With Inference questions, the correct answer has to fit the "must be true" standard, meaning that it has to be proven based on the passage; incorrect answers "could be true" but are not necessarily true based only on the information in the passage.

Here choice "The kitchen staff at Barry’s sometimes prepares multiple dishes on the same grill." fits that standard largely because of the phrase "under normal circumstances" at the end of the stimulus. If the prohibition on preparing multiple dishes on the same grill is different from "under normal circumstances," then it must be true that "sometimes" (note: "sometimes" is a very low bar to clear for proof) multiple dishes are prepared on the same grill. Choice "The kitchen staff at Barry’s sometimes prepares multiple dishes on the same grill." is therefore correct.

In contrast, notice the strong language within choice "Barry’s will not make special kitchen accommodations for diners who do not make a reservation.", that the restaurant categorically will not make kitchen accommodations (of any type) if a diner does not make a reservation. From the stimulus you know of one particular accommodation that will be made under a reservation, but you cannot conclude that there are no other possible accommodations, or that the restaurant wouldn't try to make that accommodation if someone were to arrive without a reservation.

Choice "Not all restaurants follow food allergy precautions to avoid cross-contamination between multiple dishes." could possibly be true ("not all" is another low bar of proof) but as this stimulus only tells you about one particular accommodation that one particular restaurant makes, you just do not have evidence to support this. (Note that while "not all" is a low bar, "food allergy precautions" is fairly broad: if every restaurant, for example, takes one small precaution like washing its dishes at high heat, that would be enough to rule out "Not all restaurants follow food allergy precautions to avoid cross-contamination between multiple dishes.".)

Choice "Limiting cross-contamination from multiple dishes on the same grill is the most effective way to avoid issues for diners with food allergies." is a classic example of an Inference answer choice simply going too far, using "the most effective" when you simply do not have information to rank different precautions.

And choice "Diners with food allergies are generally willing to be patient with longer wait times in order to avoid cross-contamination between dishes." is another example of a choice that might well be true, but does not have any proof in the stimulus.

This inference problem forces you to do some math to determine which answer must be true. You know from the given information that some preferences were different between the survey and the taste test (dark chocolate went from 27% to 60%, from the lowest value to the highest, so some people must have changed their preferences from either milk or white chocolate), but each answer choice will require some analysis to determine whether it "could be true" (incorrect answer) or "must be true" (correct).

Choice "The survey participants were generally inaccurate regarding their chocolate preferences." is the qualitative answer and certainly could be true, but isn't necessarily. What if this company simply has lousy white and milk chocolate, but very good dark chocolate? The respondents could have been very accurate in relaying their general preferences, but those preferences just didn't hold in this particular case. So choice "The survey participants were generally inaccurate regarding their chocolate preferences." is incorrect.

Choice "Some people who stated a preference for white chocolate in the survey preferred dark chocolate in the taste test." is more quantitative. It certainly could be true but doesn't have to be. You know that dark chocolate went from 27% to 60%, so it picked up a net gain of 33%. This could be true if some of that gain came from white and some from milk. But since you do not have the taste test totals from white and milk you can play with different combinations. Suppose all who said dark in the survey said dark in the taste test, and then 33% defected from milk to dark. That would leave white unchanged and still give you 60% dark, just with 28% white and now 12% milk. So choice B is not necessarily true and is therefore incorrect.

Choice "Some people who preferred milk chocolate in the taste test had initially stated a preference for white chocolate in the survey." does not have to be true, either. You know that 33% of respondents switched to dark chocolate, but you do not know for certain that anyone switched between white and milk. As you will see with choice "Some people who stated a preference for milk chocolate in the survey preferred dark chocolate in the taste test."...

Choice "Some people who stated a preference for milk chocolate in the survey preferred dark chocolate in the taste test." must be true. You need a net gain of 33% moving from either white or milk to dark. And since only 28% preferred white chocolate, you can't get that 33% gain unless at the very least 5% of people changed from milk to dark.

Choice "Most participants expressed a different preference in the taste test than they had indicated in the survey.​" is incorrect because, again, the minimum change is 33%. All the statements could be true if everyone who liked dark in the survey stuck with dark in the taste test, and then 33% moved to dark from milk. That case satisfies all of the facts but leaves more than half of survey responses intact, thereby invalidating choice "Most participants expressed a different preference in the taste test than they had indicated in the survey.​". Choice "Some people who stated a preference for milk chocolate in the survey preferred dark chocolate in the taste test." is correct.

