lwebb

Question Writing WorkshopStanford University Medical Education Faculty : 

Question Writing Workshop Stanford University Medical Education Faculty Lynn C. Webb, Ed.D. August 15, 2007 testing@lwebb.com

Agenda: 

Agenda Introductions Examination goals Giving away the answers 3 formats for questions Investing in future examinations The joy of feedback Questions / discussion

1. Introductions: 

1. Introductions Lynn Webb – Testing consultant Previous affiliations American Board of Psychiatry & Neurology National Board of Medical Examiners ETS (Princeton) ACT (Iowa City)

Examples of Tests: 

Examples of Tests AAN’s RITE ACP’s PRITE APA’s FOCUS AOA’s UHM ABPM AAPM&R’s SAE-R WIP CBNC ABPN Certification Recertification Subspecialty Added qualifications Non-medical exams NCLEX Solar Energy

2. Examination Goals: 

2. Examination Goals Test students’ learning Validate your teaching Prepare students for future testing Your questions can contribute to these goals, or detract from them.

Contribute to Goals: 

Contribute to Goals Questions should be aligned with the specific content Questions should be technically sound Clearly written Straight-forward, NOT tricky

Detract from Goals: 

Detract from Goals Questions don’t address teaching or learning Questions are convoluted or confusing, testing something other than the content The focus shifts to the question instead of the content

3. Giving away the answers: 

3. Giving away the answers Medical students tend to be test-wise When they don’t know an answer, they retrace your steps in writing the question to guess the correct answer Even if YOU are test-wise, blind-spots exist when it’s your material Multiple-choice questions are especially vulnerable

Let’s take a quiz: 

Let’s take a quiz Our content will be highly esoteric facts about the Alaskan Malamute Each quiz question will contain a flaw in its structure that will assist test-wise examinees. You won’t need to study the Alaskan Malamute to pass this quiz.

Boris Badenov Webb: 

Boris Badenov Webb

1. The primary function of sled dogs such as the Alaskan Malamute, Samoyed, Siberian Husky, and Eskimo Dog is to:: 

rescue drowning humans retrieve tossed objects guard property pull sleds 1. The primary function of sled dogs such as the Alaskan Malamute, Samoyed, Siberian Husky, and Eskimo Dog is to:

Test-wise examinees: 

Test-wise examinees Look for hints that the question writer overlooked between the question and the correct option Sled dogs – Pull sleds (no other options about sleds)

2. Sled dogs provided vital transport of a life-saving serum during a diphtheria epidemic in Nome, Alaska in:: 

2. Sled dogs provided vital transport of a life-saving serum during a diphtheria epidemic in Nome, Alaska in: 1492 1776 1925 1952

Several ways to approach question: 

Several ways to approach question First, eliminate options that aren’t plausible: 1492 and 1776 are too early for question content Now you have 50/50 chance between C and D. Is D plausible? Why not fly in 1952? Numeric options – when in doubt, choose C

3. Credit for the introduction of the Alaskan Malamute in the ‘lower 48’ is usually given to:: 

3. Credit for the introduction of the Alaskan Malamute in the ‘lower 48’ is usually given to: Arthur Treadwell Walden and The Seeleys Admiral Byrd and Arthur Treadwell Walden The Seeleys and Commodore Perry Arthur Treadwell Walden and Jack London

Look for repeats: 

Look for repeats Arthur Treadwell Walden and the Seeleys Admiral Byrd and Arthur Treadwell Walden The Seeleys and Commodore Perry Arthur Treadwell Walden and Jack London

More repeats?: 

More repeats? Arthur Treadwell Walden and The Seeleys Admiral Byrd and Arthur Treadwell Walden The Seeleys and Commodore Perry Arthur Treadwell Walden and Jack London

Some guys listed only once: 

Some guys listed only once Arthur Treadwell Walden and The Seeleys Admiral Byrd and Arthur Treadwell Walden The Seeleys and Commodore Perry Arthur Treadwell Walden and Jack London

