Teachers, principals, students and parents know what happens when you don't do well on tests.
For students, poor exam scores can lead to bad grades and perhaps a grounding at home.
Low scores also have consequences for teachers and administrators.
State and federal statutes detail what can happen if schools, specifically those with students from low-income families, fail to meet state benchmarks in reading and math.
Consequences range from allowing students to transfer to other schools to replacing a school's staff for chronically failing to meet goals.
Government officials, educators and parents rely on standardized tests not only as a measure of individual student achievement, but also to gauge the success of a school, a school district and individual states.
The Atlanta Journal-Constitution newspaper recently reported that test scores from hundreds of districts across the country showed the same wide discrepancies found in Atlanta public schools in 2009, where claims of cheating were investigated and confirmed by Georgia justice officials last summer.
The Journal-Constitution has said its findings don't prove other districts are cheating but that similar score discrepancies have been found in other districts -- including Kennewick, Sunnyside, Pasco and Richland -- and should be reviewed.
Mid-Columbia administrators and educators rejected the suggestion that cheating is to blame for fluctuations in test scores.
However, they said shifting expectations, a mobile student population and changing testing methods can make it difficult for a standardized test to reflect a child's knowledge.
"Why have we gotten to the point where we live or die regarding a two- to three-day event that measures, but isn't the only measure of success?" said Bruce Hawkins, superintendent of Educational Service District 123, which supports schools in Benton and Franklin counties and provides access to state and national educational programs.
Problems in Georgia
Journal-Constitution reporters used a mathematical model to compare the math and reading test scores of single groups of students, or classes, as they progressed through their education. Only test scores from 2008-11 were examined.
According to the analysis, a school district where there was no cheating would have no more than 5 percent of its classes flagged for unusually low or high performance on standardized tests.
While not proof of cheating, districts with 10 percent or more of its classes flagged consistently or with a particularly high flag rate in one year could indicate testing irregularities, the paper reported.
Based on the paper's analysis:
-- Kennewick had more than 11 percent of its classes' test scores flagged from 2008-10. Almost 8 percent were flagged in 2011.
-- Sunnyside had more than 16 percent of its classes' scores flagged in 2008 and 2009. That dropped to more than 8 percent in 2010 and about 4.5 percent in 2011.
-- Pasco had 11 percent of its classes' test scores flagged in 2010, while the other three years the number of flagged classes ranged between 3.5 percent and just under 6 percent.
-- Richland had from 6 percent to more than 8 percent of its classes test scores flagged during the four-year period.
Nathan Olson, spokesman for the Washington Office of the Superintendent of Public Instruction, told the Herald that state education officials are aware of the Journal-Constitution report. However, there are no plans for the state to review test scores, Olson said.
Fluctuating test scores
The Herald looked at test scores from two elementary schools and two middle schools in each of the districts during the four-year period. That analysis included only the math and reading scores of third- through eighth-graders and only scores from tests sanctioned by the state -- the Washington Assessment of Student Learning, or WASL, and its eventual replacements, the Measurements of Student Progress, or MSP, and High School Proficiency Exam, or HSPE.
In the analyzed Kennewick schools, it was common for one class' test performance to fluctuate as much as 20 percentage points compared to a prior year. One group of students at Park Middle School saw its reading scores drop more than 30 percentage points before climbing more than 16 points the next year.
Sunnyside test scores showed similar performance swings.
In Pasco, one group of students at Longfellow saw math scores drop 24 percentage points between the third- and fourth-grade years before rocketing up almost 36 percentage points in fifth grade.
Richland's score fluctuations went up or down from 1 to 10 percentage points between years. There were some instances when scores went up or down more than 20 percentage points, though.
Tests of integrity
District administrators said testing is an important aspect of measuring student performance, and steps are taken to ensure the integrity of tests.
Paper exams are regulated by strict protocols. Teachers undergo special training to administer tests. Materials are locked in containers inside a storage area with limited access between testing periods.
More students are taking standardized tests online, which makes them only available to the student being tested.
"It's a really secure process," said Bev Henderson, Kennewick's assessment coordinator.
But how tests are given can affect performance. Curtis Campbell, Sunnyside's executive services director, said in an email to the Herald that reading and writing scores at one of his district's middle schools went down significantly when the MSP was offered in a computer format.
"The middle school has done the MSP by paper and pencil the past two years and has seen much better results on the reading and writing portion," he said.
How schools and districts address education and the factors affecting student and school performance also affect test scores, administrators said.
