Tables, Caselets and Combined Data
CSIR NET Part A data interpretation: read tables, fill missing entries, index numbers, caselets, combined charts, Simpson's paradox and data sufficiency.
- 16 min
On this page · 16 sections
- Prerequisites
- The Hook: Better in Every Group, Worse Overall
- How to Read a Table
- Speed Techniques for Tables
- Missing Entries
- Index Numbers
- Caselets: Data Written as a Paragraph
- Combined Charts and Tables
- Is the Data Sufficient?
- Memory Tricks
- Don’t Confuse
- Common Mistakes
- Practice
- Quick Revision Table
- Key Takeaways
- What’s Next
Prerequisites
- Reading Bar, Line and Pie Charts: the reading routine and the six standard calculations are reused here.
- Weighted averages from Percentages, Ratios, Averages and Mixtures. The most important idea on this page, that an overall rate is a weighted average of group rates, comes straight from there.
- Chapter map: Graphical Analysis and Data Interpretation.
The Hook: Better in Every Group, Worse Overall
How to Read a Table
Tables are easier than charts to read exactly and harder to read fast. Before the question:
- Read the units line. “Figures in thousands”, “₹ crore”, “per 1,000 population” all change the answer.
- Find the totals. A total row or column lets you check readings and fill gaps.
- Note what each row and column is. Row = category, column = year is the usual layout, but not always.
- Read footnotes. “Excludes part-time students” or “provisional figures” can decide a statement question.
Worked data set
The table below is invented for this page: applications to and selections for a research fellowship, by region.
| Region | Applied 2024 | Selected 2024 | Applied 2025 | Selected 2025 |
|---|---|---|---|---|
| North | 1,200 | 96 | 1,500 | 105 |
| South | 1,600 | 160 | 1,800 | 162 |
| East | 800 | 48 | 1,000 | 70 |
| West | 1,000 | 90 | 1,100 | 110 |
| Total | 4,600 | 394 | 5,400 | 447 |
Derived quantities
Most questions ask for something that is not in the table but can be computed from two cells.
| Derived quantity | Formula | Example from the table |
|---|---|---|
| Rate (success rate) | selected ÷ applied | South 2024: \( \frac{160}{1600} = 10\% \) |
| Share of total | row ÷ total | South’s share of 2024 applicants: \( \frac{1600}{4600} \approx 34.8\% \) |
| Growth | (new − old) ÷ old | Total applicants: \( \frac{5400 - 4600}{4600} \approx 17.4\% \) |
| Ratio | one cell : another | Selected 2025, West : East \( = 110 : 70 = 11 : 7 \) |
| Average | sum ÷ count | Average selected per region, 2025: \( \frac{447}{4} = 111.75 \) |
Success rates for every region:
| Region | Rate 2024 | Rate 2025 | Change |
|---|---|---|---|
| North | 8% | 7% | −1 point |
| South | 10% | 9% | −1 point |
| East | 6% | 7% | +1 point |
| West | 9% | 10% | +1 point |
| Overall | ≈ 8.57% | ≈ 8.28% | ≈ −0.29 point |
Simpson’s paradox: the extreme case
In the fellowship table the overall rate moved with the larger regions. Weighting can do more than that: it can reverse every comparison. Here are the hook’s two protocols.
| Easy samples | Hard samples | All samples | |
|---|---|---|---|
| Protocol X | 80/100 = 80% | 10/50 = 20% | 90/150 = 60% |
| Protocol Y | 9/10 = 90% | 30/140 ≈ 21.4% | 39/150 = 26% |
Y is better within each group and worse overall, because 140 of Y’s 150 samples were hard. When a question asks which method is better, look for a hidden grouping variable and decide whether the comparison should be within groups or pooled. If the groups differ in difficulty, the within-group comparison is the fair one.
Speed Techniques for Tables
Compare fractions without dividing
To decide which region had the higher success rate in 2025, West \( \frac{110}{1100} \) or South \( \frac{162}{1800} \), cross-multiply: \( 110 \times 1800 = 198{,}000 \) and \( 162 \times 1100 = 178{,}200 \). West is higher. You did not need either percentage.
Round to two significant figures
\( \frac{447}{5400} \) is roughly \( \frac{450}{5400} = \frac{1}{12} \approx 8.3\% \). Exact: 8.28%. Options in Part A are rarely closer than that.
Use the total to check
If the regional figures you read do not add up to the total row, you have misread a cell. Thirty seconds of checking is cheaper than a negative mark.
Percentage of a percentage
“What percentage of all 2025 selections came from the East?” is \( \frac{70}{447} \approx 15.7\% \). It is not the East’s success rate (7%) and not the East’s share of applicants (\( \frac{1000}{5400} \approx 18.5\% \)). Three different percentages live in the same row; name the numerator and denominator before you divide.
Missing Entries
A table with row and column totals has redundancy. Each missing cell can be found if its row or column has only one unknown.