Where do repeats converge?: 

Where do repeats converge? Arthur Treadwell Walden and The Seeleys Admiral Byrd and Arthur Treadwell Walden The Seeleys and Commodore Perry Arthur Treadwell Walden and Jack London

Why do test-wise examinees do that?: 

Why do test-wise examinees do that? They are retracing your steps in creating the question That’s why you shouldn’t submit your first draft It’s easy to fix this problem

4. Champion Coldfoot Oonanik, the world’s most titled dog, was how many inches at the withers?: 

4. Champion Coldfoot Oonanik, the world’s most titled dog, was how many inches at the withers? 2.75” 25.0” 27.5” 275”

Test-wise examinees compare options: 

Test-wise examinees compare options 3 options have the digits 2 – 7 – 5 One option has different digits (B), but it’s close in size to C C must be the answer (Author had the answer 27.5, then moved the decimal place for 2 distractors, then chose another plausible one near answer)

Alaskan Malamutes are known for a:: 

Alaskan Malamutes are known for a: wooly undercoat, from 1-2” in depth when the dog is in full coat early warning system concerning burglaries or robberies interest in jumping on furniture and sleeping in beds with humans eye color of blue, brown, or one blue and one brown

Which option fits with stem?: 

Which option fits with stem? Grammar issue – “a” only works with “wooly” Also, author qualified answer in A – “when the dog is in full coat” – not all year round. The other options aren’t qualified

6. Alaskan Malamutes’ tails are:: 

6. Alaskan Malamutes’ tails are: wiry short tightly curled plumed and carried over their backs most of the time

Where did author spend time?: 

Where did author spend time? Correct answer is longest Correct answer is most detailed Correct answer is qualified

7. Which of the following is char-acteristic of Alaskan Malamutes?: 

7. Which of the following is char-acteristic of Alaskan Malamutes? They don’t shed They are always strong and silent They can survive extreme cold and usually tolerate warm climates They are always ‘one man’ dogs

Where did author spend time?: 

Where did author spend time? Correct answer is longest Correct answer is qualified Other answers include absolutes

Could these clues exist in your examinations?: 

Could these clues exist in your examinations?

Thrombotic microangiopathies are associated with:: 

Thrombotic microangiopathies are associated with: Thrombocytopenia Reduced coagulation factors Antibodies to glomerula basement membranes Anti-neutrophilic cytoplasmic antibodies Necrotizing vasculitis

A patient carefully reads the labels on all her food products and calculates that she is taking in about 3.5 grams of sodium. This is closest to:: 

A patient carefully reads the labels on all her food products and calculates that she is taking in about 3.5 grams of sodium. This is closest to: 60 mEq of sodium 100 mEq of sodium 160 mEq of sodium 200 mEq of sodium

Numeric options: 

Numeric options You may use C as the correct answer It shouldn’t ALWAYS be the correct answer

4. 3 formats for questions: 

4. 3 formats for questions Let’s talk about the 3 formats that you use for the test questions: Multiple choice questions (MCQs) True/False Matching

Multiple-choice questions: 

Multiple-choice questions The examinee’s task is to choose the one best answer of 5 options (A-E)

Multiple Choice: 

Multiple Choice Idea Correct answer Distractors

Multiple Choice Qs : Ideas: 

Multiple Choice Qs : Ideas The idea for the question is sometimes called the ‘teaching point’ You need questions for each lecture hour you present Test essential concepts

Source of ideas: 

Source of ideas Essential content Questions asked during lecture Conversations overheard Any points upon which you will build

How to start?: 

How to start? Review syllabus notes What are the most essential concepts in the lecture? The purpose of this lecture is to teach _______________________ You might have 6 concepts, or you might have 2 concepts and some sub-topics

Sampling: 

