Studying Emergent Writing: An Expedition Together with Iris Levin

Adriana G. Bus, Leiden University

When Shaul returned home after a day of work in the clinic, Iris used to say, "He worked; we were just playing." This statement always made me a bit angry with Iris, but it was undeserved. Iris was merely trying to protect herself and us against too much involvement and commitment. Her persistence in continuing the research after she became ill is another illustration of her high commitment.

Since the eighties of the last century, Iris studied how children acquire alphabetic knowledge, and I was one of the lucky people who joined her at times on this long-lasting expedition. While preparing for this day in memory of Iris, I tried to reconstruct how our collaborative projects developed and how this collaboration influences my research to this day. Iris' research has influenced the ways of teaching alphabetic knowledge to young children, and I will try to highlight how. I make no pretention to present an overview of all the work that Iris has done during her long career. My comments are a personal view of our common enterprise.

II

Iris and I first met at the ISSBD in Tours in 1985, where we both presented posters. At that time, I had a burgeoning interest in emergent reading and writing. Iris had a poster, along with Lilianna Tolchinsky-Landsmann, focusing on figurative elements in preschoolers' writing. I presume that Lilianna, Iris' first doctoral student, sparked Iris' interest in emergent writing. Lilianna, as a student, had collaborated in Emilia Ferreiro and Anna Teberosky’s Argentinian research group. Guided by Jean Piaget’s approach, their work was rooted in the belief that all knowledge, including the acquisition of writing, follows a developmental trajectory. In the process of internalizing knowledge, children reconstruct it in their own terms.

Iris and Lilianna observed a combination of features of writing and drawing among their participants. The children used letters to write dictated words, but there were also figurative elements present. For example, they might use more letters for larger objects, write the word "tomato" in red, or select forms typical for an object. An illustrative case is a girl who wrote the name of her baby brother as a series of zeros "because he is zero."

In the years following this initial exploration, Iris continued this line of research. Collaborating with Ofra Korat, she investigated, for different ages, which cues most influenced the number of letters children used when writing words. The effects of phonology increased, while the effects of semantics decreased, indicating a growth in understanding the alphabetic principle.

III

In the following years, as we met in Israel, the Netherlands, and at conferences worldwide, we began to formulate plans for a joint cross-cultural study. Most emergent writing research, including our own, had previously concentrated on four- and five-year-olds. Given the prevailing ideas at that time about the origins of writing, there was a perceived importance in extending the research to much younger children. The prevailing theory asserted that there might be some overlap between drawing and writing, but writing features are present from the beginning. Even the youngest children were assumed to exhibit writing features. However, other studies suggested that writing and drawing could not be reliably distinguished at an early age.

IV

It was Iris who took the lead in designing our joint study. She was much more aware than I was (at that time) of all the traps and pitfalls in a balanced design and what could go wrong when eliciting writing and drawing from children this young. As an undergraduate student, she received excellent training in designing experiments. Her teacher of statistics and research methods was Daniel Kahneman, the later Nobel Prize winner.

When she began her psychology studies at the Hebrew University in Jerusalem, her initial goal was to become a clinical psychologist. Her literary adventures as a teenager, such as translating Guy de Maupassant's story "The Necklace" into Hebrew at the age of sixteen or seventeen, illustrate her deep interest in the human factor. This tragic story about Madame Mathilde Loisel and her husband fascinated her and may have sparked her interest in clinical work. However, she reconsidered her plans after attending Daniel Kahneman’s courses in statistics and research methods.

Kahneman describes the group of undergraduates as very bright in his account of his early experiences teaching in the psychology department at the Hebrew University in Jerusalem in 1961. In his autobiography, he describes these experiences as consistently gratifying "because the students were so good: they were selected based on a highly competitive entrance exam, and most were easily PhD material."

From reading Kahneman’s (autobiographical) notes, I can imagine why his course in statistics and research methods was so inspiring to Iris. Kahneman states, "At that time I did a lot of serious thinking about valid intuitions on which I could draw and erroneous intuitions that I should teach students to overcome. I had no idea, of course, but I was laying the foundation for a program of research on judgment under uncertainty."

Iris adopted his strict standards for good experimental designs, as is evident in all her work, but fortunately not his policy for writing up research. Kahneman mentions about his publications at that time: "I got quite nice results in my one-question studies, but never wrote up any of the work because I had set myself impossible standards: in order not to pollute the literature, I wanted to report only findings that I had replicated in detail at least once, and the replications were never quite perfect."

V

We dictated eight words to our 2- and 3-year-old subjects, which differed in color (e.g., grass and sun), size (e.g., mother and baby, father versus bird), and number (e.g., flower versus three flowers). We asked the children to both write and draw the words. Careful consideration was given to selecting stimuli that were equally meaningful for Israeli and Dutch children.

