Riverbend Coffee Co. ran a one day sampling event across four of its branches, Riverside, Downtown, Harbor, and Uptown, to see how customers responded to three new syrup flavors.
The average weekly customer traffic across the four branches was 20,000 customers, and each branch's actual weekly traffic was within 200 customers of that average.
The trial ran in three stages. First, 1 kilogram of Vanilla Bean syrup was set out for all four branches to sample from. Four hours later, 1 kilogram of Caramel Drizzle syrup was added. Four more hours after that, 2 kilograms of Hazelnut Swirl syrup was added.
At each stage, the four branches used up the entire amount of syrup provided to them.
The charts show what share of each syrup flavor was used by each of the four branches.
Before touching any of the five choices, let's get straight on what each tab is actually giving us. Tab 1, Trial Details tells us the total amount of each syrup, 1 kilogram of Vanilla Bean, 1 kilogram of Caramel Drizzle, 2 kilograms of Hazelnut Swirl. Tab 2, Usage Charts tells us what share of each of those totals a given branch used. Neither tab alone gets you a weight in kilograms, you need the total from one tab and the percentage from the other, multiplied together.
Since the question asks for the greatest of five listed values, and every one of those values needs its own multiplication, there's no shortcut that skips checking all five. Guessing based on which slice looks biggest on the chart is risky here, because the three charts represent different sized totals, a big looking slice of the smallest pie can still lose to a modest looking slice of the biggest pie.
Vanilla Bean's total is 1 kilogram. Scanning the four Vanilla Bean slices on Tab 2, Downtown's is the largest at 33 percent.
So the biggest single Vanilla Bean value is 0.33 kilograms.
This one needs two lookups added together, Harbor's Vanilla Bean share and Harbor's Caramel Drizzle share, each against its own 1 kilogram total.
0.18 plus 0.30 comes to 0.48 kilograms.
Caramel Drizzle's total is also 1 kilogram. The largest Caramel Drizzle slice belongs to Uptown, at 38 percent.
That gives 0.38 kilograms, close to choice B but still smaller.
Here's where the size of the total actually matters. Hazelnut Swirl's total isn't 1 kilogram like the other two flavors, it's 2 kilograms, since Tab 1 tells us twice as much of it was set out. Uptown's slice of that pie is 32 percent.
32 percent of a 2 kilogram total is 0.64 kilograms. Because this percentage is being taken of a bigger pot than Vanilla Bean or Caramel Drizzle, a percentage that looks unremarkable on the chart turns into the largest raw weight in the whole list.
Uptown's Vanilla Bean share is 25 percent of 1 kilogram, and its Caramel Drizzle share is 38 percent of 1 kilogram.
0.25 plus 0.38 is 0.63 kilograms. This is the closest competitor to choice D, just one hundredth of a kilogram behind it, which is exactly why eyeballing this question is dangerous. Choice E looks like it should win because it is stacking up two decent sized percentages, but choice D wins because its single percentage is being applied to a much larger total.
"Percent greater" is a comparison between two specific amounts, not a comparison of the two percentages shown on the pie chart. We need the actual kilogram values for Uptown and Downtown's Hazelnut Swirl usage first, and only then compare them using the percent change idea, where you take the difference between the new and old values and divide by the old value.
It would be tempting to just subtract the chart percentages, 32 minus 20, and call it 12 percent. That skips a step. The chart percentages are shares of the same 2 kilogram total, not percentages of each other, so subtracting them directly doesn't answer what the question is asking.
Tab 2, Usage Charts shows Uptown at 32 percent of the Hazelnut Swirl pie, and Tab 1 tells us that pie represents 2 kilograms total.
Uptown used 0.64 kilograms.
Downtown's slice of the same pie is 20 percent.
Downtown used 0.40 kilograms. This is our comparison base, the "old" value in the percent change formula, since the question asks how much greater Uptown is compared to Downtown.
Percent greater means new minus old, divided by old, times 100. Uptown, 0.64, is the new value. Downtown, 0.40, is the old value, the one we're measuring the increase against.
0.64 minus 0.40 is 0.24. Dividing 0.24 by 0.40 gives 0.6, and multiplying by 100 gives 60 percent.