As with any inference question, your job here is to understand the information given and to choose an answer choice guaranteed by the text. You are told in this stimulus to this question that free trials are meant to increase sales of the full version of an online service by giving users who would not buy the service without trying it first a chance to experiment with it. Choice "The number of people who see the free trial as an acceptable replacement for buying the online service is not greater than the number of people who buy the online service because of their experience within the free trial." is the only answer choice that is guaranteed by the text. If the number of people who find that the free trial was a good substitute is bigger than the number of people who are incentivized to buy the full online service because of the free trial, then the ability to experiment before you try the full service would not only be meaningless, it would be counter to the reason that companies offer free trials.

Among the other answers, choice "Because the cost of offering a free trial can be high, companies are often resistant to offering free trials, especially free trials that offer all features included within the paid version of the online service." can be eliminated because there is no information about what makes companies more or less likely to offer free trials. Choice "In calculating the total number of an online service sold, free trials are generally included as zero-dollar sales rather than as a separate category." can be eliminated because there is no information given about the spread of companies’ free versus paid sales. Choice "The number of sales for a given online service is directly proportional to the number of visitors to the online service's website, a number that tends to increase if a free trial is offered." can be eliminated because there is no information about whether the two values are directly proportional at all. Choice "Online services that are easily adapted to free trial versions sell better than do online services that are not readily distributed as free trials." can be eliminated for similar reasons to choice "Because the cost of offering a free trial can be high, companies are often resistant to offering free trials, especially free trials that offer all features included within the paid version of the online service.". there is no information about the importance of the ease of creating a free trial.

The answer to this Inference problem is "Last year there were more copies of accounting software sold to non-accountants than in any previous year.". Remember - in an Inference question the correct answer must be true based on the premises, and "Last year there were more copies of accounting software sold to non-accountants than in any previous year." can be proven by the facts. You know that 1) the total number of copies of accounting software was its greatest ever and that 2) the percentage that non-accountants purchased was its greatest ever (the first time over 50%). So non-accountants purchased their greatest-ever share of the greatest-ever total, meaning that they must have purchased their greatest number of copies of accounting software ever.

Among the incorrect answer choices:

"Last year more accounting software was sold to corporations than in any previous year." very well might be true, but cannot be proven. What if the growth in accounting software was entirely due to non-accountants (perhaps this was the first-ever year that a program like TurboTax was available, and so the non-accountant software surged while several accountants went out of business and didn't purchase anything)?

"At least some non-accountants purchased the most-purchased copy of software last year." also could be true, but you certainly cannot prove it. What if the most-sold software was a must-buy for any corporation but had no appeal to individuals?

"More non-accountants purchased accounting software last year than in any previous year." is close, but note the precision in language there: all the premises are about the number of copies sold, whereas "More non-accountants purchased accounting software last year than in any previous year." draws a conclusion about the number of purchasers. What if the number of purchasers stayed the same or even decreased, but each purchaser bought multiple different copies (maybe TurboTax came with a "add on Quicken for a dollar" promotion and almost everyone who purchased one piece of software last year bought two this year?).

And "Last year fewer copies of accounting software were purchased by accountants than in the previous year." of course does not have to be true as there is no proof for it anywhere. You know that the highest total number of copies of accounting software was sold so it is difficult to believe that fewer were sold to non-accountants, and that's the only real evidence you have to get close to this conclusion.

With any Inference question, you must select the answer choice that must be true based on the information in the passage. Here, several choices might seem very likely, but the "must be true" standard is crucial for inferences.

Choice "Daily meditation sessions of an hour or longer can increase one's life expectancy." must be true. The premises state that "while longer sessions produce better results, all of the above benefits (including a longer life) are possible from daily-ten minute sessions." From that, you can infer that longer sessions (an hour vs. ten minutes) would at least produce the same benefits, if not better. Additionally, note the easier-to-prove word "can" in "can increase one's life expectancy." This is much easier to prove than "will" or "only," words you see in other answer choices.

Among the other choices, choice "Mediation is only effective if it is performed on a daily basis." goes too far with "only." While the last sentence suggests that daily sessions are effective, the previous sentence uses "regular meditation" (so not necessarily "daily"), and ultimately there is nothing to suggest that even infrequent sessions are completely ineffective.

