Chapter 17: Problem Solving and Data Analysis

Welcome to Last Minute Lecture.

This free chapter overview is designed to help students review and understand key concepts.

These summaries supplement not replaced the original textbook and may not be redistributed or resold.

For complete coverage, always consult the official text.

Have you ever felt like you're trying to crack some kind of complex code, especially when it's something as high stakes as a test like the SAT?

Oh, definitely.

All those numbers, the graphs, different scenarios,

I can feel like a tangled mess.

Totally.

Well, if you've ever wished for a shortcut,

a way to untangle that information, really get it and feel empowered,

then you are definitely in the right place today.

That's right.

Our mission today is basically to become data detectives.

We're doing a deep dive into Chapter 17 of the official SAT study guide.

Yep.

The big one.

Focusing specifically on problem solving and data analysis.

And it's not just about like crunching numbers.

It's really understanding the story the data is trying to tell you.

Absolutely.

And by the end of this deep dive, you're going to have a much clearer picture, a really solid understanding of the types of questions you'll face, the core skills you need and the specific strategies to master what is honestly a huge chunk of the SAT math test.

And it really is significant.

We're talking 17 out of the 58 math questions.

That's almost 30%.

Wow.

Yeah.

30%.

Plus, maybe some good news here.

You get to use your calculator for all of these questions and it counts as one of your three math test subscores.

Okay.

That's good to know about the calculator.

What's really interesting here, I think, is that the core purpose of this whole section,

it isn't just rote math.

It's actually testing your ability to use math understanding to solve real world problems.

Okay.

It zeros in on how well you can create representations of problems, understand quantity apply statistical principles and interpret data from, you know, everyday contexts.

These are skills you'll use way beyond test day.

That makes a lot of sense.

Okay.

Let's start decoding them.

Let's get into the foundational language of numbers here.

We're diving into mastering ratios, proportions, units, and percentages.

The building blocks.

Exactly.

These aren't just abstract concepts from a textbook.

They're fundamental for how we describe relationships in the world.

Stuff like scaling a recipe or even understanding economic trends.

Indeed.

At its heart, a ratio is just a comparison, right?

A relationship between two quantities.

But the key thing to remember is it's the relationship, not the actual amounts themselves.

Right.

A proportion is just when two ratios are equal.

And a unit rate,

that tells you how many units of one thing correspond to one unit of another,

like miles per hour.

Gotcha.

And don't forget, fractions are a super effective way to represent and work with ratios.

It's kind of like finding a comet denominator for your comparison.

Okay.

So let's make this concrete.

Let's jump into an example.

It might seem simple, but it illustrates this perfectly.

Imagine 240 adults and children at a show.

The ratio of adults to children is five to one.

Problem is, how many children attended?

Right.

So a five to one ratio means you've got five parts adults for every one part children.

That totals up to six parts altogether.

Okay.

Six parts.

So to find the value of one part, you just divide the total number of people.

240 by the total parts, six.

240 divided by six.

That's 40.

Exactly.

40.

And since children represent one part, there were 40 children.

And you could double check.

Five parts adults would be five times 40, which is 200.

200 plus 40 is 240.

It works.

Makes sense.

And just a heads up, the SAT might show ratios differently, like three to one or three colon one colon one, or sometimes even just the number three.

Recognizing all those formats is pretty key.

Good tip.

Okay.

Building on that, we often see proportionality, especially with unit conversions, like an architect's drawing where maybe one inch represents three feet.

Sure.

Scale drawings.

Or those classic shadow problems.

A 12 foot tree casts an 18 foot shadow.

You might need to figure out the actual area from the scale drawing or find an unknown length based on that proportion.

And what's absolutely critical there is keeping your units consistent.

If one inch represents three feet, it's often really helpful to convert those three feet to 36 inches first.

So everything's in inches.

Yeah.

Avoid mixing them up.

It's such a common trap.

And in that shadow example,

the ratio of, say, shadow length to height L over H, which is 18 over 12 or three over two,

interestingly, that ratio itself, 32, often has no units in the problem context.

It's just describing the relationship.

That can sometimes throw people off.

Huh.

Good point.

That focus on unit consistency is great advice, though.

Avoid silly mistakes.

But sometimes you get these multi -step unit conversions, right?

It can feel a bit like a complex puzzle.

Oh, yeah.

Like this one.