All three of these statements can be checked directly against one tab or the other, no combining required. That's worth noticing on its own, since not every MSR question needs the full cross tab treatment, sometimes a single well chosen tab settles the whole thing.
"Most" means more than half. So the real question here is whether Downtown's slice plus Harbor's slice of the Vanilla Bean pie clears 50 percent. We don't even need the 1 kilogram total for this one, since we're comparing a share of the whole pie to another share of the same whole pie, the percentages already put everything on the same footing.
Downtown's share is 33 percent, and Harbor's share is 18 percent.
33 plus 18 is 51 percent, just over half. So yes, those two branches together used most of the Vanilla Bean syrup.
Correct: TrueThis one lives entirely on Tab 1, Trial Details, no chart needed. The tab states the average traffic was 20,000 customers, and each branch's actual traffic was within 200 customers of that average. So the question is just, what percent of 20,000 is 200?
200 divided by 20,000 is exactly 1 percent. The tab's own wording matches the statement precisely.
Correct: TrueThis is a direct factual check against Tab 1, which states plainly that at each stage, the four branches used up the entire amount of syrup provided to them. There's no calculation here at all, just matching the statement's wording to the tab's wording.
Correct: TrueNorthgate Transit Authority surveyed 1,800 commuters about whether they had used the night bus service in the past year. Of those surveyed, 1,200 said they had never used it.
The 1,200 who had never used the night bus were then asked whether they would consider using it in the next three months. Their responses are shown below.
| Would consider | Would not consider |
|---|---|
| 400 | 800 |
The 400 commuters who said they would consider using the night bus are called Group C, and the 800 who said they would not are called Group D.
Group C and Group D were each randomly split into 4 equal subgroups. For the next six weeks, each subgroup received a different treatment:
• Free night fares
• A printed route map mailed to their home
• Both free fares and a mailed route map
• Neither treatment
The trial covered weeknights only. Weekend ridership was excluded from all counts reported after the trial.
The chart shows, for each subgroup, the percent who rode the night bus weekly, the percent who rode occasionally, and the percent who did not ride at all during the six week trial.
"At least once" covers two of the three outcomes shown on the chart, riding weekly and riding occasionally. Only "did not ride" is excluded. So for each of Group C's four subgroups, we need to add the weekly percentage and the occasional percentage together, then apply that combined percentage to the subgroup's actual size.
Group C has 400 people, straight from the tab 1 table.
400 people split into 4 equal subgroups means each subgroup has 100 people.
Free fares: 32 percent rode weekly and 41 percent rode occasionally, 73 percent combined. Route map: 14 plus 29 is 43 percent. Both treatments: 45 plus 33 is 78 percent. Neither treatment: 9 plus 24 is 33 percent.
Applying each of those combined percentages to the 100 person subgroup gives 73, 43, 78, and 33 people.
Adding those four subgroup counts together gives the total for all of Group C.
This is a percent greater question, which means we need two actual head counts first, not just the two chart percentages. Percent greater is calculated as the difference between the new and old amounts, divided by the old amount, then multiplied by 100. Here, "both treatments" is the amount we're checking, and "free fares" is the baseline we're checking it against.
800 divided by 4 is 200 people per subgroup.
The chart shows 26 percent of the both treatments subgroup rode weekly.
That's 52 people.
The chart shows 18 percent of the free fares subgroup rode weekly.
That's 36 people, our comparison baseline.
52 is the new value, 36 is the old value we're measuring against.
52 minus 36 is 16. Dividing 16 by 36 and multiplying by 100 gives approximately 44.4 percent.
All three of these statements draw on more than one tab, which is worth noticing since it's easy to assume a chart heavy question set only needs the chart.
This statement is comparing actual head counts across two groups, not percentages, so the trap here is stopping at the chart and comparing 26 percent against 45 percent, seeing Group C's percentage is higher, and calling the statement false without ever converting to people.
Group C's both treatments subgroup has 100 people, and 45 percent of them rode weekly, which is 45 people. Group D's both treatments subgroup has 200 people, twice as many, and 26 percent of them rode weekly.
52 is greater than 45, so more people really did ride weekly in Group D's both treatments subgroup, even though its percentage was lower. The larger subgroup size more than makes up for the smaller percentage.