Choice "People who meditate for ten minutes each day will live longer than those who meditate less frequently." goes too far with the prediction "will" - for one, the argument doesn't give enough information to compare daily ten-minute sessions with, say, five-days-per-week hour-long sessions. But just as damning is the word "will" - predictions are just very hard to prove. Can you conceive of a situation in which people who meditate for ten minutes each day live shorter (too much radiation from their Headspace app?)? If so, "will" is not necessarily true.

Choice "It is possible to achieve as much of a gain in life expectancy from ten minutes of meditation per day as from less frequent meditation sessions of an hour or longer." is wrong for similar reasons as "Mediation is only effective if it is performed on a daily basis." is wrong: the hard fact is that "regular" meditation is more important than the duration of each session, but "regular" does not necessarily mean "daily" so this comparison is impossible to make without further information. For the same reason, choice "Meditating less frequently than once per day will lead to less positive benefits than meditating daily." is also incorrect.

This Inference problem demonstrates the importance of the "Must Be True" standard for GMAT inferences. With Inference problems, you want to attack the answer choices looking to exploit small flaws, and eliminate accordingly.

Choice "People whose diets are high in vitamin C are less likely to contract periodontitis than those whose diet are low in vitamin C." is too general and emphatic a conclusion. Even though vitamin C itself can help to prevent or treat these conditions, one cannot conclude that those who consume vitamin C will be less likely to contract those conditions. Consider a hypothetical: it could be that vitamin C alone would help, but that vitamin C is often present in sugar-containing foods and most people with high vitamin C levels are also guilty of a high sugar diet that leads to even quicker gum disease. Choice "People whose diets are high in vitamin C are less likely to contract periodontitis than those whose diet are low in vitamin C." may very well be true, but if you can create a hypothetical with a case in which it would not be true, you can eliminate it.

"Periodontitis is a condition only contracted by those who have previously contracted gingivitis." is similar: it seems like it's probably true, since you're told that gingivitis "can lead to periodontitis" but you don't know that it's the only thing that can lead to the condition (as choice "Some people suffering from periodontitis do so without having contracted gingivitis." suggests). Since you're unsure whether gingivitis is the only cause, or just one of multiple potential causes, you can eliminate both "Periodontitis is a condition only contracted by those who have previously contracted gingivitis." and "Some people suffering from periodontitis do so without having contracted gingivitis.".

Choice "A periodontitis treatment plan that does not include vitamin C is less effective than a plan that does." is also not proven. There may be other plans that do not include vitamin C but that are extremely effective. Note the language in the last sentence of the stimulus, that vitamin C "can help to treat periodontitis." "Can help" is soft language that leaves plenty of room for another treatment program to be even more helpful.

Choice "Increasing the amount of vitamin C in one’s diet has helped some periodontitis patients reduce the severity of that condition." is correct, in large part because of similarly soft language. If vitamin C can help treat the condition, that means that it must have helped at least some patients in treatment. That's an easy bar to get over, and since you know for a fact that vitamin C is helpful, you can clear that bar. Choice "Increasing the amount of vitamin C in one’s diet has helped some periodontitis patients reduce the severity of that condition." is correct.

The stimulus states that "...health insurance companies paid for a smaller percentage of claims last year than they did ten years ago." This means that the companies had a greater percentage of unpaid claims last year. Put another way, they had a smaller percentage of unpaid claims ten years ago, which is what choice "the percentage of health insurance claims that were unpaid was less than last year's percentage." says.

Choice "fewer people made health insurance claims than was the case last year." is incorrect because the number of people making claims is never mentioned. One person can make several claims, so we cannot draw any inferences about the number of people.

"more claims were not paid by insurance companies than were not paid last year." is wrong because the overall number of claims and the percentage of claims not paid were both lower ten years ago. Since they were both lower, their product (Total * Percentage) would also be lower, disproving the statement that more claims were not paid.

Answer "health insurance companies paid a greater percentage of their claims than they paid twenty years ago." is incorrect because we do not have any way of knowing the difference in the percentage of claims paid for these two periods (ten years ago versus twenty years ago).

Answer "profits made by health insurance companies were similar to profits made by health insurance companies last year." is incorrect since we already know that the companies have compensated for paying more real dollars by hiking their rates, but there are too many other factors involved in determining profit to make this comparison.

Whenever you are asked to make an inference from an argument presented on the GMAT, remember that inferences don't need to be interesting or surprising - they only need to be guaranteed.

The argument here states that a television network has been criticized for not giving as much time to climate change deniers as it does to those who believe in climate change even though they do give equal time to all sides of the debates surrounding other issues like tax reform and immigration. The network claims that this is because they only give equal time if one side of the debate cannot be definitively proven by science.