The Pacific Plate moves 1 ,060 kilometers in 10 .3 million years, and you need to calculate its average speed in centimeters per year.

OK.

Yeah, that's got a few steps.

This is where really carefully tracking your units is essential.

You start with kilometers.

Right.

Convert that to meters, multiplying by 1 ,000, then convert meters to centimeters, multiplying by 100.

Got it.

Kilometers to meters, meters to centimeters.

And then you divide that whole quantity by the time 10 .3 million years.

So the calculation looks like 1 ,060 times 1 ,000 times 100 divided by 10 ,300 ,000.

OK, that's a big fraction.

It is.

But when you crunch the numbers,

you get about 10 .3 centimeters per year.

The big takeaway is just pay super close attention to your units.

Be sure the intermediate ones cancel out nicely.

It's like following a chain reaction.

Like dimensional analysis from chemistry class.

Exactly like that.

Yeah.

So how do these kind of skills play out in those student produced responses?

The grid -ins, where there are no multiple choices.

Good question.

Take population density.

Let's say county Y has two districts.

You're given the square miles and the population density for each district, plus the total area for the whole county Y.

OK.

You might need to find the overall population density for county Y itself.

Right.

For those, you have to think about building up the totals first.

You can't just average the densities usually.

Why not?

Because the districts might have different areas.

So you calculate the total population for each district first.

That's density times area.

OK.

Population for district one, population for district two.

Then you add those populations together to get the total population for county Y.

And then you divide that total population by the total area of the county.

Ah, I see.

You need the overall total population and the overall total area.

Precisely.

These questions really test if you can work through the steps without the safety net of multiple choice options.

Makes sense.

And speaking of everyday math, percentages.

We deal with these constantly sales, tips, taxes.

The SAT definitely loves to test percent increase, percent decrease, or just calculating a percentage of something.

Absolutely.

A good rule of thumb to remember is that percent increase or decrease is always the difference between the two quantities divided by the original quantity and then times 100.

The original amount is key.

So important.

Let's look at a furniture store example.

Say they usually charge 75 percent more than what they paid for a table.

The cost.

OK.

A 75 percent markup.

But during a sale, they only charge 15 percent more than cost.

Now, if the sale price is $299, what was the usual non -sale price?

Ooh.

OK.

This feels like working backwards and then forwards.

Yeah.

It's exactly.

It's like a little financial detective story.

You need to unravel it step by step.

So let's break it down.

If $299 is 15 percent more than the cost.

That means $299 is actually 115 percent of the cost or like 1 .15 times the cost.

Right.

So to find the original cost, you divide $299 by $1 .15.

Which give you $260.

That's the cost.

$260.

OK.

That's step one.

Now, the usual price is 75 percent more than that cost.

So you take the cost $260 and multiply it by 1 .75, representing 100 percent of the cost plus the 75 percent markup.

And 1 .75 times $260 is $455.

$455.

That's the usual price.

Yeah.

It's a great example of how you might have to string together a couple of percentage calculations.

Definitely test your understanding.

OK.

So we've got ratios, proportions, units, percentages down.

Let's shift gears a bit.

Let's do it.

Now we're moving into understanding how these numbers, these quantities sort of talk to each other.

This is section two.

Decoding relationships and data.

OK.

This is all about seeing how two variables relate, often shown visually in things like scatter plots, graphs, tables, or represented by equations.

The SAT math test really expects you to understand and analyze these relationships.

And also, importantly, the conditions under which a model like an equation is actually a good fit for the data.

And the properties of the functions used, too, like linear, quadratic, exponential.

Exactly.

Those are the main types you'll see.

So it sounds like we're moving from individual clues, like the ratios, to piecing together a bigger picture from the data itself.

That's a great way to think about it.

And interpreting scatter plots and lines of best fit is a perfect example.

A scatter plot just shows you visually how two variables might be related or correlated.

And the line of best fit is just that straight line drawn through the dots that best captures the overall trend.

Precisely.

Let's imagine a scatter plot showing raspberry sales.

On one axis, you have the price per pint, and on the other, the number of pints sold.

And there's a line of best fit given by the equation, say, y equals 23332x.

OK.

y equals 23332x.

So x is price, y is pint sold.

Correct.

Now, questions might ask you to predict sales based on a certain price, or maybe interpret what the slope means or what the y -intercept means.

OK.

Let's dig into that, because this is where it gets really interesting, I think.

