That question is a bit tough.

Consider A again: Consumer's perceptions about the frequency with which an advertisement appears are their primary consideration when evaluating an advertisement's claims about product quality.

There are a few things wrong with this answer.

First, the answer states that Kirmani would agree that the frequency of the ad is the PRIMARY concern of the consumer.

Second, Kirmani's research suggests that excessive adverstising can be considered unreasonable by the consumer and lead to consumers sensing low manufacturer confidence, but not necessarily to consumers evaluating advertisements claims about product quality.

Now consider D again.

"The frequency of advertisement is not always perceived by consumers to indicate that manufacturers are highly confident about their products' quality."

The "is not always perceived" probably threw you off. The test makers use the negative instead of the positive verbiage. "Is not always perceived" can be restated as as the opposite: "sometimes can be perceived" or "may be perceived." Also, if you are going to switch is "not always perceived" to "is sometimes perceived" you can switch "that manufacturers may not be highly confident."

In other words if you re-read D you can INFER (remember this is an inference question):

"The frequency of advertisement may be perceived by consumers to indicate that manufacturers may be less than highly confident about their products' quality."

This is in line with Kirmani's research.

Even if that seems like a lot of work, the best option is to learn to eliminate answers like A that sound really good, but when truly evaluated are out of scope and make new assumptions.

I agree that the 2 answer choices are the most compelling, however, the author indicates that the frequency of advertisement is a signal but no where in the passage is it described as the primary signal. All I am saying is that it is abit stronger for an inference question.