Correct: True"Most" means more than half of Group D's 800 people, which is more than 400. So we need Group D's full at least once total across all four subgroups, the same weekly plus occasional addition from question 1, but for Group D's percentages instead of Group C's.
Free fares: 18 plus 27 is 45 percent. Route map: 11 plus 21 is 32 percent. Both: 26 plus 35 is 61 percent. Neither: 7 plus 16 is 23 percent.
Applying those to the 200 person subgroups gives 90, 64, 122, and 46 people.
322 is less than 400, so it is not most of Group D, it's only about 40 percent.
Correct: FalseThis one doesn't need any chart reading at all, it's a direct check against a sentence on Tab 2, Trial Design, which states plainly that the trial covered weeknights only, and that weekend ridership was excluded from every count reported after the trial.
Correct: FalseFor years, Fernwood Middle School noticed that a large share of its students skipped breakfast, and staff worried this was hurting focus during first period classes. Last spring, the district's nutrition coordinator surveyed 900 students and found that 540 had not eaten a school provided breakfast at all during the previous term.
District nutrition staff proposed two competing explanations for this pattern. One group of staff argued that students skipped breakfast mainly because they doubted it would actually help their schoolwork, and that a persuasion campaign explaining the benefits would fix the problem. A second group argued that the deeper issue was simply forgetfulness and inconvenient timing, not doubt, and that practical reminders would work better than a persuasion campaign. This second group predicted that reminders alone would outperform persuasion alone if the two were ever tested side by side.
District leadership designed a small study to test both theories before committing to a district wide program. Of the 540 students who had not eaten breakfast, 180 said they would be willing to try some kind of support program, and 360 said they would not.
The 180 willing students are called Group W, and the 360 unwilling students are called Group U.
Even the strongest results from this study came with a caveat. The coordinator noted that the study ran during a single term, and that any effects seen might fade once the novelty of the new outreach wore off.
Group W and Group U were each randomly split into 3 equal arms. Over the course of Term 1 and Term 2, each arm received a different form of outreach: reminder texts sent to a student's phone before the school day, a peer buddy assigned to walk with the student to the cafeteria, or both reminder texts and a peer buddy together.
Definition. A student is counted as achieving "regular attendance" in a given term only if that student ate a school breakfast on at least 4 of the 5 school days in a typical week during that term. A student who ate breakfast less often than that, even occasionally, does not meet the regular attendance standard.
The table shows, for each arm within Group W and Group U, the percent of students who achieved regular attendance, the percent who attended occasionally but did not meet the regular standard, and the mean number of days per week the arm attended, with the standard deviation shown in parentheses.
The Term Results table gives us a percentage, not a headcount, so before that percentage means anything we need the actual size of Group U's peer buddy arm. That size isn't sitting anywhere on Tab 3 itself, it has to be built from the other two tabs.
360 students split into 3 equal arms means each arm has 120 students.
Cross referencing Group U's row with the peer buddy arm and the Term 2 column, the regular attendance figure is 30 percent.
This question is about total people, not the highest percentage. Since Group W and Group U are different sizes, sixty students per arm against a hundred twenty students per arm, the arm with the single highest percentage on the table doesn't automatically win once you're counting actual students. The only reliable approach is to convert every arm's percentage into a headcount, for both groups, and add each arm's two headcounts together before comparing.
Group W's reminder texts arm, 60 percent of 60 students. Group U's reminder texts arm, 35 percent of 120 students.
36 plus 42 is 78 students combined.
Group W's peer buddy arm shows the highest single percentage on the whole table, 85 percent of 60 students. Group U's peer buddy arm is 30 percent of 120 students.
51 plus 36 is 87 students combined, a strong number, but notice it's driven mostly by Group W's small arm hitting a very high percentage.
Group W's texts and buddy arm is 70 percent of 60 students. Group U's texts and buddy arm is 55 percent of 120 students.
42 plus 66 is 108 students combined, the largest of the three totals, even though neither individual percentage in this row was the highest on the table.
These three statements pull from three different places, a table cell, two table columns, and the argument laid out in the background passage.
