \times 100\%$

WHY? Allows comparison across datasets of different sizes. "11 out of 23" is hard to compare with "15 out of 50", but 47.8% vs 30% is instant.

[!mistake] Common Mistakes (and why they feel right)

Mistake 1: Forgetting to cross every 5th tally

Wrong: ||||||||| (10 individual marks)
Steel-man: It feels faster to just keep drawing marks without crossing.
Why it fails: With 50+ tallies, you'll miscount. Your eyes can't reliably group random marks.
Fix: Cross every 5th mark religiously. The groups of |||| are instantly recognizable as 5.

Mistake 2: Treating secondary data as always reliable

Wrong assumption: "It's published, so it's perfect."
Steel-man: Published data feels authoritative and official.
Why it fails: Secondary data may have collection bias, outdated methods, or errors. You didn't witness collection.
Fix: Check the source's credibility, methodology, and date. Note limitations when using secondary data.

Mistake 3: Forgetting the "Total" row

Wrong: Frequency table without a sum row.
Steel-man: The individual frequencies feel complete.
Why it fails: You can't verify correctness or calculate relative frequencies without the total.
Fix: Always add a Total row: $\sum f$. Use it to check your counting and as the denominator for percentages.

Mistake 4: Mixing primary and secondary data without distinction

Wrong: Using your survey data + government census data as one dataset without labeling sources.
Steel-man: "It's all data about the same topic."
Why it fails: Different collection methods = different reliability and potential bias. Your analysis loses credibility.
Fix: Clearly label which data is primary, which is secondary, and note any methodological differences.

[!recall]- Feynman Explanation (Explain to a 12-year-old)

Imagine you're at a cricket match and want to count how many times each batsman hits a four vs a six. You could try to remember every shot, but you'd forget!

Instead, you make a tally chart: every time a four happens, you make a little mark | in the "Fours" row. After four marks, the next one crosses them |||| so you see a bundle of 5 at a glance. This is like bundling sticks — much easier to count than loose sticks everywhere.

At the end of the match, you count the bundles: 3 bundles of 5 = 15, plus 2 extra marks = 17 fours. Now you write "17" in a clean table called a frequency table. That's your final report!

Primary vs Secondary data: If YOU sat in the stadium counting, that's primary (you did the work). If you read the scorecard online that someone else made, that's secondary (you're borrowing their work). Both are useful, but knowing the difference helps you trust the right data for your project.

[!mnemonic] Memory Aids

TALLY = Take A Look, Lines Yield (groups of 5 lines give you quick totals)

Primary = Personal Recollection (you're Present for data collection)
Secondary = Someone Else's Collection (already Stored somewhere)

Frequency Table columns mnemonic: "CFRF" → Category, Frequency, Relative Freq (Common Full Reporting Format)

Connections

• Pictorial representation — bar charts and pictograms — Frequency tables are the numerical base; bar charts are the visual translation
• Measures of central tendency — mean median mode — Mode is the category with highest frequency in your table
• Grouped data and class intervals — When data has many values, you group into ranges (extending frequency tables)
• Probability from frequency — experimental vs theoretical — Relative frequency becomes experimental probability
• Sampling methods — Primary data collection must follow proper sampling to be representative

#flashcards/maths

Concept Map

Hinglish (regional understanding)

Hinglish (regional understanding)