Administrators in Pasco, Kennewick and Sunnyside said the emphasis districts place on various areas of education can influence student test performance.
In Sunnyside, the use of curriculum aligned to state standards was incorporated in the past few years. Those new guidelines improved learning, but Campbell said many teachers were involved in their development. That took the teachers out of the classroom and left instruction to substitutes, which could have dragged down scores.
In Kennewick, teachers focus on getting students to read at their grade level starting with the third grade, but educators don't have the resources to do the same for fourth- and fifth-graders, said Superintendent Dave Bond. That means those students could see their reading scores drop after the third grade. The district also encourages teachers to try new teaching methods, which sometimes don't work.
"We push for innovation, and that leads to drastic changes," said Greg Fancher, Kennewick's assistant superintendent for elementary education.
If administrators notice a group of kids isn't doing well in a certain subject, more emphasis is placed on the struggling area, leading to boosted scores in that area during the next school year, Campbell said.
Student mobility can play a role in a school's test scores, as can the departure of long-term staff.
In Pasco, as many as one in three third-graders at Longfellow Elementary School will move before the fifth grade, said Mark Garrett, who oversees testing in Pasco schools. Incoming students may score lower than expected on standardized tests until educators learn of deficiencies in their prior education.
"An apartment complex can spring up over the summer and change the whole profile of a school," said Pasco spokeswoman Leslee Caul.
In Kennewick, two experienced fifth-grade teachers moved to a newly opened school in the district in recent years. They were replaced by two teachers new to the fifth-grade curriculum and test scores dropped the following year, but they eventually recovered, Fancher said.
Out with the WASL
Then there are the tests themselves.
Washington students were tested using the controversial WASL from 1997 to 2009. Since then, students have been tested with the MSP and HSPE, with high school students also taking End Of Course, or EOC, assessments.
The change in tests lead to some score shifts, administrators said. They also said that while the tests are a tool for determining student success, the results have a wide margin, making it more difficult to help kids on the fringe.
Educators in Richland, Pasco and Kennewick said they rely more on the Measures of Academic Progress, or MAP, tests, administered by the Northwest Evaluation Association.
It is easy to see MAP is being embraced. While it takes months for districts to receive MSP results, MAP test scores are available within a day of turning in the exams.
Kennewick's MAP scores, while not immune to big changes in student performance year-to-year, do show more consistency, particularly among middle school students. Data provided by the district indicate no middle school class saw MAP scores increase or decrease by more than 10 percentage points between years.
But the state doesn't recognize the MAP test, and it is not included in test score data available through the state superintendent's website.
Hawkins, the ESD 123 superintendent, said that is because MAP scores are better at showing student growth rather than student performance comparisons between schools and districts.
District administrators said the state also has been dedicated to having its own tests and not one administered by an outside organization.
More tumultuous score data may be looming. Mid-Columbia districts, and others across the country, are preparing to switch in 2015 to the Common Core Standards, a testing format that seeks to align the varying curricula state-to-state.
'An unfortunate situation'
While administrators deny there are any problems with their student test scores, they said poor test scores prompt action. Low performance leads to conferences with parents and students and prompts scrutiny for teachers and schools.
The important thing is to make decisions based on multiple sources of data, administrators said. Fancher said the Atlanta Journal-Constitution's analysis was based on a superficial level of information, and it presented educators with a Catch-22: If scores go up too quickly, you must be helping them cheat, but if they fall too quickly, you must be covering something up.
"As an educator, it's just a slap in the face to see that kind of stuff," Fancher said, referring to the accusation.
Bond said factors that contributed to the cheating scandal in Atlanta -- financial and employment incentives for teachers to make sure their students test well -- don't exist in his district or others in the state. However, beginning in 2015, student performance data is expected to be a "substantial factor" in how teachers are evaluated throughout the state.
Hawkins said factors outside of education have led some teachers to see no choice but to inappropriately assist their students.
He cited one example of a superintendent learning that an instructor allowed students to change their responses to questions after the teacher reviewed them. Upon investigation, it was found that the teacher didn't allow the cheating to make himself look like a better teacher or for any personal benefit, but because he felt the students didn't understand the questions thoroughly and couldn't answer honestly, he said.
Hawkins agreed that while student performance should be assessed, it should not be the be-all, end-all goal of education. But not all students who test well succeed in life, nor do all students who earn low test scores go on to low-paying jobs, Hawkins pointed out.
"We have created an unfortunate situation," he said.