Sampling Testing is always a sampling of knowledge Even the USMLE samples knowledge If you make a list of 12 things that MUST be tested from a lecture, perhaps ½ of them could be used now and ½ of them could be used later. (More on this in Topic 5 – Investing in the Future)

Characteristics of GREAT questions: 

Characteristics of GREAT questions Clearly written Examinee knows what is being asked before reading the options All the information needed to answer is in the question (options don’t build upon each other) Simplest phrasing possible No extraneous information

MCQs: Correct answer: 

MCQs: Correct answer Each question should have one answer that is best When you’re writing, the best sequence is Teaching point (What you want to ensure they know) Correct answer Distractors

Other Sequence: 

Other Sequence (Not recommended) Vague concept List of options Choose one to be correct OR revise some options to be either correct or incorrect

Example: FK506 (Tacrolimus):: 

Example: FK506 (Tacrolimus): is an anti-proliferative drug binds and inhibits the stimulation of a calcium activated phosphatase, thereby blocking the induction of the early response genes in the immune response is less commonly used than cyclosporin because of its lower potency is preferred over cyclosporin since it has no nephrotoxicity is most commonly used to treat auto-immune disorders

If you didn’t know anything about FK506….: 

If you didn’t know anything about FK506…. Which option would you choose Hint: retrace the steps of the question writer

FK506 (Tacrolimus):: 

FK506 (Tacrolimus): is an anti-proliferative drug binds and inhibits the stimulation of a calcium activated phosphatase, thereby blocking the induction of the early response genes in the immune response is less commonly used than cyclosporin because of its lower potency is preferred over cyclosporin since it has no nephrotoxicity is most commonly used to treat auto-immune disorders

What to do? : 

What to do? Choose a specific teaching point from the lecture notes – what is it that you want to be sure the students know about FK506? Form that teaching point into a question Write the correct answer Write the distractors

Multiple Choice Qs: Distractors: 

Multiple Choice Qs: Distractors The distractors should be: - plausible - less correct than the answer They are called ‘distractors’ because their intent is to DISTRACT examinees who do not know the concept

Sources for Distractors: 

Sources for Distractors Common misconceptions (“Why do some of the students think that ___________ “) Logical thinking, but it doesn’t apply in this case

FK506 (Tacrolimus) binds and inhibits the stimulation of:: 

FK506 (Tacrolimus) binds and inhibits the stimulation of: calcium activated phosphatase

Better format: 

Better format But is it an essential teaching point?

(or) What happens when FK506 (Tacrolimus) binds and inhibits the stimulation of calcium activated phosphatase?: 

(or) What happens when FK506 (Tacrolimus) binds and inhibits the stimulation of calcium activated phosphatase? The induction of the early response genes in the immune response is blocked

Your item here?: 

Your item here?

Why change format?: 

Why change format? When questions are written clearly, it’s more likely that responses will reflect examinees’ knowledge on this topic, rather than their care in reading the question Typical of the type of questions they will see on USMLE and specialty certification examinations

USMLE & others do NOT use:: 

USMLE & others do NOT use: Which of the following is TRUE? Which of the following is FALSE? Each of the following statements about X is true EXCEPT:

Which one is true (or false): 

Which one is true (or false) Is not a multiple-choice question format It’s a multiple true/false format Options tend to be heterogeneous (mixing apples and oranges) because the teaching point isn’t focused

True/False Questions: 

True/False Questions + Much easier/faster to write - Less information about what the students know They have a 50/50 chance of answering correctly, just by answering (MCQs were only 20% chance)

When to write a T/F question?: 

When to write a T/F question? When the content is essential, but shallow When the multiple choice question didn’t work

FK506 (Tacrolimus):: 

FK506 (Tacrolimus): is an anti-proliferative drug binds and inhibits the stimulation of a calcium activated phosphatase, thereby blocking the induction of the early response genes in the immune response is less commonly used than cyclosporin because of its lower potency is preferred over cyclosporin since it has no nephrotoxicity is most commonly used to treat auto-immune disorders

True/False Qs: Ideas: 

True/False Qs: Ideas (same as Multiple Choice Qs) Essential content Questions asked during lecture Conversations overheard Any points upon which you will build

True/False Qs: Correct answer: 

True/False Qs: Correct answer Be clear Don’t be tricky If something is generally true (say in 999 of 1,000 cases) do you want them to say TRUE for the 999 cases or FALSE for the 1 case?

