Free Voluntary Reading and Autonomy in Second Language Acquisition: Improving TOEFL Scores from Reading Alone

The procedure was simple. Students were given access to the library of graded readers that had been available to them during their classes. In contrast to reading done as part of the classes they took, readers were not asked to keep any records of how much or what they read, although some did so. The instrument used was the ITP (Institutional Testing Program) TOEFL (Test of English as a Foreign Language). This test was constructed from previously administered TOEFL tests, and is available from the Educational Testing Service for local use for institutions for placement, awarding credit, as a final exam, etc.. It consists of three parts, Listening Comprehension, Structure and Written Expression, and Reading. Test administration takes about two hours, and multiple forms are available. Reliability of the TOEFL is very high (for the ITP-TOEFL, total reliability = .95; listening comprehension = .90; structure = .87, reading = .88; TOEFL, 2005). (Note: A TOEFL score of 550 is thought to represent enough English competence to study in an American university.)

RESULTS

Table 1 presents gains made by each subject for each component of the TOEFL, as well as weeks spent reading and the average gain per week. The average gain per student was 3.5 points per week.

Table 1: Gains made by students

Name	Test date	Listen	Grammar	Reading	*Total	Gain	Weeks	**Pts/Wk
Noriko	1/17/01	51	44	41	453
	4/6/01	51	52	47	500	47	11 wks	4.2
Sumiyo	1/22/03	46	44	42	440
	4/5/03	50	49	44	477	37	10 wks	3.7
Yoko	7/12/03	44	45	48	457
	8/6/03	46	49	50	477	20	3.5 wks	5.7
U	7/12/03	46	51	47	480
	10/25/03	49	50	55	513	33	15 wks	2.2
Yu	7/12/03	44	42	46	440
	10/25/03	46	44	51	470	30	15 wks	2
Kenji	3/31/05	n.a.	n.a.	n.a.	467
	6/10/05	n.a.	n.a.	n.a.	500	33	10 wks	3.3

*Total TOEFL scores are arrived at by calculating the mean of the three components and multiplying by ten: eg, 51 + 44 + 41 = 136/3 = 45.3 *10 = 453.
**The calculation of points per week may over-estimate the amount of reading done per week because the time period included, in four cases of out six, the final two weeks of the semester during which final examinations were administered. n.a. = not available

It was not possible to calculate the amount gained per page for all subjects, as only three subjects provided information necessary for this calculation. Yoko reported reading 300 pages, and gained 20 points, a gain of about 1 point for each 15 pages read, and U read 1300 pages, gaining 33 points, a gain of about 1 point for each 40 pages read. Kenji reported that he focused exclusively on the work of Sidney Sheldon and read up to 200 pages per day.