We faced challenges when initially coding the completeness of written and drawn representations of children this young. I travelled multiple times to Israel to discuss coding issues face to face. Iris and Shaul always welcomed me into their apartment. We worked from early morning to late evening at Iris' dinner table, taking breaks to walk the two dogs, Jazz and Blues.

VI

Our findings deviated from the dominating idea at that time that writing and drawing differ from early on. Up to age 3, almost all children produced scribbles or what we termed a "good form," making it impossible to distinguish between products intended as writing or drawing for us, the mothers of the children, and the children themselves. The data did not support our expectation that the origin of drawing and writing was different.

However, we did observe an improvement in the skill to draw print in three-year-olds. Features typical for print were present at this early stage. Children segmented writing, used increasingly complex forms, and many other typical characteristics of writing were increasingly evident in their products. Similar to how they drew a flower by representing features like the stem and leaves, they drew the two-dimensional object of print by creating something linear composed of different units that, with age, varied more and became more letter-like. Our main conclusion was that children as young as 2 to 3 years are familiarizing themselves with the visual elements of print. In 2003, we finally published our paper in Developmental Psychology.

VII

A logical follow-up to the finding that visual processing is changing in the early years was: Does visual processing of print further refine when children are four and five, and does the way in which children visually process print predict letter knowledge and learning letters? I will quickly present a few results from a study using eye-tracking that we recently conducted in Leiden to demonstrate that this early visual development is important for developing alphabetic knowledge.

From research with adults comes evidence that expert readers recognize letters by fixating only on small areas of the letters (referred to as distinctive areas). Similarly, we used eye-tracking to determine whether four- and five-year-olds exhibit a similar pattern. We created "heat maps" to visualize which parts of the letter attracted the most attention from our subjects and could thus be seen as the most distinctive areas in letters.

Using heat maps, we concluded that, just as found in adults, small parts are distinctive features for letter recognition (cf. Dunn-Rankin’s findings in adult groups). In letters such as "T," especially angular intersections are fixated. Almost all participants fixated on the same small part of letters, but the duration of fixation varied. We found that this variation in how long children look at distinctive features predicts letter learning. The briefer the children fixated on the distinctive areas, the more letters they knew, and the more they benefited from repeated readings of an alphabet book.

Initially, it surprised us that briefer, not longer, fixations on distinctive areas predicted letter knowledge and letter learning. However, this makes sense if we assume that critical dimensions of letter forms are engraved in children’s memory as a result of previous encounters with letters, and consequently, a glance is sufficient to recognize the letter.

VIII

Iris was fascinated by the special position of the proper name in young children’s development of writing skills. Her early writing research with Lilianna had shown that any 4- to 5-year-old can print their name in conventional spelling (Tolchinsky–Landsmann & Levin, 1985). She also found that by the age of 4–6, children are more proficient at naming letters of the proper name than other letters (Levin & Aram, 2004). Similar observations were reported by Becky Treiman for American children (Treiman & Broderick, 1998; Treiman & Kessler, 2003).

The gap between the levels of name writing and word writing was, however, never tested. Iris initiated, therefore, in 2003 a secondary analysis in which data in various age groups, collected in Israel and the Netherlands, and coded in a similar way, were synthesized.

In this study, four data sets were included, comprising about 450 children in total. The conclusions supported the gap between the name and other words: Whatever the age or culture, the names had more characteristics of writing compared to other words. When children wrote random words, children showed very similar behaviour, as indicated by the high correlations between each dictated word and all other words – on average, .71. The correlations between the name and random words were either low or insignificant – on average, .30.

IX

Originally, the focus in the writing research had been on children’s endogenous motivation to experiment with words. Learning was not seen as the result of caregivers’ input, feedback, or guidance. Young children may know more about name writing than writing other words because they often come across their name on personal possessions—like their bedroom door, drinking glass, and cubby hole at school—and this may explain the advantage found for the name over other words.

Iris argued that a preference for the name stems primarily from differential mediation. Caregivers may typically select the name as a starting point to direct children’s attention to print, to encourage them to mimic it by scribbling or producing marks, to practice writing a letter string, and to name letters and become acquainted with their sounds.

At the end of the eighties, Iris began pioneering research with Dorit Aram on how parents support young children’s writing. Maybe my work contributed to her interest in this aspect of early writing. In 1988, she asked me to give a talk about parental mediation in her department at Tel Aviv University. I had just published a paper about mother-child interaction, attachment, and emergent literacy in Child Development.

