Random sampling and inference Knowledge-Based Study

Study notes

1. Understanding Populations and Samples

   In statistics, a population represents the entire group of individuals, objects, or events about which we want to draw conclusions. Because surveying or measuring an entire population is often impractical, costly, or time-consuming, researchers examine a smaller subset called a sample.

   Why Random Sampling Matters

   For a sample to provide trustworthy information about the entire population, it must be representative. A simple random sample ensures that every member of the population has an equal chance of being selected. When selection is random, it minimizes selection bias.

   • Convenience Sampling: Choosing individuals who are easiest to reach (e.g., asking only people standing near a school entrance). This often produces biased results that do not represent the broader group.
   • Voluntary Response Sampling: Allowing participants to self-select (e.g., an optional online poll). This typically overrepresents individuals with strong or extreme opinions.
   • Simple Random Sampling: Using objective tools, such as random number generators or lotteries, so that personal preferences or accessibility do not influence who is chosen.

   From Sample Statistic to Population Parameter

   A value calculated from a sample is called a statistic, whereas the true value describing the entire population is a parameter. We use the sample statistic to estimate or make an inference about the unknown population parameter.

   Proportional Reasoning

   If a representative random sample of size $n$ contains $x$ individuals with a specific trait, the sample proportion is given by:

   $$\hat{p} = \frac{x}{n}$$

   To estimate the total number $X$ with that trait in a population of total size $N$, we set up a proportion:

   $$\frac{x}{n} = \frac{X}{N} \implies X = N \times \frac{x}{n}$$

   Worked Example

   A quality control inspector inspects a random sample of 250 light bulbs from a factory batch of 10,000 bulbs and discovers that 6 are defective.

   • Sample proportion of defectives: $\frac{6}{250} = 0.024$ (or 2.4%)
   • Estimated defective bulbs in the entire batch: $10{,}000 \times 0.024 = 240$

   Because the sample was chosen randomly, 240 is a reasonable point estimate for the total number of defective bulbs in the batch.

   Sampling Variability

   If two different researchers independently select two distinct random samples of the same size from the same population, their sample statistics will rarely be identical. This natural difference among sample statistics is called sampling variability.

   The Role of Sample Size

   The size of a random sample directly influences how closely sample statistics tend to cluster around the true population parameter:

   • Small Samples: Exhibit greater variability. A few unusual values can dramatically skew the sample estimate.
   • Large Samples: Exhibit lower variability. Larger sample sizes tend to produce estimates that are more consistently close to the true population value.

   Evaluating Claims

   When assessing a statistical inference, always check:

   1. Was the sample selected randomly from the population of interest?
   2. Is the sample size large enough to avoid extreme variation?
   3. Does the conclusion apply strictly to the population from which the sample was drawn, rather than an unrelated group?