"At least four of five school days" is the exact wording Tab 2, Study Arms uses to define regular attendance. So this statement is really just asking, does the regular attendance percentage for this cell clear 50 percent?
The Term 2 regular attendance figure for Group U, reminder texts, is 35 percent. The occasional attendance figure sitting right beside it is 30 percent. Adding those two together gives 65 percent, which might look like "most students attended somewhat regularly," but occasional attendance does not satisfy the four of five days standard that Tab 2 defines. Only the 35 percent regular figure counts toward this statement.
35 percent is well under half.
Correct: FalseThis statement is comparing two different columns, the mean days column and the standard deviation column, and asking whether the same single cell tops one list and bottoms the other.
Scanning the Term 2 mean days column, the highest figure is 4.6 days, belonging to Group W's peer buddy arm. Scanning the Term 2 standard deviation column, the lowest figure is 0.5, belonging to Group W's texts and buddy arm. Those are two different cells.
Correct: FalseThis statement doesn't need any arithmetic at all, it's a direct paraphrase check against Tab 1, Program Background's description of the second staff group's position.
Tab 1 states that the second group argued forgetfulness and timing, not doubt, were the real problem, and that practical reminders would work better than a persuasion campaign, and specifically that this group predicted reminders alone would outperform persuasion alone if the two were ever tested side by side. That's exactly what the statement says, just restated.
Correct: TrueLarkspur Community Clinic contacted 2,400 residents to check seasonal vaccination status. Of those contacted, 1,440 had not received the seasonal vaccine.
Those 1,440 unvaccinated residents were asked whether they would consider receiving the vaccine if the clinic followed up. Their responses are shown below.
| Would consider | Would not consider |
|---|---|
| 480 | 960 |
The 480 residents who said they would consider it are called Group R, and the 960 who said they would not are called Group S.
Group R and Group S will each be split into 3 equal arms. Each arm will receive one of three outreach methods: a home visit, a phone campaign, or a mailed voucher.
Note that the mailed voucher arm's results should be read as a proportion of residents who were actually reachable in that arm, since a portion of that arm could not be reached by mail at their listed address.
Due to the mailed voucher method's poor reach, that arm was discontinued after week two for both groups. Members of Group R's mailed voucher arm were all reassigned to the home visit method. Members of Group S's mailed voucher arm were split evenly between the home visit and phone campaign methods.
Final outcomes for the two remaining methods, home visit and phone campaign, are shown below. Each pie shows the share of that arm's final membership who were vaccinated, scheduled for a future appointment, declined the vaccine, or could not be reached at all.
The word "final" in this question is doing real work, and it's the reason this question needs all three tabs read in order, not just the two that seem to have numbers on them.
Group R has 480 residents, straight from the tab 1 table.
Tab 2 splits that into 3 equal arms.
So originally, home visit, phone campaign, and mailed voucher each had 160 residents.
Tab 3 states that Group R's entire mailed voucher arm, all 160 residents, was reassigned to home visit, and that phone campaign was left alone.
Group R's final home visit arm has 320 residents, twice its original size.
The home visit, Group R pie shows 42.5 percent vaccinated.
Just like question 1, this needs the final, revised arm sizes for both groups, and the revision affected Group R and Group S differently, which is exactly what makes this version of the question harder than it first looks.
As found in question 1, Group R's home visit arm absorbed its entire 160 person voucher arm, landing at 320 residents, and 42.5 percent of that arm was vaccinated.
136 residents.
Group S starts with 960 residents, 3 equal arms of 320 each. Its voucher arm split evenly between home visit and phone campaign, meaning half of 320, or 160, joined home visit.
Group S's final home visit arm has 480 residents. The home visit, Group S pie shows 17.5 percent vaccinated.
84 residents, our comparison baseline.
136 is the new value, 84 is the old value.
136 minus 84 is 52. Dividing 52 by 84 and multiplying by 100 gives approximately 61.9 percent.
Every statement in this grid depends on catching the 22 March revision. This is deliberate, since the whole teaching point of this set is that a later, dated memo can override an earlier one, and answer keys need to name exactly which memo each step is using.
Under Tab 2, dated 9 March alone, home visit and phone campaign were equal, 160 residents each, since all three original arms were the same size. A student who stops at Tab 2 would answer this statement false.
