FAQ on Exams

2.1. How much do LPI exams cost?

Our intention has always been to provide testing at the lowest possible cost. As of January 1, 2009 our exams will cost $160 US in most locales with the exception of our LPI-301 exam which is usually priced at $260 US. Prices may vary in some jurisdictions due to currency conversions and exchange rates.

2.2 Why do I have to pay in US$ when taking the test elsewhere?

Our exams are delivered through VUE and Prometric. They may charge their fee per exam in US$ wherever it is taken.

2.3 How many tests does it take to be certified?

The answer depends on how many levels of LPI certification you intend to pursue. There are currently 2 exams for the first level (LPIC-1) and 2 exams for the second (LPIC-2). LPIC-3 only requires one exam--however, it does have one specialty designation: "Mixed Environment" (LPI-302). A second Level 3 "specialty" on "Security" (LPI-303) will be released in February 2009.

2.4 Where can I take the exams?

Our exams are offered through the testing centers of both Pearson VUE (www.pearsonvue.com) and Prometric (www.prometric.com). Visit their web pages to find a testing center near you. Pearson VUE has over 3700 locations in 145 counties and Prometric has over 3500 locations in 120 countries, so it should be easy to find a center near you.

2.5 How can I offer the exams to my students/clients?

You will need to become either a VUE or a Prometric testing center. You may also work with a Pearson VUE or Prometric testing center near you. Local testing centers may offer tests at your facility under certain conditions. However, if there is an LPI exam proctor in your area, the LPI proctor may be able to provide exam labs with paper based tests. Please contact our operations office for more information.

2.7 Why aren't you offering exams across the web?

Unfortunately, for web based exams in unproctored situations, you cannot get around the fact that someone else could take the exam for you - or be right with you as you take the exam. At this time, there's really no way to securely offer exams online without a proctor present.

2.8 Do all LPI exams use computer-based testing?

In order to deploy our exams globally and to include the largest number of people, LPI delivers its exams using computer-based testing through Pearson VUE and Prometric. Occasionally, we do hold exam labs where paper-and-pencil exams are given (see 2.5 above). We have not ruled out the possibility that future examinations might include some different type of examination method, but that is something that will have to be evaluated at some later date.

2.11 What about the testing obstacles presented to individuals for whom English is not their native tongue?

LPI recognizes this as a demand issue. Some time ago we created Japanese and German translations of our exams and we now have Chinese, French, Portuguese and Spanish versions available for some of our exams. As you might imagine, there are many costs associated with translation and publication of additional exam forms, but if you believe there is sufficient demand for a translation to a specific language please help us to identify this need and we will consider it.

2.13 Why do the number of items vary on an exam?

Under our pilot item testing process, English-language exams administered at Prometric and VUE testing centers are extended to 120 minutes in order to allow a limited number of new pilot questions to be seeded into each exam. While scores on these new items are not included as part of candidate scores, they are used to determine item validity and quality before the items become part of the pool of official, scored exam questions. In order for the evaluation of new questions to be valid and useful, LPI cannot indicate which items on each exam are new items. In addition to this change, the exam forms for LPI exams are changed with greater frequency.

By frequently rotating exams and growing our item pool this way, we enhance the security of our exams (by impeding the abilities of cheaters and brain dumps) and thus help protect the value of LPI's certification for everyone who completes our program.

2.14 There are many questions about obscure options for commands. Why are we supposed to know them? I am an experienced Linux professional, and I still consult the man pages!

The goal of LPI's tests is to identify competent, experienced people from among unqualified people. Experienced people tend to know more of these options than inexperienced people. For example, an experienced IT professional might answer 70% of these questions correctly versus only 30% by novices. To ensure the validity of these questions during our item pilot testing process we investigate the quality of new pilot (unscored) test items. If nobody answers the item correctly (or if everybody does), the item is useless and is removed. All scored test items on our exams have characteristics that are useful in making the distinction between qualified and unqualified persons.

2.15 How are the test scores computed? How many questions do I need to answer correctly to pass? What is the maximum score? How good is my score of ...?

Initially for both levels, an Angoff (see http://www.afte.org/AssociationInfo/certification/Files/Appendix%20G.pdf for a generalized description of this method) standard setting project was completed . For LPIC-1, this project was last completed in 2003 and for LPIC-2 the project was last completed in 2005. In these standard setting projects, a number of experienced Linux system administrators were contacted to participate in the standard setting study. Average Angoff ratings were calculated for each item and these ratings were utilized to set the minimum passing score for each form. During the standard setting process and to create a passing standard, we utilized an analysis method called Rasch IRT (Item Response Theory). Rasch (http://en.wikipedia.org/wiki/Rasch_model) methods create examinee scores in a scale called logits (log-odds unit) which are especially useful for this kind of project. For forms created after these initial standard setting projects, statistical equating is completed to maintain the passing standard at the same level, so that all examinees are held to the same standard, regardless of the form they take.

Rather than providing logit measures to examinees and to create a more useful interpretation of scores, LPI provides scaled scores to examinees. LPIC-1 and LPIC-2 scaled scores are created in such a way that the minimum passing score is always 500, the standard deviation is 100 and scores generally fall between 200 and 800. Below is a table that describes a few 2005 test forms and the minimum passing scores. Please keep in mind that although these tables list the number of items scored, there are additional items that are presented on the examinations that are being pilot tested and thus are not scored.

2.16 So how good are your tests?

We feel that our exams are very good, since we apply psychometric principles throughout the exam development and scoring process and have experienced, dedicated Linux community members to assist us as subject matter experts for our projects.

In terms of statistics in item analysis we ran in late 2005 on LPIC-1 and LPIC-2 data, the overall exam reliabilities from all the forms ranged from .88 to .92, which indicates that our exams measure the Linux professional abilities of examinees consistently well. Reliability coefficients range from 0 to 1, and a reliability of .8 or higher is generally expected for high-stakes exams.

Other statistics that we scrutinize for individual items are the point-biserial statistic and a Rasch fit statistic. These statistics indicate if there is some unusual or unexpected response pattern for an item, such as when the failing examinees get an item correct more often than do the passing examinees. We use these statistics to aid our subject matter experts when revising items or in the process of selecting items for a new form. In analysis, all forms generally had no more than 5% of the items with unusual statistics, indicating that the exams and the measurement scaling are behaving as we would expect it should.

Because of these statistics and the excellent psychometric properties of our exams, we feel our exams are tough, but fair.

2.17 Aren't hands-on certification tests (like Red Hat's) better than multiple choice tests like LPI is using?

The short answer is: No. For a detailed answer, please see this article noted below:

.
2.18 How long will it take for LPI to send out my certificate once I've passed?

The total time to receive your certificate once you've been notified that you have passed an exam is between 2 to 5 weeks. It will take approximately 1-2 weeks for processing and up to an additional 3 weeks for international mail delivery. Please note, however, that it is critical that you notify LPI if you change mailing addresses during this time.

2.19 Do LPI certifications expire?

LPI's recertification policy is outlined in our policies section (see item #6).