Probability and Statistics: Praxis Algebra I 5162
The 5162 outline includes probability and statistics in its Number and Quantity; Probability and Statistics domain.
- Define the sample space, distinguish independent from mutually exclusive events, choose summaries suited to the distribution, and interpret regression with units and limits.
- State the reference group, units and limits of any model-based conclusion.
On this page15 sections
- Independence and disjoint events
- With and without replacement
- Summarize data responsibly
- Regression and residuals
- Residuals and extrapolation
- Read the event model before multiplying
- Calculate a conditional probability
- Compare theoretical and observed probability
- Interpret a regression equation with units
- Expected counts versus guaranteed outcomes
- Use boxplots and histograms for different questions
- Conditional probability uses a restricted group
- Compare center and spread together
- Interpret slope and residuals carefully
- Use a two-way table to distinguish conditional and joint probability
Independence and disjoint events
A and B are independent if the occurrence of one does not change the probability of the other; then P(A∩B)=P(A)P(B). Mutually exclusive events cannot occur together, so the intersection has probability zero. If both have positive probability, mutually exclusive events are not independent.
On one fair die, A={2,4,6} and B={5,6}. Their probabilities are 1/2 and 1/3; intersection is {6}, probability 1/6, equal to the product. Thus they are independent. Even and odd outcomes are mutually exclusive with positive probabilities, so they are not independent.
With and without replacement
A bag holds five orange and three purple counters. Drawing two orange without replacement has probability (5/8)(4/7)=5/14. With replacement it is (5/8)(5/8)=25/64 because the composition returns to its initial state before the second draw.
For small sample spaces, list outcomes or use a tree to ensure the denominator matches the experiment. If outcomes are not equally likely, favorable-count over total-count is not sufficient; use the outcomes’ actual probabilities.
Summarize data responsibly
The mean uses every value and can shift with an extreme value. The median is the middle ordered value and is more resistant. A boxplot shows median, quartiles, spread and possible outliers; a histogram reveals distribution shape. Select the summary that communicates the data rather than automatically using the mean.
If data are skewed, report center and spread in a way that is not distorted by one unusually large observation. Compare groups only when their units and collection methods are comparable.
Regression and residuals
For a model predicted minutes=12+4x, where x is miles, a 3-mile trip predicts 24 minutes. If actual time is 27 minutes, residual observed minus predicted is +3 minutes. The positive residual means the observation lies above the model prediction.
A fitted association does not establish causation. Interpret slope with its units and avoid prediction far outside the observed range. The model is an estimate, not a guarantee for each individual case.
Residuals and extrapolation
For predicted y=12+4x, if x is 3 and observed y is 27, residual is 27−24=3. A negative residual would mean the observation is below prediction. Residuals can help evaluate a model’s fit, but one residual does not explain why a point differs.
A line fitted over x-values from 2 to 8 may not be reliable at x=50. Extrapolation assumes the same relationship persists beyond observed data. State the range and limits when interpreting predictions.
Read the event model before multiplying
The multiplication rule for independent events uses the product of their separate probabilities. For dependent events, use a conditional probability for the second event. Without replacement, the composition changes after the first draw. With replacement, it resets. Write the first and second probabilities separately before multiplying.
In a data problem, identify whether values are a sample or full population, and whether the plot or summary supports the proposed conclusion. A regression line summarizes a relationship in observed data; it does not imply that changing one variable will cause a change in the other.
Calculate a conditional probability
A bag has 5 orange and 3 purple counters. Given that the first counter was orange and it is not replaced, the remaining bag has 4 orange of 7 total. Thus P(second orange | first orange)=4/7. The unconditional chance that both are orange is (5/8)(4/7)=5/14.
The condition changes the sample space. With replacement, the conditional second-draw chance stays 5/8 and the joint probability is 25/64.
Compare theoretical and observed probability
Theoretical probability follows a model of equally likely outcomes or known event probabilities. Observed frequency is based on data. In a small number of trials, observed results can differ from the model without disproving it.
If 12 heads appear in 20 flips, the relative frequency is 0.60, while a fair coin model gives probability 0.50. More trials may move the observed share, but no finite sample guarantees exactly half.
Interpret a regression equation with units
If predicted weekly cost is 18+2.5x, where x is miles, the intercept is $18 at zero miles and the slope is $2.50 per mile within the model. Check whether the zero-mile interpretation is meaningful for the sampled setting.
A fitted line is a summary. Do not extrapolate beyond the data range without support, and do not infer causation from the slope alone.
Expected counts versus guaranteed outcomes
For 100 fair coin flips, the expected number of heads is 50. That does not guarantee exactly 50; random outcomes can vary. Expected value describes a long-run average under a model, not a promise for one sample. A question about observed data should distinguish theoretical probability from the sample’s actual relative frequency. Both can be valid quantities and answer different questions.
Use boxplots and histograms for different questions
A boxplot highlights median, quartiles and potential outliers. A histogram shows the shape of the distribution and where values cluster. Neither display alone establishes why the data took that shape. When comparing groups, align axes and scales. A truncated axis can visually exaggerate differences, so inspect labels and units before describing a shift.
Conditional probability uses a restricted group
Suppose 40 students are surveyed; 24 play a sport, 15 take music, and 9 do both. Among the students who play a sport, the proportion taking music is 9/24=3/8. This is P(music given sport), not 15/40, which is the overall music proportion. The condition changes the reference group.
A two-way table makes the denominator visible. Put the condition group in the total being used, then count the intersection in the numerator. In a verbal item, mark phrases such as “among,” “given that,” or “of those who.” These signal that the probability is conditional.
Compare center and spread together
Two classes can have the same mean but very different distributions. Class A scores 70, 72, 74, 76, 78 and class B scores 50, 60, 74, 88, 98; both have mean 74, but B is much more spread out. A center alone does not describe consistency. Report a measure of spread such as range, interquartile range, or standard deviation when relevant.
The median and quartiles are resistant to extreme values compared with the mean and standard deviation. For skewed data or clear outliers, median and interquartile range can better describe a typical value and middle spread. The choice depends on the display and the question, not on a universal rule.
Interpret slope and residuals carefully
If a regression model predicts test score ŷ=62+2.4h from study hours h, the slope says the predicted score rises 2.4 points for each additional hour in the fitted model. For h=5, predicted score is 74. If the observed score is 78, the residual y−ŷ is 4 points.
A residual describes the difference between one observed value and its model prediction. Positive means the observation is above the predicted value; negative means below. A pattern in residuals may indicate that a linear model misses structure. The fitted slope alone cannot establish causation or guarantee an individual’s result.
Use a two-way table to distinguish conditional and joint probability
In a group of 50, suppose 20 take art, 18 take music, and 8 take both. The joint probability of art and music is 8/50. The probability of music given art is 8/20. The probability of either is (20+18−8)/50=30/50. Each calculation uses a different denominator or overlap rule.
Write the denominator before simplifying. “Among art students” restricts the group to 20, whereas “randomly selected from all 50” keeps the whole group. A two-way table shows both totals and the intersection and helps prevent substituting the overall proportion for a conditional one.
For a sample mean, the calculation depends on all observations: (2+4+4+6)/4=4. The median is also 4, while the range is 6−2=4. Add an extreme value such as 20 and the new mean becomes 7.2, while the median becomes 5. These changes show why a summary should be interpreted alongside the distribution and sample size.
Common questions
Are mutually exclusive events independent?
Not if both have positive probability.
What is a residual?
Observed value minus predicted value.
Does regression prove causation?
No. Association alone does not establish a cause.