Introduction to Probability: Theory vs. Reality
In the world of Statistics, we often try to predict the future. We use probability to describe how likely it is that an event will happen. But there is a big difference between what should happen on paper and what actually happens when we run an experiment. This chapter explores the relationship between Theoretical Probability and Experimental Probability, and how we can tell if a situation is "fair" or biased.
1. Theoretical Probability: What "Should" Happen
Theoretical probability is based on logic and math. We use it when we assume that all outcomes are equally likely (like a fair coin or a balanced die). It is the probability we calculate without actually carrying out an experiment.
The Formula:
\( P(\text{Event}) = \frac{\text{Number of successful outcomes}}{\text{Total number of possible outcomes}} \)
Example: If you roll a fair 6-sided die, the theoretical probability of rolling a 4 is \( \frac{1}{6} \). This is because there is only one "4" and six possible numbers in total.
Key Takeaway: Theoretical probability is the "ideal" result. It doesn't change no matter how many times you run the experiment.
2. Experimental Probability (Relative Frequency): What "Actually" Happens
Experimental probability is based on data we have already collected. In the Pearson Edexcel syllabus, this is often called Relative Frequency. We use this when we don't know the theoretical probability (for example, the probability of a drawing pin landing point-up) or when we want to test if a tool is fair.
The Formula:
\( \text{Relative Frequency} = \frac{\text{Frequency of the event}}{\text{Total number of trials}} \)
Example: If you flip a coin 10 times and get "Heads" 7 times, the experimental probability (relative frequency) of Heads is \( \frac{7}{10} \) or \( 0.7 \).
Quick Review: Remember that probabilities are always between \( 0 \) (impossible) and \( 1 \) (certain). You can write them as fractions, decimals, or percentages!
3. The Law of Large Numbers: Closing the Gap
You might notice that in the example above, the experimental probability (\( 0.7 \)) is different from the theoretical probability (\( 0.5 \)). Does this mean the coin is rigged? Not necessarily!
The Rule: As the number of trials increases, the experimental probability tends to get closer and closer to the theoretical probability. This happens as long as all variables remain random.
Did you know? If you flipped that coin 1,000 times instead of 10, you would likely find the relative frequency is much closer to \( 0.5 \). If you only do a few trials, your results are often affected by "luck" or random chance. This is why larger sample sizes lead to more reliable estimates.
4. Identifying Bias
Bias in statistics means that the outcomes are not equally likely when they should be, or the experiment is designed in a way that favors a certain result. We can use the difference between experimental and theoretical results to spot bias.
How to spot bias in an exam question:
1. Look at the theoretical probability (e.g., \( \frac{1}{6} \) for a die).
2. Calculate the relative frequency from the data provided.
3. Compare them. If the difference is very large after many trials, the tool (the die/coin/spinner) is likely biased.
Example: A spinner has 4 colors. Theoretically, each should land \( 25\% \) of the time. If after 500 spins, "Red" has landed \( 48\% \) of the time, the spinner is likely biased toward Red.
Common Mistake to Avoid: Don't claim a die is biased after only 6 rolls just because you didn't get one of each number! You need a large number of trials to make a confident claim about bias.
5. Higher Tier Focus: Commenting on Bias
(Higher Tier Only topic)
At the Higher Tier, you may be asked to comment more formally on the differences between experimental and theoretical values. This includes comparing collected data with expected frequencies.
Expected Frequency: This is how many times we expect an event to happen based on its probability.
\( \text{Expected Frequency} = P(\text{Event}) \times \text{Total number of trials} \)
Higher Tier Tip: You might be asked to compare experimental results to a Binomial Model. If the data looks significantly different from what the Binomial Model predicts (e.g., way more successes than \( np \) would suggest), you should comment that the model may not be suitable or the process might be biased. Note: You do not need to perform formal significance tests for this GCSE.
6. Summary and Quick Tips
• Theoretical Probability: Uses math/logic. \( \frac{\text{Successes}}{\text{Total Outcomes}} \).
• Experimental Probability: Uses data. \( \frac{\text{Counted Frequency}}{\text{Total Trials}} \).
• Reliability: More trials = more reliable results.
• Identifying Bias: If Experimental Probability \(\neq\) Theoretical Probability after a large number of trials, bias is likely present.
• Expected Frequency: Probability multiplied by the number of trials.
Don't worry if the numbers don't match perfectly! In statistics, we expect some variation. The key is looking for patterns that stay consistent over a long period of time.