Examples: 

Examples Menarche is usually the first sign of puberty in females Endometrial adenocarcinoma typically presents as high stage disease.

Great material for T/F: 

Great material for T/F Can be found in unfocused multiple choice questions

Which statement is true about Bladder carcinoma?: 

Which statement is true about Bladder carcinoma? Bladder carcinoma is not associated with smoking Recurrence and progression are very rare after local excision of low grade papillary carcinomas “Field effect” refers to the concept that toxins in wheat grain pesticides may cause prostate carcinoma Bladder carcinoma is usually adeno-carcinoma Squamous cell carcinoma is associated with chronic irritative conditions such as chronic catheterization

Any questions about the True/False format?: 

Any questions about the True/False format?

Matching: 

Matching A series of 2 or more questions that use the same set of up to 10 possible answers (listed once as a heading). The answers may be used once, more than once, or not at all There is only ONE correct answer per question

Matching: Ideas: 

Matching: Ideas Essential content Several concepts that are alike – typically testing recognition more than application of knowledge Lists of diseases, microorganisms, medications, etc.

Matching: Correct answers: 

Matching: Correct answers This format requires more thought about correct answers After preparing a matching set, review each description again (left side) comparing to each of the options on the right. You must be sure that only ONE is the best answer

Matching: Distractors: 

Matching: Distractors There are typically fewer distractors to generate for Matching than for MCQs, because the other correct answers are included in the distractor list Additional distractors that are added should be common misconceptions about descriptions on left or plausible-sounding alternatives from within the category of things listed.

Any questions about the matching format?: 

Any questions about the matching format?

5. Investing in Future Examinations: 

5. Investing in Future Examinations You might decide to work ahead for future examinations when you’re working on your assigned questions. Work ahead when you have an abundance of ideas “Clone” your questions

Cloning Multiple Choice Qs: 

Cloning Multiple Choice Qs Especially easy to do when starting from a question that had a negative context (NOT, EXCEPT, etc.)

Which one of the following obstetric risks is NOT commonly associated with mullerian anomalies?: 

Which one of the following obstetric risks is NOT commonly associated with mullerian anomalies? Intrauterine growth restriction (IUGR) Pre-eclampsia Malpresentation Preterm labor

4 options: 

4 options 3 are associated with mullerian anomalies 1 is NOT associated with mullerian anomalies

Create a positive item: 

Create a positive item Which of the following obstetric risks is commonly associated with mullerian anomalies? Pre-eclampsia (now a distractor, instead of correct answer) Intrauterine growth restriction (IUGR) (now the correct answer) (add distractor here) D. (add distractor here)

Clone new item: 

Clone new item Same item, use same distractors, but insert new correct answer

Which one of the following obstetric risks is commonly associated with mullerian anomalies?: 

Which one of the following obstetric risks is commonly associated with mullerian anomalies? Pre-eclampsia (still a distractor, instead of correct answer) Malpresentation (new correct answer) (add same distractor here) D. (add same distractor here)

Clone second item: 

Clone second item Same process Keep same distractors Insert new correct answer

Which one of the following obstetric risks is commonly associated with mullerian anomalies?: 

Which one of the following obstetric risks is commonly associated with mullerian anomalies? Pre-eclampsia (still a distractor, instead of correct answer) Preterm labor(new correct answer) (add same distractor here) D. (add same distractor here)

Another example: 

Another example Change negative context Multiple choice question to positive, then clone it.

