|
|
|
Research Findings‹ Successful programs | Writing Placement | References › Solid formal research studies exist, both questioning and defending writing-placement methods. Administrators and faculty will listen when research findings are presented. Here is a sampling of studies. More studies can be found annotated in Speck (1998), and the trends in the research are synopsized in Haswell (2004) http://comppile.tamucc.edu/writingplacementresearch.htm. Bers, Trudy H.; Kerry E. Smith (1990). Thirty-seven percent of the students who failed the regular writing courses had been placed into it by a local holistic rating of a placement essay. Blakesley, David (2002). Describes diirected self-placement at Southern Illinois State University with some careful validation of it. Argues that the system is no worse than previous systems that placed students by exam. Breland, Hunter M. (1977). Correlations with course performance and three kinds of tests: TSWE .51; SAT Verbal .44; essay .47. One of many studies that show direct testing with little greater predictive power than indirect testing. Bridgeman, Brent (1991). An essay correlated with an end-of-the-semester essay.at .56 and with grade at .27; same with SAT verbal, .44 and .22; same with TSWE, .51 and .25. Ericsson, Patty; Richard H. Haswell (Eds.) (forthcoming). Contains a number of studies and discussions of research findings showing that automated essay scoring systems sold by ETS, ACT, and the College Board are no better at predicting success in writing courses than are the standardized tests. Galbato, Linda; Mimi Markus. (1995). On evidence of writing during the first week of class. a third of students were recommended to change the course that a standardized test (ASSET, SAT, TSWE, or ACT) had placed them in, Hughes, Ronald Elliott; Carlene H. Nelson (1991). The correlation of ASSET with passing the freshman course was .23. In head count, 60 of 162 students passed but would have been predicted as not passing. This is an important study because it tests predictability in a way few other studies dare to do: get predictive scores, then put all students regardless of score into the mainstream course and see what happens, then compare what happened to had the pre-course scores been used. Huot, Brian (2002). Describes the portfolio placement system at the University of Louisville, now dismantled (pp. 156–163). During the first five years of the program students who chose the portfolio placement route also took the ACT. If the two placement decisions differed, the student was allowed to pick, and almost always picked the higher placement: “for the most part they were successful in the courses they chose” (p. 161). Larose, Simon; Donald U. Robertson; Roland Roy; Fredric Legault (1998). SAT plus high-school rank accounted for only 4.2% of GPA with high-risk students. One of a number of studies that show that the predictive validity of scores produced by minority students is lower than scores produced by majority students. Leonhardy, Galen; William Condon (2001). Describes difficult decisions as raters are deciding placement from a two-sample direct placement testing at Washington State University. McKendy (1992). Reviews twelve recent studies of the predictability of holistic scores of placement essays as well as another review of indirect tests published in 1954.. He shows that a single-sample writing essay correlation with course grade typically runs from .2 to .4, and best efforts to create multiple-regression combination other predictors such as high-school GPA can raise this only to around .6. All told, a very strong argument against using a single test or single essay to place a student. Meyer, Russell J. (1982). Compared take-home essays, written and revised at the student’s leisure and then read by faculty, with objective tests: 3% were placed in a higher course than the objective test would have done, 44% were placed in a lower course. A number of studies concur that faculty seem to be tougher with placement decisions than off-campus commerical scores. Olson and Martin (1980). Compared three different methods of placement testing and found that 1,002 (61%) of their entering students would be placed differently by indirect proficiency testing than by teacher rating of an essay, and 1,051 (64%) would be placed differently by that teacher assessment than by the student’s self-assessment. Shell, Duane F.; Carolyn Colvin Murphy; Roger Bruning (1986). Three types of self-efficiency (student’s estimate of their own writing abilities) correlated at .32, .17, .13 with a 20-minute essay, scored holistically and analyticallly. Shermis, Mark D.; Jill Burstein (Eds.) (2003). Defends computer scoring programs largely by comparing the machine’s score with a human holistic score. In the entire book there is not one report of a completed study of instructional validity. Smith, William (1993). The best and most thorough study of predictive validity and faculty rating of student essays, at the University of Pittsburgh from 1985–1993. Stock, William P.; Juan M. Flores; Linnea M. Aycock (1986). At California State University, Fresno, correlation of final course grade in freshman composition was .24 with the CSU English Placement Test; .29 with the SAT Verbal score; and .35 with the Test of Standard Written English. ‹ Successful programs | Writing Placement | References › |