# Read a quantitative paper critically

A ten-step reading sequence that moves from the recorded variables and descriptive patterns to the tested parameter, uncertainty, model assumptions and prediction claims. Use the linked causal and transparency readings for questions the statistical calculation cannot answer.

1. [Lecture 07 : Variables and Types of Data](https://nptel.ac.in/courses/111104120)
   Start with what the paper actually recorded, before reading its conclusion.
   Task: List each main variable, how it was obtained and the proposed interpretation.

2. [Lecture 08 : Absolute Frequency, Relative Frequency and Frequency Distribution](https://nptel.ac.in/courses/111104120)
   Check who or what contributes to each percentage.
   Task: Write the numerator, denominator and missing-data count for one reported proportion, or mark the missing information.

3. [Lecture 18 : Range, Interquartile Range and Quartile Deviation](https://nptel.ac.in/courses/111104120)
   Add spread to the reported centre.
   Task: Find a spread measure and state whether it describes individual variation or uncertainty in an estimate.

4. [Lecture 25 : Univariate and Bivariate Scatter Plots](https://nptel.ac.in/courses/111104120)
   Inspect the pattern before accepting a single association summary.
   Task: Sketch what the scatterplot would need to reveal, including unusual points or separate groups.

5. [Hypothesis testing of variance and proportions](https://nptel.ac.in/courses/103106120)
   Identify which kind of parameter the analysis tests.
   Task: Decide whether the reported test concerns a mean, a variance or a proportion; record its assumptions rather than guessing from the p-value.

6. [p value](https://nptel.ac.in/courses/103106120)
   Interpret the p-value as a conditional calculation.
   Task: Rewrite the p-value sentence without assigning a probability that the hypothesis is true.

7. [Hypothesis testing using confidence interval](https://nptel.ac.in/courses/103106120)
   Relate the test to the effect estimate and matching interval.
   Task: State which effect sizes remain compatible with the reported uncertainty under the chosen method.

8. [Statistic for linear regression](https://nptel.ac.in/courses/103106120)
   Inspect the statistical model behind a coefficient.
   Task: List the outcome, predictors and one error-model assumption; label a causal claim separately.

9. [Hypothesis testing in linear regression](https://nptel.ac.in/courses/103106120)
   Separate fitting the observed data from predicting new observations.
   Task: Identify whether an interval refers to a fitted mean or a new case, and what evidence of out-of-sample performance is available.

10. [Statistics for Hypothesis Testing - Part 1](https://nptel.ac.in/courses/103106120)
   Return to the sampling model that makes the inference meaningful.
   Task: Write a five-sentence appraisal: observation, estimate, uncertainty, assumption and unresolved reporting item. Consult TOP before claiming the result is reproducible.

Original route: Evidence Press, CC BY 4.0. Linked teaching material retains its own rights. No videos or YouTube API payloads are included.