The table below is invented: CSIR NET qualifiers from one university, by subject and year. Letters mark missing cells.
| Subject | 2023 | 2024 | 2025 | Total |
|---|---|---|---|---|
| Physics | 40 | a | 55 | 150 |
| Chemistry | b | 60 | c | 170 |
| Biology | 70 | 80 | d | 240 |
| Total | 160 | 195 | e | 560 |
If every row and column still has two unknowns, the table cannot be completed from totals alone, and a “cannot be determined” option may be correct.
Index Numbers
An index number expresses a value relative to a base period, which is set to 100.
$$\boxed{\text{Index}_t = \frac{\text{Value}_t}{\text{Value}_{\text{base}}}\times 100} \qquad \boxed{\text{Value}_t = \text{Value}_{\text{base}}\times\frac{\text{Index}_t}{100}}$$The data below are invented: a price index for a reagent kit, base year 2020 = 100.
| Year | 2020 | 2021 | 2022 | 2023 | 2024 |
|---|---|---|---|---|---|
| Index | 100 | 108 | 120 | 126 | 135 |
| Change on previous year | — | +8% | ≈ +11.1% | +5% | ≈ +7.1% |
Caselets: Data Written as a Paragraph
A caselet gives the numbers in sentences. The fastest method is always the same: draw the table first, then answer the questions from the table.
graph LR
A[Read the paragraph] --> B[Identify the two ways of classifying]
B --> C[Draw rows x columns with a Total row and column]
C --> D[Fill every number given]
D --> E[Complete cells by subtraction]
E --> F[Check the grand total]
F --> G[Answer each question as one division]For three overlapping categories (for example, students who took courses A, B and C), a Venn diagram replaces the table. Overlaps follow inclusion–exclusion: \( n(A \cup B) = n(A) + n(B) - n(A \cap B) \), the counting form of the probability addition rule on the Counting, Probability and Number Properties page.
Combined Charts and Tables
Some questions give two displays and need both. Typical pairings:
| Display 1 | Display 2 | What you combine |
|---|---|---|
| Pie of shares | Total stated in the title | Share × total = amount |
| Table of applicants | Line of success rates | Applicants × rate = selections |
| Bar of totals by year | 100% stacked bar of composition | Total × share = count of one part |
| Table of values | Index numbers | Deflate or inflate values to a common base |
Is the Data Sufficient?
Some Part A items ask whether a conclusion follows from the data, or offer “cannot be determined” as an option. Ask two questions:
- Do I have both the numerator and the denominator? A share without a total gives no amount; a count without a population gives no rate.
- Is the data at the level the question needs? A regional total cannot tell you about one district; an annual figure cannot tell you about one month.
| Question about the fellowship table | Sufficient? | Why |
|---|---|---|
| Overall success rate in 2025 | Yes | 447 and 5,400 are both given |
| Number of women selected from the South in 2025 | No | No gender split is given |
| Whether the East’s applicants grew faster than the West’s | Yes | 25% against 10% |
| Success rate of first-time applicants | No | No split by attempt |
Memory Tricks
Don’t Confuse
| This | vs | That |
|---|---|---|
| Success rate (selected ÷ applied in that region) | share of selections (that region’s selected ÷ all selected) | |
| Rise of 15 index points | rise of \( \frac{15}{\text{starting index}}\times 100\% \) | |
| Overall rate (weighted by group size) | average of the group rates | |
| “Cannot be determined from the data” | “false” (a statement can be undecidable without being wrong) | |
| Better within every group | better overall (Simpson’s paradox can reverse it) |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper. Data sets are invented.
Level 1: Foundation
- A price index (base 2020 = 100) is 135 in 2024. The 2020 price was ₹400. What is the 2024 price?
- A row is 18% of a column total of 2,500. What is the row’s value?
- Two categories are in the ratio 7 : 9 and make up the whole. What percentage is the first?
Answers
- ₹540. \( 400 \times 1.35 \).
- 450. \( 0.18 \times 2500 \).
- 43.75%. \( \frac{7}{16}\times 100 \).
Level 2: Part A Style
Questions 1 to 4 use the table below: production of a reagent (kg) at four plants over four quarters of a year.
| Plant | Q1 | Q2 | Q3 | Q4 | Total |
|---|---|---|---|---|---|
| P | 120 | 140 | 150 | 190 | 600 |
| Q | 200 | 180 | 210 | 210 | 800 |
| R | 90 | 110 | 100 | 100 | 400 |
| S | 150 | 170 | 190 | 190 | 700 |
| Total | 560 | 600 | 650 | 690 | 2,500 |
Practice Q1. Plant P’s share of the year’s total production is:
(a) 20% (b) 24% (c) 25% (d) 30%
Answer: (b) 24%
\( \frac{600}{2500}\times 100 = 24\% \).
Practice Q2. In which quarter was the percentage increase in total production over the previous quarter the highest?
(a) Q2 (b) Q3 (c) Q4 (d) The increase was the same in all three quarters
Answer: (b) Q3
- Q2: \( \frac{40}{560} \approx 7.14\% \).