Since the network does not give equal time to both sides of the climate change debate, that means that people at the network believe that it fits the exception given and that one side (the side that believes in climate change) has been definitively proven by science, which matches answer choice "The television news network believes that the existence of climate change has been definitely proven by existing scientific research.".

Among the other answers, "The television news network believes that it is important to avoid debate on scientific discoveries." can be discarded because while the network does not give equal time for all sides of the debate in some cases, it does not provide a blanket dismissal of debating all scientific discovery. Choice "There are no effective counter arguments against climate change that might sway viewers to believe that climate change is not occurring." can be eliminated since while the network believes the science has been settled, this is not the same thing as claiming that no one will be swayed by the arguments against climate change. Choice "No individuals who spoke on the network and who did not believe climate change was occurring were scientists." is a bit harder to eliminate since the stimulus tells you that the people who believe that climate change is occurring are scientists, but it doesn't specifically say that the individuals who don't believe in climate change aren't scientists, so "No individuals who spoke on the network and who did not believe climate change was occurring were scientists." can be eliminated. Choice "If the news network gave the same time it gave to scientists to individuals who don’t believe in climate change, it would increase its ratings." can also be eliminated since even though some people criticize the network for its current policy, it is unknown how a change in policy would affect overall criticism of the network (and in turn how that would effect ratings).

The correct answer to this question is "It is not possible for any top computer hacker to gain access to a computer equipped with the recognition software solely by virtue of skill in replicating the structure of fingerprints.". This is an INFERENCE question, requiring the test taker to choose the correct answer that must be true based on the information provided in the stimulus. "Use of the recognition software is largely impractical due to the time it takes to record and analyze a fingerprint." is incorrect as the passage provides no information with regard to the speed of recording and analyzing the fingerprint; as such, no related conclusion can be drawn. "Computers equipped with the recognition software will soon be installed in most financial firms that deal with sensitive electronic information." is incorrect as the passage provides no information with regard to the installation of computers that possess the software in specific locations; as such, no related conclusion can be drawn. "It is not possible for any top computer hacker to gain access to a computer equipped with the recognition software solely by virtue of skill in replicating the structure of fingerprints." This is the correct answer. The passage states that the software detects more characteristics than those that the most successful hackers are able to duplicate; as such, we can conclude it would be impossible for any top hacker to gain access to a protected computer solely by replicating one of multiple characteristics analyzed by the software. "Fingerprint recognition software has taken many years and tremendous investment to develop and perfect." is incorrect as the passage provides no information with regard to the time and investment costs associated with the development of the software; as such, no related conclusion can be drawn. "The fingerprint recognition software is so sensitive that many authorized users are often denied legitimate access." is incorrect as the passage provides no information with regard to errors produced by the software; as such, no related conclusion can be drawn.

Because this is an Inference question, the degree of proof for the correct answer is that the correct answer MUST BE TRUE. Because of that:

Choice "Unless the minimum wage increases, MacDowell’s will continue to remain profitable." is incorrect because you're not told what happens if the minimum wage does not increase. This prediction is hard to make, then: suppose the minimum wage stayed flat but a disease was traced to MacDowell's ingredients or a fire burned down its top-grossing store. There are plenty of ways for profitability to be cut even if the minimum wage stays flat.

Choice "If the minimum wage increases, MacDowell’s will no longer be able to remain profitable." is incorrect because it goes too far. You know that profitability will decrease, but not that it will go away entirely.

Choice "MacDowell’s will see its profitability increase if the minimum wage does not increase." is incorrect for similar reasons to choice "Unless the minimum wage increases, MacDowell’s will continue to remain profitable.". There are plenty of factors aside from the minimum wage that could decrease profitability, so choice "MacDowell’s will see its profitability increase if the minimum wage does not increase." is not necessarily true.

Choice "MacDowell’s will be unable to maintain its current profitability if the minimum wage increases." is correct. Because you're told in the argument that, of the two options that would face MacDowell's in the event of a minimum wage increase, both will decrease profitability, you know it to be true that a wage increase will cut profitability.

Choice "If MacDowell’s sees a reduction in its profitability, that means that the minimum wage has increased." is incorrect for similar reasons to "Unless the minimum wage increases, MacDowell’s will continue to remain profitable." and "MacDowell’s will see its profitability increase if the minimum wage does not increase.". Plenty of other factors can lead to a decrease in profitability, so that decrease on its own does not allow you to infer that it was specific to a minimum wage hike.