SAT Data Analysis and Statistics: Interpreting Tables, Graphs, and Probability

Data analysis questions ask you to extract information from tables, compute quantities like means and medians, and make inferences from the data presented. When reading a table, first understand the row and column headers to identify what each entry represents. For a two-way frequency table showing gender and preference, the cell at row "Female" and column "Chocolate" tells you how many females prefer chocolate. Some questions ask you to calculate totals, like adding across a row to find the total number of males. Others ask for conditional probabilities or ratios. Always read the question first before working with the table, so you know which specific values or calculations you need and can focus on the relevant sections without scanning the entire table. Tables on the SAT are usually small (three or four rows and columns), so you can often answer questions by locating a single cell or adding a few numbers. Careless errors like reading the wrong row or column are common, so point to the relevant cell with your pencil or cursor before reading its value.

Scatterplots show the relationship between two variables. Each point represents an observation with values on both axes. Questions might ask you to identify the relationship (positive, negative, or no correlation), estimate values for points not explicitly marked, or determine which line of best fit most closely matches the data. Positive correlation means both variables increase together. Negative correlation means one increases while the other decreases. No correlation means the points show no clear pattern. To estimate a value, find the relevant x-value on the horizontal axis, move vertically to the line of best fit or the cloud of points, then move horizontally to the y-axis to read the approximate value. This visual process is quick and avoids calculation entirely.

The mean (average) of a dataset is the sum of all values divided by the number of values. For {2,3,5}, the mean is (2+3+5)/3=10/3≈3.33. The median is the middle value when data are ordered. For {2,3,5}, the median is 3 (the middle value). For an even number of values like {2,3,5,7}, the median is (3+5)/2=4 (average of the two middle values). The mode is the value that appears most frequently. The range is the maximum value minus the minimum value. Understanding which measure is most appropriate for describing a dataset is as important as computing the measures themselves. The median is less affected by extreme values (outliers) than the mean, so if a dataset has one very large or very small value, the median may better represent the "typical" value. Quartiles divide data into four equal parts, and the interquartile range (IQR) is the range of the middle 50% of the data. These appear on box plots, which display quartiles and outliers visually.

Standard deviation measures how spread out the data are. A small standard deviation means values are close to the mean. A large standard deviation means values are far from the mean. The SAT rarely asks you to calculate standard deviation by hand, but it may ask you to interpret which dataset has larger standard deviation by comparing the spread of values. A dataset that is clustered tightly around the mean has smaller standard deviation than one with values scattered widely. Comparing standard deviations is often easier by looking at the data visually rather than calculating. When the SAT provides standard deviation values, you can use them to estimate what range contains approximately 68%, 95%, or 99.7% of the data if it is normally distributed (the empirical rule). For instance, if the mean is 100 and standard deviation is 10, about 68% of values fall between 90 and 110.

Probability measures how likely an event is, on a scale from 0 (impossible) to 1 (certain). For a simple scenario with n equally likely outcomes, the probability of a specific outcome is 1/n. The probability of an event is (number of favorable outcomes)/(total number of possible outcomes). If a bag contains 3 red and 2 blue marbles (5 total), the probability of drawing red is 3/5. Multiple-event probabilities are calculated differently for independent and dependent events. For independent events (like two coin flips), P(A and B)=P(A)*P(B). For coin flips, P(heads and heads)=0.5*0.5=0.25. Two-way frequency tables directly show the counts needed to calculate probabilities and conditional probabilities without needing to remember formulas. If a table shows 20 males and 30 females surveyed, with 10 males preferring chocolate, the probability that a randomly selected male prefers chocolate is 10/20=0.5. Always use the table directly rather than trying to apply formulas from memory.

Relative frequency is the proportion of observations in a category, calculated as (count in category)/(total count). For a dataset of 100 people where 35 prefer coffee, the relative frequency is 35/100=0.35. This is another form of probability: if you randomly select someone from the dataset, the probability they prefer coffee is 0.35. Conditional probability answers "given that" questions. The probability that someone prefers chocolate given that they are female uses only the female row of a two-way table. From the female row, count females who prefer chocolate and divide by the total number of females. This conditional approach is more straightforward than trying to apply formula notation, especially under time pressure.

Some data analysis questions ask you to determine trends, draw conclusions, or identify which statement is best supported by the data. For trend questions, look at how values change as the x-variable increases. For a scatterplot showing years on the x-axis and temperature on the y-axis, a clear upward trend suggests rising temperatures over time. The slope of a line of best fit quantifies this trend numerically. Be cautious about cause and effect: a correlation between two variables does not mean one causes the other. Both might be caused by a third variable, or the correlation might be coincidental. When asked which statement is best supported by the data, eliminate statements that require information not present in the data or that go beyond what the data actually show. A dataset showing a correlation between ice cream sales and pool attendance does not prove that eating ice cream causes people to go to pools. It might be that warm weather causes both.

Predictions or extrapolations using a line of best fit are valid only within a reasonable range. If a line of best fit is based on data from 1990-2020, you can reasonably predict values for 2025 by extending the line slightly, but predicting for 2100 using the same line is unreliable because so many things might change. Questions ask you to identify what conclusions can be drawn with confidence (those directly supported by the data) and what conclusions require caution (those based on extrapolation or assumption). Reading the question stem carefully prevents you from selecting conclusions that overstate the data's implications. Data literacy on the SAT rewards careful reading and skepticism about claims that extend beyond what the evidence shows.