Negative context question: 

Negative context question Post-streptococcal glomerulonephritis is associated with all EXCEPT Subepithelial humps Pharyngitis Low serum complement levels Low anti-streptolysin antibody titers* Hematuria

Post-streptococcal glomerulonephritis is associated with :: 

Post-streptococcal glomerulonephritis is associated with : Low anti-streptolysin antibody titers (now a distractor) Subepithelial humps (now correct) (add a distractor) (add a distractor) (add a distractor)

Clone 3 more items: 

Clone 3 more items Keep the 4 distractors (1 old, 3 new) the same Change the correct answer each time, alternatively using: Pharyngitis Low serum complement levels Hematuria

Do students share exams?: 

Do students share exams?

Many topics have more than one compenent: 

Many topics have more than one compenent Think of the questions you used to write as “All of the above” - Those kinds of topics are great for cloning Write the question with only one of the possible correct answers Work hard to develop plausible distractors Keep the question and distractors; alternate the correct answers

Use 1 per test: 

Use 1 per test Store the others for the future

Cloning True/False: 

Cloning True/False Write the True/False question Clone it to the opposite answer

TRUE Original: 

TRUE Original The most common cause of congenital adrenal hyperplasia is 21-hydroxylase deficiency Now, clone to FALSE

Clone to FALSE: 

Clone to FALSE The most common cause of congenital adrenal hyperplasia is 3-beta-HSD deficiency Or The most common cause of congenital adrenal hyperplasia is 11-hydroxylase deficiency

Cloning Matching : 

Cloning Matching Not as easy to do as for Multiple Choice or True/False Easiest when the list of possibilities is very long Try to utilize the options that were not used in the original set, if possible.

6. The joy of feedback: 

6. The joy of feedback Writing questions is difficult Writing questions alone is even more difficult It’s unlikely that your first draft of any question should be submitted

(Clinical presentation was here): 

(Clinical presentation was here) Which one of the following is the best explanation for the serum sodium concentration and the concentrated urine?

Options in first draft : 

Options in first draft Arginine vasopressin release caused by hypotonic serum. Arginine vasopressin release caused by hypertonic urine. Arginine vasopressin release caused by low effective blood volume. Arginine vasopressin release caused by a combination of hypertonic urine and low effective blood volume

Your colleague recommends tidying the options: 

Your colleague recommends tidying the options The phrase “Arginine vasopressin release caused by” repeats in all the options Move it up to the question

The best explanation for the serum sodium concentration and the concentrated urine is that arginine vasopressin release was caused by: : 

The best explanation for the serum sodium concentration and the concentrated urine is that arginine vasopressin release was caused by: hypotonic serum hypertonic urine low effective blood volume a combination of hypertonic urine and low effective blood volume

1st drafts can include too many details: 

1st drafts can include too many details When it’s your patient, many of the details will come to mind when writing the test question It’s easier for your colleague to weed out any information that isn’t needed to answer the question

Sample first draft question: 

Sample first draft question Miss Smith, a previously healthy 22-year-old woman, is brought to the ER by her overly-protective mother. Miss Smith has a one day history of dysuria, urinary frequency, urgency, and supra-pubic tenderness. She noted cloudy urine, but denied fever. Her urinalysis showed many polymorphonuclear leukocytes and bacteria. Which of the following best characterizes her condition?

Edited question: 

Edited question Miss Smith, a previously healthy 22-year-old woman, presents to the ER with a one day history of dysuria, urinary frequency, urgency, and supra-pubic tenderness. She noted cloudy urine, but denied fever. Her urinalysis showed many polymorphonuclear leukocytes and bacteria. Which of the following best characterizes her condition?

Questions?: 

Questions?

Should we write some questions together?: 

Should we write some questions together?

What problems have you faced when working on test assignments?: 

What problems have you faced when working on test assignments?

THANK YOU: 

THANK YOU Lynn C. Webb, Ed.D. testing@lwebb.com