And it reveals something super important.

If you want to predict sales for, say, $4 .50 a pint,

you just plug $4 .50 in for x in the equation.

So y equals 233 minus 32 times 4 .50.

Which comes out to 89 pints.

Simple prediction.

Right.

But what about the slope and the y -intercept?

What do they actually tell us in this context?

Great question.

The slope here is negative 32.

That means for every $1 increase in the price x, the number of pints sold y is expected to decrease by 32.

OK, that makes sense.

Higher price, fewer sales.

Pretty intuitive.

But now the y -intercept, that's the value of y when x is zero.

Here, it's 233.

So mathematically, it means that a price of $0 per pint, the model predicts 233 pints would be sold.

But wait a minute.

A price of $0?

That's unrealistic.

You wouldn't sell raspberries for free in this context.

Exactly.

And this highlights a crucial limitation of the model.

The y -intercept, while mathematically correct for the line, might not make any sense in the real world, or might fall outside the range of prices the data was actually collected for.

The store only sold them between $2, $5.

So the model might not be valid or meaningful weighed down at $0?

Correct.

Understanding these limitations, knowing when a model is useful and when it breaks down, is just as important as being able to do the calculation.

It's critical thinking about the data, not just number crunching.

That's a fantastic point about model limitations.

Really key.

It's about what the numbers mean and crucially, what they don't mean in the real world.

Okay, moving on from lines.

Let's distinguish linear versus exponential relationships.

People mix these up all the time.

Yeah, it's a key difference.

With a linear relationship, you have a constant difference added each time step.

Think constant speed, straight line on a graph.

For exponential growth, though, the quantity increases by a constant ratio or factor each time step.

It's multiplying, not adding.

Multiplying.

Got it.

Consider a table showing bacteria growth.

Maybe at time zero, you have 4 ,000 bacteria.

At time one hour, you have 1 .6 by 104, which is 16 ,000.

If you find the ratio between time one and time zero, that's 16 ,000 divided by 4 ,000, you get four.

If that ratio holds for the next hour or two, then you have exponential growth.

So the function would be something like initial amount times the ratio raised to the power of time.

Exactly.

And ti equals 4 ,000 times 4t.

That constant ratio, four in this case, is the signature of exponential models and is why exponential growth can seem so sudden or explosive, think bacteria or compound interest or even viral posts online.

Yeah, compound interest.

That's a perfect real world example of exponential relationships that hits our wallets.

Like a bank offers $1 ,000 certificates, maybe 4 % simple interest per year or 4 % compounded semi -annually.

How would you calculate the total with compounding?

This is where that famous compound interest formula comes in.

A equals P, 1 plus Rn and T.

Okay, let's break that down.

A is the final amount.

P is the principal, the initial amount, $1 ,000.

R is the annual interest rate, 4 % or 0 .04.

N is the number of times the interest is compounded per year and T is the number of years.

So if it's compounded semi -annually, our n would be 2.

Right, compounded twice a year.

So that 4 % annual rate gets applied as 2%.

Our n deal 0 .042 equals 0 .02 every half year and the exponent becomes 2 times T.

Understanding how that formula works shows you the power of compounding earning interest on your interest.

It's exponential growth in your savings account.

Very practical.

And sometimes the relationship is purely visual, right?

Like just reading a graph, say a graph showing Maria's speed versus time walking, jogging, running segments.

And the question is just identify the segment where her speed was greatest.

For that, it's usually straightforward graph reading.

You just look for the highest point on the vertical axis, the speed axis, find the corresponding time interval on the horizontal axis.

If the graph peaks at 8 miles per hour between, say, 1934 minutes, that's your answer.

So just accurately reading the visual information like looking at a car's speedometer.

Yeah, pretty much.

Quick and accurate interpretation.

All right, let's dive into our final big section then.

This is where we really put on our data detective hats.

Diving deeper into data and statistics.

Okay, the deep end of the data pool.

Yeah, this is where you really flex those analytical muscles, looking at data in tables, bar graphs, histograms, line graphs, all sorts of displays and drawing meaningful conclusions.

And a crucial part of this is probability,

often shown in two -way tables.

Probability is basically just a measure of how likely an event is to happen.

Right.

So maybe you have data from a shoplifting alarm system.

A two -way table shows when the alarm sounded or didn't sound versus whether a shoplifting attempt actually happened or not.

