Sitonce
Country: US
Show exams for United States Hong Kong
Sign in

Data Interpretation: Praxis Core Academic Skills For Educators: Mathematics 5733

Updated 10 min read
Key takeaway

Read a data display by checking its title, labels, units, scale and denominator before calculating.

  • Then choose a measure or comparison that answers the question.
  • For Praxis 5733, keep counts, rates and percentages distinct, and describe a pattern without claiming that an association proves cause.
On this page10 sections
  1. Read the display before choosing an operation
  2. Compare counts with rates and percentages
  3. Choose the right measure of center
  4. Describe spread and unusual values
  5. Interpret scatterplots and association
  6. Read probability and sampling summaries carefully
  7. Avoid misleading comparisons
  8. A repeatable data-question method
  9. Check the denominator in a weighted mean
  10. Check how a graph encodes change

Read the display before choosing an operation

A table or graph gives numerical evidence, but its labels establish what the numbers mean. Read the title, categories, units, time interval, source population and scale before calculating. A value of 25 might mean 25 people, 25 dollars, 25 percent or a rate of 25 per 1,000. The same arithmetic can produce a mathematically correct answer to the wrong question if the units are ignored.

For a bar chart, inspect the category labels and vertical axis. A truncated axis can make a modest difference appear visually large. For a line graph, identify the time step between points and whether the line connects measured observations or only guides the eye. For a scatterplot, each point represents a paired observation, so the horizontal and vertical variables must be read together.

A quick data checklist is: What is being measured? In whom or what? Over what period? In which unit? Is the axis scale linear? What comparison does the prompt request? Answer these questions before using a formula. They help detect when two groups have different denominators or when a chart reports a rate rather than a raw count.

Compare counts with rates and percentages

A count reports a number of cases. A rate compares a count with an exposure, population or time interval. A percentage is a proportion multiplied by 100. These measures answer different questions. A large group can have more total events but a lower rate than a smaller group.

Imagine School A records 24 absences among 120 students, while School B records 18 among 60 students. School A has more absences as a count. Its rate is 24/120=20%; School B’s rate is 18/60=30%. If the question asks which school had the larger share of students absent, compare the rates. If it asks for the larger number of absences, compare the counts.

Percent change uses a specified baseline: (new value−original value)/original value×100%. If a monthly total rises from 80 to 92, the increase is 12 and the percent increase is 12/80=15%. Dividing by 92 instead gives a different comparison. Circle or label the original value before dividing.

Be careful with percentage points. A change from 34% to 41% is an increase of 7 percentage points and a relative increase of 7/34, or about 20.6%. If a problem asks for the difference between two rates, percentage points may be intended. If it asks how much one rate increased relative to its starting level, calculate percent change.

Choose the right measure of center

The mean is the arithmetic average: add all observations and divide by the number of observations. The median is the middle value after sorting; for an even number of observations, average the two middle values. The mode is the most frequent value. Each measure describes a different feature of a distribution.

For the data 4, 5, 5, 6, 20, the mean is 40/5=8, the median is 5 and the mode is 5. The large value 20 pulls the mean upward, while the median stays at the central position. If the question asks for the average in the arithmetic sense, use the mean; if it asks for a resistant middle value, the median can be more informative.

A weighted mean gives observations different weights. Suppose a course grade is 60% exams and 40% projects, and the scores are 80 and 95. The weighted result is 0.60(80)+0.40(95)=86. A simple average of 87.5 would incorrectly treat the two categories as equally weighted. Read the weights and confirm that they total 100% or 1.

For grouped data, a table may provide intervals and frequencies instead of individual values. An exact mean may not be recoverable from intervals alone unless the problem instructs you to use class midpoints as estimates. Do not claim more precision than the supplied information supports. Follow the requested method and identify when an answer is an estimate.

Describe spread and unusual values

The range is maximum minus minimum. In the sorted data 4, 5, 5, 6, 20, the range is 20−4=16. The range is simple but depends only on the endpoints, so it does not describe how observations are distributed between them. Other measures of spread may appear in a table or be defined by the question.

An outlier can affect the mean and range strongly. If one observation changes from 20 to 200, the mean and range rise substantially, while the median remains 5. Before deciding whether a center measure represents the data well, look at the full distribution and consider whether unusual observations are errors, valid extremes or part of a separate group.

A frequency table organizes how often values or categories occur. The total frequency is the number of observations. A relative frequency is category frequency divided by that total. To compare groups of different sizes, relative frequencies often provide a fairer comparison than raw counts, but the question must specify or support that interpretation.

Interpret scatterplots and association

A scatterplot displays paired values for two quantitative variables. A positive association means larger values of one variable tend to occur with larger values of the other; a negative association means larger values of one tend to occur with smaller values of the other. The pattern can be weak or strong and may be curved rather than linear.

Association does not establish causation. If two variables move together, a third factor may influence both, or the relationship may reflect selection, timing or chance. A scatterplot alone usually cannot determine why the pattern exists. Describe what the data show and avoid turning a relationship into a causal claim without suitable evidence.