But Tab 3, dated 22 March, is the more recent memo, and it states Group R's entire voucher arm moved into home visit, while phone campaign was left untouched.
320 is greater than 160, so the statement holds once the most recent information is used.
Correct: True"Most" means more than half, and since vaccinated and scheduled are both shares of the same pie, we can check this directly against the percentages without needing the arm's actual size at all.
The home visit, Group S pie shows 17.5 percent vaccinated and 22.5 percent scheduled.
40 percent is under half.
Correct: FalseThis one needs the same final arm size from question 1, 320 residents, combined with the unreachable slice of the home visit, Group R pie.
That slice is 10 percent.
10 percent of 320 is exactly 32, matching the statement precisely.
Correct: TrueHalvorsen Aquaculture tracked 5,600 fish across two ponds. At the start of the trial, 3,360 of those fish were below their target weight for their age.
Those 3,360 fish were split into two groups based on how they had responded to earlier feed adjustments.
| Responsive (Group L) | Non responsive (Group M) |
|---|---|
| 1,120 | 2,240 |
Group L had responded to prior feed changes, and Group M had not.
Group L and Group M were each split into 4 equal cohorts. Each cohort received a different daily feed dose, measured in grams per fish per day:
• Cohort 1: 2.0 grams per fish per day
• Cohort 2: 3.0 grams per fish per day
• Cohort 3: 4.0 grams per fish per day
• Cohort 4: 5.0 grams per fish per day
The trial ran for eight weeks, with fish weighed weekly.
The scatterplot shows, for each cohort, the daily feed dose against the percent of that cohort that reached target weight by the end of the trial.
The line chart shows mean fish weight across the eight week trial, tracked separately by pond and by group.
The scatterplot gives us a percentage for each cohort, plotted against that cohort's feed dose. To turn that percentage into an actual fish count, we need the cohort's size, which comes from combining the other two tabs.
Group M has 2,240 fish, straight from the tab 1 table.
2,240 fish split into 4 equal cohorts means each cohort has 560 fish.
Group M's Cohort 3 point sits at a dose of 4 grams per fish per day, and its height on the vertical axis shows 45 percent reached target weight.
Notice something important about this question before doing any arithmetic. It never touches the scatterplot or the line chart at all. Tab 2, Feed Protocols gives the daily dose, and the same 560 fish cohort size from question 1 gets us the rest of the way. Tab 3 is simply not needed here, and part of building good MSR instincts is noticing when a tab you'd expect to use turns out to be irrelevant to the specific question in front of you.
Group M's Cohort 3 has 560 fish, and Tab 2 sets its dose at 4.0 grams per fish per day.
2,240 grams per day for the whole cohort.
One week is seven days, so we multiply the daily total by 7.
15,680 grams per week.
There are 1,000 grams in a kilogram, so we divide by 1,000.
These three statements draw on the scatterplot, the feed dose list, and the line chart in turn, and the second one hides the same unit mismatch that question 2 already warned us about.
"Most" means more than half of Group L's 1,120 fish, so more than 560. Since Tab 2 tells us all four of Group L's cohorts are the same size, there's a shortcut available here, we can average the four scatterplot percentages directly instead of converting each one to a fish count and adding.
Group L's four cohort percentages are 35, 55, 70, and 75.
58.75 percent is above half, so more than half of Group L reached target weight. Converting to actual fish confirms it, 98 plus 154 plus 196 plus 210 comes to 658 fish out of 1,120, comfortably over half.
Correct: TrueThis is the same gram to kilogram, day to week conversion from question 2, applied to a different cohort. Cohort 1's dose is 2.0 grams per fish per day, and Group L's Cohort 1 has 280 fish.
3.92 kilograms per week, which is under 4, not over it.
Correct: FalseThis statement is about a crossing point between two of the four lines on the growth chart, so it needs two specific weeks checked, not the whole eight week span.
At week 0, Pond B Group L starts at 172 grams, while Pond A Group M starts at 175 grams, so Pond B Group L begins lower.
At week 8, Pond B Group L ends at 302 grams, while Pond A Group M ends at 258 grams, so Pond B Group L finishes higher.
Correct: True