Okay, I can picture that table.

A typical question might be calculate the probability that a customer did not attempt to shoplift given that the alarm sounded, given that part is key.

Ah, conditional probability.

That sounds tricky.

It can seem that way, but it's about narrowing your focus.

You're not looking at all customers, only the ones where the alarm sounded.

That becomes your new total, your denominator.

Okay, so I only care about the row or column where the alarm sounded.

Exactly.

Then, within that specific group, you find the number of instances where the customer did not attempt to shoplift.

That's your numerator.

So if the table shows, say, three instances where the alarm sounded, but no shoplifting occurred, and 56 total instances where the alarm sounded.

Then the probability is three divided by 56.

Which is about 0 .0536 or roughly 5 .4%.

Precisely, you nailed it.

It's about calculating probability within a specific condition.

Super common in real world stats like medical test accuracy or risk assessment.

Got it.

Okay, beyond probability, we need to talk about describing the data itself, measures of center and spread.

The basics,

mean, median, and range.

Mean is just the average.

Median is the middle value when the data is all ordered up.

And range is simply the difference between the biggest and smallest values.

So they tell you where the data clusters and how spread out it is.

Exactly.

They give you a quick snapshot of the data set's main features.

But what about those outliers, the weird values far away from everything else?

Do they mess things up?

That's a really great question because, yes, outliers do matter and they affect things differently.

An outlier, a value much larger or smaller than the rest, can really pull the mean towards it.

Skews the average.

Exactly.

Makes it less representative of the typical value sometimes.

However, the median, being the middle value, is much less affected by outliers.

It's more robust, as statisticians say.

That's why it's important to look at both.

Think about average income versus median income.

A few billionaires can pull the average way up, but the median tells you more about what the typical person earns.

That makes perfect sense.

Let's take a histogram example.

Say it shows the number of workers versus the hours they worked per week.

You've got 40 employees total.

How do you find the mean and median hours from that?

Okay, for the mean from a histogram, you need to estimate the total hours worked.

You'd multiply the midpoint of each hour range bar by the number of workers in that bar.

Sum those up.

So hours, workers, for each bar,

add them all together.

Right.

Then divide that grand total of hours by the total number of workers, 40 in this case.

That's your mean.

Okay, and the median?

For the median, you need the middle value.

With 40 employees, the data's ordered.

The median is the average of the 20th and 21st person's hours.

You'd have to figure out which bar contains the 20th and 21st workers by adding up frequencies.

So you count in from the beginning until you hit worker 20 and 21.

Exactly.

If, say, both the 20th and 21st workers fall into the bar representing 20, 30 hours, and let's say the data points within that range lead to both being 22 hours, based on the raw data, if we had it, or an assumption within the histogram,

then the median would be 22 hours.

Got it.

Find the middle position.

Okay, one more statistical measure, standard deviation.

Sounds intimidating.

It does sound technical, but the concept is actually pretty intuitive, and that's what the SAT tests the concept, not usually the calculation itself.

Okay, so what is the concept?

Standard deviation simply measures how spread out the data values are from the mean.

Think of it as the average distance from the average.

A larger standard deviation means the data points are more spread out, more varied.

A smaller standard deviation means the data points are more tightly clustered around the mean, more consistent.

So less spread equals smaller standard deviation, more spread, larger.

You got it.

For example, imagine two dot plots showing quiz scores for Class A and Class B.

If Class A's dots are all bunched up near the average score, while Class B's dots are spread all over the place.

Then Class A has the smaller standard deviation.

Their scores are more consistent.

Exactly.

You just need to visually or conceptually grasp which data set is more spread out.

Okay, that makes sense.

This all seems to lead towards sampling and inference, right?

Using sample data to say something about a bigger group.

Precisely.

And the key distinction here is population parameter versus sample statistic.

Okay, what's the difference?

A population parameter is a number that describes the entire population you're interested in, like the true average height of all adult women in a country.

We usually don't know this exactly.

You can't measure everyone.

So instead, we take a sample, a smaller group.

A sample statistic is a number that describes that sample, like the average height of 1 ,000 randomly chosen women.

We use the sample statistic to estimate the unknown population parameter.

And that brings us to random sampling.

Why is the random part so important?

Because random selection is the cornerstone of being able to generalize your sample results back to the entire population.

If your sample isn't chosen randomly, it might be biased.

And you can't confidently say your findings apply to the whole group.