A linear model predicts one variable from another using a line. In a model y=12+4x, the slope 4 means the predicted y increases by 4 units for each one-unit increase in x, within the model’s context. The intercept 12 is the predicted y when x=0; it may have no practical meaning if zero is outside the observed range.

For an original example, suppose a simple model predicts delivery time in minutes as 12+4x, where x is distance in miles. At 3 miles the predicted time is 24 minutes. If an observed delivery takes 27 minutes, the residual is observed minus predicted: 27−24=+3 minutes. A positive residual means the observation is above the model’s prediction, not that distance caused exactly three minutes of delay.

Use fitted models cautiously outside the observed range. Extrapolating can produce implausible predictions because the relationship may change. A line that fits nearby values does not guarantee the same trend at much larger distances, different populations or another time period. Check the data context and the limits stated in the question.

Read probability and sampling summaries carefully

A sample proportion is the number with a characteristic divided by the sample size. It estimates a population proportion only under an appropriate sampling process. A small or biased sample may not represent the broader group. If a question supplies a sample, calculate from the numbers given but do not automatically generalize to every member of a population.

If 32 of 80 survey respondents prefer option A, the sample proportion is 32/80=0.40, or 40%. This is an observed fraction within that sample. A claim that exactly 40% of an entire city prefers the option would require information about how the sample was selected and its uncertainty.

Probability statements also depend on assumptions. A theoretical probability for equally likely outcomes may differ from an observed frequency in a finite sample. Repeated trials can make the empirical proportion more stable over time under a fixed model, but a short run can vary. Distinguish the model’s probability from the sample’s observed rate.

Avoid misleading comparisons

Check whether categories overlap before adding their values. If the same person can appear in two categories, their counts cannot simply be summed as if the groups were disjoint. Check whether a displayed percentage is rounded. A total of 99% or 101% can result from rounding rather than an error or an omitted category.

Averages can hide differences between subgroups. Two groups can have the same mean but very different ranges or distribution shapes. Conversely, a combined mean can differ from the average of two subgroup means if their sample sizes differ. Weight subgroup means by the number of observations when calculating an overall mean.

For example, Group A has 10 values with mean 80, while Group B has 30 values with mean 90. The combined mean is (10×80+30×90)/40=87.5, not the unweighted average of 85. The larger group contributes more observations and therefore receives more weight. This calculation assumes the two groups measure the same quantity on the same scale.

A repeatable data-question method

First restate what the question wants: a count, difference, rate, center, spread, probability or prediction. Next copy the relevant labels and values, including units and denominator. Select the operation and calculate. Finally, interpret the result in a sentence that uses the correct unit and does not claim more than the data support.

For an answer choice, test whether it confuses count with percent, changes the denominator, reverses the comparison or gives a causal explanation for an association. For numeric entry, estimate the size first and check the final entry. These habits are especially useful when calculations are straightforward but the wording contains several categories or reference values.

The ETS 5733 companion is the authority for the current exam outline. The OpenStax statistics text provides explanations of regression and residuals. The numerical displays and model used here are original teaching examples; they do not represent real school or delivery data and are not ETS questions.

Check the denominator in a weighted mean

A combined mean must account for how many observations each group contributes. If two classes have 12 and 28 students and means of 76 and 88, the combined mean is (12×76+28×88)/40=(912+2,464)/40=84.4. Averaging 76 and 88 to get 82 gives each class equal weight, which is wrong because the second class has more observations.

Before combining statistics, confirm that both groups measure the same variable on the same scale and that the sample sizes refer to the observations behind the means. A weighted mean of incompatible quantities, such as dollars and percentages, has no clear interpretation. The denominator should be the total number of observations represented by the weighted sums.

A scatterplot and a fitted line answer different questions. The plot shows paired observations and their pattern; the line summarizes a relationship under a model. A slope is measured in y-units per x-unit, and a residual is observed y minus predicted y. Keep those units when interpreting whether a point lies above or below a prediction.

Check how a graph encodes change

A line graph can show absolute counts while a second graph shows a percentage index. A rising line does not mean an accelerating rate unless the increments themselves grow over equal time intervals. Compare the change between successive points and inspect the horizontal scale. If intervals differ, a steeper visual segment may simply cover a shorter period.

When a prompt asks for a prediction from a fitted line, substitute the stated x value and interpret the output in units. For the model y=12+4x, at x=3 the prediction is 24. If the observed value is 27, the residual is +3. The model predicts an average relationship; it does not explain the cause of that individual difference.

Common questions

When should I use a median instead of a mean?

The median can better describe a typical middle observation when data are skewed or contain extreme values. The mean uses every value and is sensitive to those extremes.

What is a residual in a regression model?

A residual is observed value minus predicted value. A positive residual means the observation is above the model prediction.

Does correlation show causation?

No. A relationship between variables does not by itself establish that one caused the other. Consider other explanations and the evidence available.