The questions that are not scored

Everyone quotes 130 questions. The number on the screen will be 145.

Fifteen of them do not count toward your result. They are pretest items, which is the industry's word for questions being trialed on live candidates before they are used for real, and you get no indication of which ones they are.

Why any exam does this

A question cannot be trusted until it has been answered by a lot of people. The vendor needs to know how many candidates got it right, whether strong candidates did better on it than weak ones, and which of the wrong options actually attracted anybody. None of that can be worked out in an office.

So new items are seeded into live forms, answered by real candidates under real conditions, and analyzed afterward. The ones that behave well graduate into the scored pool on a future form. The ones that do not are rewritten or thrown away.

This is also why the passing score is reported on a scale rather than as a raw count. Different forms are not exactly equally difficult, so scores are equated onto a common scale before anyone decides whether you passed. TDI sets that passing score and does not publish it as a percentage.

What it costs you, in minutes

This is the practical consequence and it is the one nobody mentions.

Divide 150 minutes by the 130 scored questions and you get 1.15 minutes each, which is the figure we use across this site and which is our division rather than Pearson's. Divide the same 150 minutes by the 145 questions you actually sit and you get a shade over a minute.

Budget the published figure across every item you meet and you have planned a paper roughly a quarter of an hour longer than the one in front of you. That is a real amount of exam and it is where the last few questions go missing.

It changes almost nothing about how you answer

The obvious question is whether you can spot them. You cannot, and trying would cost more than it could possibly return.

Pretest items are written to look exactly like scored ones because that is the whole point of the exercise: an item behaves differently if candidates know it does not count, and the data would be worthless. Any hunch you form on the day about which questions are throwaways is a guess dressed up as insight.

So answer all 145 the same way. There is exactly one behavior this fact should change, and it is at the margin.

Where they sit in the paper

The pretest items are split the way the paper is: 10 among the general questions and 5 among the Texas ones. Proportionate, in other words, which makes sense given the vendor is trialing items for both portions of future forms.

It also means the Texas half is slightly denser than the count suggests. Thirty questions count, thirty-five appear, and since those items should be running at recall speed anyway the extra five cost you less than the ten extra general ones do.

What this does to your practice

• Train at a minute a question, not at 1.15. The buffer is what pays for the general questions that need thinking
• Sit at least one full-length form end to end, so the length is not a surprise on the day
• Expect to meet something you do not recognize. On a 145-question paper with 15 trial items, that is normal rather than a sign your preparation failed
• Do not let one strange question change your pace for the next ten

Our own mock forms are 130 questions rather than 145, and that is a deliberate choice we are slightly unsure about. Matching the scored count keeps the score meaningful and the weighting honest against the published section counts. Matching the administered count would rehearse the real clock better. We picked scoring accuracy over stamina, and a candidate who wants the harder version can simply sit a form with fifteen fewer minutes on it.

One thing we cannot tell you: whether the 15 are distributed evenly through the paper or clustered. Nobody here has sat this exam and Pearson does not publish item placement, so anyone describing where the pretest questions appear is telling you something they do not know.