Does Pre-Assessment Work?

The February edition of Educational Leadership includes an article by Thomas R. Guskey entitled "Does Pre-Assessment Work?" which made me think about the practices I have seen and used in the classroom. Guskey first admits to the lack of research on the value of pre-assessment. This makes it a hard sell in many cases.

Next he discusses the purposes of pre-assessment.

1. Identify exceptional learners: the gifted community embraces this purpose to enable clustering for differentiated instruction and curricular compacting toward acceleration. For example,
   Brulles, D., Cohn, S. & Saunders, R. document research on the effectiveness of clustering and differentiating instruction in their article here. While pre-assessment might be used for placement decisions, especially in math, there is no universal acceptance or practice in this area.
2. establish a baseline for performance to measure growth: in sports programs we often see coaches record times, number of baskets made, distances thrown or jumped, number of balls hit, etc and then use that to measure progress. We see teachers measure score on a spelling list or math facts and then ask students to complete instruction and various assignments that will, hopefully increase the score.

The second type of pre-assessment lends itself to misuse when students are graded on improvement (I can score a 0 on the pre-assessment and then not do well on the post-assessment and still ace the class) or teachers are evaluated on student gains (the pre-assessment doesn't matter kids, don't worry about it verses this is very important and will be graded). Other confounding factors include:

• natural growth and maturity- a six year old will hold a pencil better and have more writing stamina than a 5 year old, a student who grows five inches in a year should be able to jump farther and sprint faster, a student in the concrete stage of development will see the world differently than one at an abstract stage.
• life circumstances- a trauma like a family divorce, accident or death, or exposure to violence will likely influence student performance. Getting adequate sleep and a good diet will help student perform better. Students on medication for focus, anxiety, depression or other mental health concerns will do better if they do not forget to take their meds.
• instructional practices around the testing- "This one does not matter," versus "This one is a major grade in the book." My son was exposed to many of these pre-assessments. He made it one of his missions to come up with the craziest responses imaginable (too much Calvin and Hobbes) on pre-tests. 

Then Guskey talks about measuring different types of learning goals. Cognitive tests assess academic goals (what causes the seasons, how many regular figures can you name, What is a simile), affective ones assess attitudes, beliefs and interests (do a project on snakes, sharks or flowers or how and what do they read); behavioral ones assess skills (athletic or musical performance, ability to cut on a line or use a protractor). Often these goals might appear on the same pretest- for example I want to know if you can identify various angles, use a protactor and how you feel about geometry.

Then he gets to the nitty gritty of the forms of pretests. Again there are overlaps.

• prerequisite- What do you need to be able to do before you start this unit? Dribble before you play a game of basketball. Have good reading decoding before you read a chapter book. Use a ruler before we measure volume of prisms. Have you ever cared for a pet? before learning about personal responsibility.
• present- measure current knowledge. What do you know about Shakespeare? Can you multiply by fives? What do you know about adding fractions? Tell me what you know about glaciers. Tell me about a book you read that you enjoyed.
• Preview- comparisons of student growth come from this category. If you know how to add fractions, I do not need to cover that in my instruction. If you have no idea about the i before e spelling rule, I should probably teach it before I assess your spelling of words like conceive, weigh and thief. If most kids think that the seasons are based on closeness to the sun rather than tilt of the planet, I do not need to waste time to pre-assess it.

Guskey then presents one high quality study of pretests from 1983. Leyton-Soto randomly divided students into four groups: pretest with traditional instruction, pretest with instruction in prerequisite skills, pretest with mastery learning instruction, pretest with both prerequisite skill instruction and mastery learning. He worked in Algebra 2 and second year foreign language classes with clear prerequisites, assisted in the development of pretests and provide support in the instructional phase to ensure design integrity. Students in the prerequisite skill instruction groups received two weeks per semester of mastery learning on missing skills. Results indicated that all three interventions significantly outperformed the traditional approach. The combination of mastery learning and prerequisite instruction resulted in students performing at mastery level on a comprehensive final exam with almost eight times the frequency of students in the traditional group. Yes, this was two subject areas with clearly defined prerequisites with students who probably did not have lots of learning problems or issues with success- using upper level classes tends to weed out poor performers. That being said, a two week intervention in teaching prerequisites seems really doable if it results in a three x increase in success. Using mastery learning seems like a good choice of instructional techniques if it results in a 5 x increase in results.

Providing effective pre-assessments, intervening with instruction in prerequisites and using mastery learning techniques appears to be a highly successful plan to increase performance of our students. Since that is the name of the game it makes sense to use them for instructional purposes. If we grade based on growth or hold teachers accountable based on student gains, however, we pollute the data and invite misuse. We might as well just ignore the pretest all together. If you will not use the pretest information to inform instruction skip it and use the time a different way.

When my daughter scored an 83 on the Earth Science pretest in September and then had to sit through ten months of instruction with the whole class in order to achieve a 98 on the final I was disappointed. Her teacher did not offer her much in that time. Although she made some growth, for most kids a 13 point increase would be a devastating failure. (The class average on the pretest was under 50.) The entire course was really a waste of her time. We suggested letting her use a college level text and assignments rather than the ninth grade one- it did not happen. The pretest merely made me angry because there was no benefit to her- the teacher was required to administer it for his professional rating and her score as a significant outlier was eliminated from the data set of even that activity. It is time to get serious about using our data in a smart way rather than a poor formulaic manner.