It needs to be representative.

Exactly.

Which ties directly into the idea of margin of error.

Ah, yes, margin of error.

You hear that all the time with polls.

So if a random sample of 80 students has an estimated mean internet time of 14 hours per week with a margin of error of 1 .2 hours,

what does that actually mean for the whole school, say all 1 ,200 students?

Good question.

It means we're reasonably confident that the true mean internet time for all 1 ,200 students is somewhere between 12 .8 hours, that's 14 minus 1 .2, and 15 .2 hours, 14 plus 1 .2.

So it gives you a range, a plausible range for the real population average.

Exactly, it's a confidence interval.

And two main things affect how wide that margin of error is.

First, how much variability or spread there is in the data itself.

More spread usually means a wider margin of error.

Second, the sample size.

Generally, a larger sample size leads to a smaller margin of error, giving you a more precise estimate.

Makes sense, more data equals more confidence.

But remember, and this is crucial,

the margin of error applies to the estimate of the population mean, not to any individual student in the sample.

It's about the group average.

Got it.

Okay, this brings us to a really, really key distinction.

And honestly, it feels like a common trap people fall into.

Random selection versus random assignment.

What's the difference and why is this like the critical point for drawing cause and effect conclusions?

This is honestly one of the most important concepts in this whole domain.

Random selection, like we just discussed, is about how you choose your sample from the population.

If it's random, you can generalize your findings back to that population.

Okay, selection lets you generalize.

Random assignment, on the other hand, happens after you have your sample or group of volunteers.

It's about how you assign those individuals to different treatment groups in an experiment, like who gets the real drug versus who gets the placebo.

Ah, okay, different process.

And this random assignment is what allows you to potentially draw cause and effect conclusions.

It helps ensure the groups are similar to begin with.

So any difference you see later is likely due to the treatment itself.

So let's take an example.

A community center offers a Spanish course.

Some students decide to use extra audio lessons on their own.

Can the center conclude the audio lessons caused better scores?

Not necessarily, because the students weren't randomly assigned to use the lessons or not.

The ones who chose the lessons might have been more motivated or had more free time or maybe better prior language skills.

Those other factors could explain the difference in scores.

So you can't separate the effect of the lessons from the effect of, say, motivation.

Exactly.

But if the center had randomly assigned students, you use the lessons, you don't.

Then if the lesson group scored significantly better, they could more confidently conclude the lessons caused the improvement.

Okay, so random selection generalizability.

Random assignment, cause and effect.

That's the fundamental difference.

If subjects are randomly assigned, you can talk cause and effect.

If they were randomly selected, you can generalize to the population.

If neither was random, you have to be really careful about your conclusions.

It's the difference between saying these two things are related and this thing causes that thing.

Wow, okay, that is a super important distinction.

You really can't just assume cause and effect without that random assignment piece.

All right, so we have navigated the ins and outs of problem solving and data analysis, chapter 17 of the official SAT study guide.

We covered a lot of ground.

We really did.

From basic ratios and proportions all the way through to complex stats like inference, standard deviation, interpreting graphs, we've really tried to become data detectives here.

And it's worth remembering, this domain isn't just a minor part.

It's almost a third of your SAT math score.

The real goal here, I think, isn't just memorizing formulas.

Right.

It's genuinely understanding why these concepts work, how they apply.

That's where those aha moments happen.

And that's what leads to real mastery, not just guessing.

Absolutely.

And honestly, these skills interpreting data, spotting proportionality, understanding the difference between correlation and causation, they aren't just for a test, are they?

Not at all.

They're vital for navigating the just constant stream of information we see every day.

News headlines about polls, economic reports, personal finance decisions, even figuring out if some health trend is actually backed by good evidence.

Every day critical thinking.

Exactly.

So maybe a final thought for you listening.

As you go about your day today or tomorrow, what's one piece of data you encounter, a new stat, an ad, something at work or school that you can now look at through this deep dive lens?

Yeah, ask yourself, what's the real meaning here?

Are there limitations?

Could there be cause and effect?

Or is it just correlation?

How is the data gathered?

Great questions to ask.

We really encourage you to go back, reexamine these concepts, practice using the official guide, keep sharpening those data detective skills.

Definitely.

Well, thank you so much for joining us on this deep dive into mastering the SAT math test, specifically problem solving and data analysis.

We hope you found it helpful.