Gunn, Smolkowski, Biglan, Black, and Blair, (2005) studied the effects of supplemental reading instruction on struggling Hispanic and non Hispanic readers over a period of four years. Subjects were chosen from 14 schools in four Oregon communities. From a population of 4,004 students, 359 families were recruited after an assessment of reading difficulties or social skills problems. 60 subjects dropped out of the study for various reasons. 299 participants, 159 of whom were Hispanic and 140 of whom were not Hispanic, completed the study. There were 161 boys in the study and 138 girls.

There were 51 kindergartners, 87 first graders, 90 second graders, and 71 third graders. The information about ethnicity and language use came from parent interviews. 94% of Hispanic students were from Mexico, the rest from Central America or other Latin American countries. 9% of the Hispanic students were born in the United States and 85% were born in Mexico. Eighty four percent of the Hispanic families spoke only or mostly Spanish at home. Parents were paid to participate in the study (Gunn et al. 2005), $30 to complete a parent questionnaire at the end of each year, and $15 for providing information on the social behavior of their children at the beginning of the study, the end of year two, and the end of year four.

Subtests of the Woodcock-Johnson Tests of achievement and the Walker- McConnell Test of Social Skills were given to the subjects (Gunn et al. 2005). Students in each social skills category were grouped in matched pairs by ethnicity, starting with the pair of least skilled readers, then assigned randomly to a condition: either intervention was provided or not. Supplemental instruction was provided 30 minutes daily, as well as parent training and social skill intervention for students in the experimental group.

Half the students received 6 months of supplemental reading instruction in the first year of the study, and received supplemental instruction for the entirety of the second year (Gunn et al. 2005). At the beginning of the first year, all students were assessed using the Woodcock Johnson Tests of Achievement, and then were again assessed at the end of the year for 4 years. Assessors were not aware of the children’s membership in control or intervention groups. Supplemental instruction was only provided for the first two years, but testing continued to assess if there were long term effects of the supplemental instruction.

Students were pulled for supplemental instruction during times that did not interfere with classroom instruction (Gunn et al. 2005). The instruction was conducted in their home rooms. Nine instructional assistants worked with the students in small groups of two to three. For supplemental instruction, Reading Mastery was used for first or second grade students. The program consists of phonemic awareness, phoneme and grapheme correspondence, and decoding. Corrective Reading was used when the students reached third and fourth grades. This program is for older students who do not have the primary skills. It covers the same components as Reading Mastery, with topics interesting to older students, and moves at a faster rate. No information was given as to the program used with kindergarten intervention subjects. These groups usually spent five to seven minutes on phonics, 10-15 on word reading and spelling, and the rest of the session on reading passages to build fluency and accuracy.

Contingencies for Learning Academic and Social Skills and Dina Dinosaurs Social Skills and Problem-Solving Curriculum were used to help students in the intervention group improve their social skills (Gunn et al. 2005). These programs are designed to reinforce positive behaviors and help children reduce inappropriate behavior.

Gunn et al. (2005), concluded that intervention students gained much faster (p=.0052) than controls, who started at the same place at T1. Though at T3 there was not a significant difference between groups (p=.0887), the students letter- word identification did grow faster than control group ( p=.0092). At T5 (the end of the fourth year), showed that even two years after the end of intervention, these students were still significantly ahead of the control group in word attack (p=.0346). The group was no longer improving at a greater rate than the control group (p=.5461), but they still scored above the control group on the assessment.

Word attack scores showed that students in the intervention group performed significantly better (p= .0013) at T3 than students in the control group (Gunn et al. 2005). At T5, there was not a significant difference between the groups (p= .8274). Control groups were actually increasing at a higher rate (t=8.40, p<.0001) than intervention students (t= -4.23, p<.0001). There was a significant difference in the rate of improvement between ethnic groups.  Hispanic students scores started off lower than their non Hispanic counterparts. Results showed that non-Hispanic students in the experimental group improved at a greater rate (p<.0001) than non-Hispanic control groups, and Hispanic students with the same treatment improved even more quickly (p=.0228) than their matched pair. This phenomenon faded at T3, where there was not a significant difference between races (p=.0954).

For oral fluency, students in the intervention group improved at a faster rate than the control group (p=.0129) (Gunn et al. 2005). At T3, the difference in abilities between groups was significant (p=.0356). At T5 the difference was even greater (p=.0144). “…Intervetnion students at T1 read less than 2 cwpm faster than controls, but by T5, they read almost 14 CWPM faster” (Biglan et al., 2005. 78.)

The study (Gunn et al. 2005) concludes that the results showed that intervention helped for all students on word attack and comprehension. On word attack, Hispanic and non-Hispanic students both improved more than their matched pair in the control group, even though the slopes started to come together at T5. Hispanic students improved faster than non-Hispanic students. At the final assessment (T5), at the end of four years and two years after the end of the intervention) “students differed by condition on letter-word identification (d=0.25), oral reading fluency (d=0.29), and reading comprehension (d=0.29)” (Gunn et al. 2005, p. 82)

The study (Gunn et al. 2005) shows that intervention in small groups can affect the reading achievement of both Hispanic and non-Hispanic students who struggle with reading. Though Hispanic students started at a lower level (possibly because less than 15% spoke English at home), they improved at greater rates than their non-Hispanic counterparts. This is not only due to their increased familiarity with English, as the control group also became more familiar with the language during this period. The researchers suggest that it is helpful to provide supplemental instructions in reading English even before the students are proficient speaking it.

Paying the subjects to participate may have helped the researchers with their study, but brings questions into the reliability of the information. The parents may just be doing it for the money and not because they care about accuracy. Parents who have children with poor social skills or low levels of reading achievement may feel ashamed and not want to be honest, and so fill out a survey dishonestly trying to make their child look more competent than they are. They might not want to subject their family to scrutiny, and be participating just to get the money. No information was given as to the socioeconomic status of the participants.

Intervention in comprehension was not provided. If reading is the making of meaning, then this was a major omission on the part of the researchers. The researchers (Gunn et al. 2005) tested on comprehension, but comprehension strategies were not part of the intervention treatment. Testing for something that was not made part of the control is not reliable. The reader does not know what may have caused these changes. Though intervention students improved more from T1 to T3 (p= 0357 at T3) Control and intervention students improved at the same rate from T3 to T5 (p=.3703). This could be expected for something that there was no specific intervention addressing. It seems control and intervention groups may have differed on this because of a phonic ability to read the words, and thus have a chance to comprehend them, but once the control group caught up in that sense there were no strategies taught that would have kept the intervention group ahead of controls. There was also no extra help in the intervention in vocabulary, but it was tested for as well. Intervention subjects scored only slightly better (p=.0446) at T3 and even less so at T5 (p=.0571.)