- Q3: \( \frac{50}{600} \approx 8.33\% \).
- Q4: \( \frac{40}{650} \approx 6.15\% \).
Q2 and Q4 both added 40 kg, but Q4 started from a larger base, so its percentage is smaller.
Practice Q3. The average quarterly production of plant S is:
(a) 170 kg (b) 175 kg (c) 180 kg (d) 190 kg
Answer: (b) 175 kg
\( \frac{700}{4} = 175 \) kg. Use the total column; there is no need to add the four quarters again.
Practice Q4. Which plant showed the largest percentage increase in production from Q1 to Q4?
(a) P (b) Q (c) R (d) S
Answer: (a) P
- P: \( \frac{70}{120} \approx 58.3\% \).
- Q: \( \frac{10}{200} = 5\% \).
- R: \( \frac{10}{90} \approx 11.1\% \).
- S: \( \frac{40}{150} \approx 26.7\% \).
P also has the largest absolute increase (70 kg), so here both readings agree.
Level 3: Multi-Step (Part C depth)
Practice Q5. Use the fellowship table from this page (applied and selected, by region, 2024 and 2025). Consider the statements:
- The success rate rose in exactly two regions.
- The overall success rate rose from 2024 to 2025.
- Total applications rose by more than 15%.
- Total selections rose by more than 15%.
Which statements are correct?
(a) 1 and 3 only (b) 1, 2 and 3 only (c) 3 and 4 only (d) 1, 3 and 4 only
Answer: (a) 1 and 3 only
- Rates: North 8% → 7%, South 10% → 9%, East 6% → 7%, West 9% → 10%. Exactly two rose: statement 1 is true.
- Overall: \( \frac{394}{4600} \approx 8.57\% \) to \( \frac{447}{5400} \approx 8.28\% \). It fell: statement 2 is false.
- Applications: \( \frac{5400}{4600} \approx 1.174 \), a rise of about 17.4%: statement 3 is true.
- Selections: \( \frac{447}{394} \approx 1.135 \), a rise of about 13.5%: statement 4 is false.
Check of 2 against 3 and 4: if applications grew faster than selections, the overall rate must have fallen.
Practice Q6. A survey covers 400 PhD scholars, of whom 240 are in the sciences and the rest in the humanities. 150 of the 400 hold a fellowship, and 120 of the fellowship holders are in the sciences. What percentage of humanities scholars hold a fellowship?
(a) 12.5% (b) 18.75% (c) 20% (d) 37.5%
Answer: (b) 18.75%
Build the table:
| Fellowship | No fellowship | Total | |
|---|---|---|---|
| Sciences | 120 | 120 | 240 |
| Humanities | 30 | 130 | 160 |
| Total | 150 | 250 | 400 |
Humanities with a fellowship \( = 150 - 120 = 30 \), out of 160: \( \frac{30}{160} = 18.75\% \).
Option (c) is \( \frac{30}{150} \), the humanities share of fellowship holders. Option (d) is \( \frac{150}{400} \), the overall rate. Option (a) is \( \frac{30}{240} \), the wrong group’s total.
Practice Q7. In the missing-entry table of CSIR NET qualifiers on this page (Physics, Chemistry, Biology; 2023 to 2025), what was Biology’s share of the 2025 qualifiers, to the nearest whole percent?
(a) 40% (b) 42% (c) 44% (d) 46%
Answer: (c) 44%
Fill the gaps first: \( a = 55 \), \( b = 50 \), \( c = 60 \), \( d = 90 \), so the 2025 column is \( 55 + 60 + 90 = 205 \).
Biology’s share \( = \frac{90}{205} \approx 43.9\% \), which rounds to 44%.
A candidate who takes 2025 as \( \frac{560}{3} \approx 187 \) (an average year) gets about 48%, which is not offered; one who divides by the Biology row total 240 gets 37.5%, also not offered. When none of your wrong routes match an option, that is a good sign the remaining one is right.
Quick Revision Table
| Idea | Rule |
|---|---|
| Before the question | Units line, totals, row and column meaning, footnotes |
| Rate vs share | Rate = part of its own group; share = part of the overall total |
| Overall rate | Weighted average of group rates; follows the large groups |
| Simpson’s paradox | Better in every group can be worse pooled when group sizes differ |
| Compare fractions | Cross-multiply instead of dividing |
| Missing cells | Start with a row or column that has one unknown; check the grand total |
| Index number | \( \frac{\text{value}}{\text{base value}}\times 100 \) |
| Change between two non-base years | \( \frac{\text{index difference}}{\text{starting index}}\times 100 \) |
| Caselet | Draw the table first |
| Sufficiency | Need both numerator and denominator at the right level of detail |
Key Takeaways
What’s Next
- Graphs of Functions and Scientific Plots: from tables of numbers to curves, slopes and areas.
- Histograms, Distributions, Scatter and Box Plots: grouped data, averages from a frequency table, and correlation.
- Graphical Analysis: Quick Revision.
- Back to Reading Bar, Line and Pie Charts.