The study with Dorit highlighted the important role of mothers in the development of emergent writing skills. Mothers not only stimulate children to write, but they also elevate their child’s writing level at the same time. Dorit and Iris’ findings evidence that maternal support makes a huge difference. When mothers provide help in shaping the letters or suggest phonemic analysis—such as asking, "Which letter do you hear?"—children outperform their less fortunate peers who receive less (adaptive) support from their mothers.

X

In Leiden, we delved into the special position of the name in writing development. We tested the hypothesis that name writing might help develop phonetic writing. My PhD student Anna Both-de Vries argued that adults may provide children with fairly substantial amounts of direct instruction about letters as symbols for sounds when talking about children’s own or other people’s names. They say things like: “Look. That’s your letter” or “That’s /pi/ of Peter” or “M of mama.” Adults may thus unintentionally instruct children on how letters of their own name and other people’s names sound in words.

There was prior research showing that children use letters from their own name more often than other letters when they produce strings of letters to write words. Iris and Dorit (Levin & Aram, 2004) had found that young speakers of Hebrew show elevated letter-sound knowledge for the first letter of their own name. We went one step beyond these results and tested that the first letter of the name may be the first one to be used phonetically in children’s writings, whatever the letter is. More than other words, the name may thus be the start of symbolic writing.

Treiman (2001) had not found a special effect of letters from the child’s name in kindergarten, first grade, and second grade. Her observations did not confirm that early phonetic writing was limited to letters from the name. We argued that in Treiman’s 2004 sample, the effect of name letters might not have been present because her group included many quite advanced children. It seemed plausible that only in the very beginning stage of phonetic writing would phonetic writing be limited to the first letter of the name.

XI

So we focused on a group in the very beginning stage of phonetic writing. In their writings, letters from the proper name often occurred, and, in particular, the first letter. This letter appeared more often than any other letter of the name or non-name letters. So, Tom may often use the letter T and rarely one of the other name letters.

The most interesting result was that they used the first letter of the name not randomly but in most cases correctly, while all other letters were used incorrectly. Linea Ehri described emerging spelling skills in terms of an increasing ability to parse spoken text and decide what units align with the written form (Ehri & Wilce, 1985). We refined this theory by showing that identifying phonemes starts with the first letter of the name.

Recently, Chris Lonigan (2012) extended the finding to a group of English-speaking preschoolers. A substantial proportion of full-name writers were able to use their advanced knowledge about the first letter in the name to represent phonemes when spelling words, consistent with our findings with Dutch preschoolers.

XII

Iris felt compassion for disadvantaged children who often fail to learn to read and, therefore, struggle at school and later in society due to low academic performance and clumsiness in situations such as job interviews. She aimed to advance reading instruction, especially for low-SES children. In her valedictory lecture in 2006, she talked about her attempts to improve the teaching of reading and writing by designing interventions in schools instead of just conducting more controlled research on basic processes. She went to schools in Jaffa to help the teachers improve their teaching.

Rereading the sheets, it struck me how uncertain Iris was about the value of her hard work at the schools in Jaffa. In her lecture, she concluded that it is worthwhile to initiate interventions that promote literacy among preschoolers in a low SES population via the teacher, as she did because “research with teachers may form a bridge between basic research and introducing reforms into the educational system.” The studies thus directly contribute to the participating children and teachers. And more importantly, the conclusions are accessible to policymakers and practitioners. This way, you may affect the educational system more than with basic research alone.

XIII

Iris did not believe that these interventions would add in important ways to the improvement of teaching. She found more improvement in reading and writing in the experimental than control classrooms, but in her valedictory address, she was rather desperate about the interpretation of her findings, and she gave several good reasons to be desperate. There was, for instance, no easy way to know what the intervention had been. Looking back, Iris concluded that it had been the result of many decisions, some evidence-supported, but more often based on common sense.

She also mentioned: Interventions in classrooms carried out by teachers need strong collaboration with teachers if the teacher is implementing the intervention, but they have their own (conflicting?) agenda and simultaneously use other programs that often compete with the intervention program. So you don’t know what exactly happened neither in the experimental nor in the control group, was another devastating observation.

XIV

It is a pity that we cannot know where Iris’ thinking would have taken her to enable a more experimental approach in schools. Maybe she would have joined me in attempts to use computer interventions. We talked about the possibility of thus creating a real match at school for the long stretch of early literacy experiences across the preschool years at home. Unfortunately, this research and the investment in computer programs for young children cannot be done with our low budgets. Compared to society’s funding for finding the Higgs particle or sending rockets to the moon, investment in studying teaching is peanuts.

In Leiden, we had a chance to bring the results of the basic research of Iris and other people in a more controlled way into classrooms. I came in touch with a private company that aimed at developing educational software for very young children. They responded enthusiastically to my proposal to build a computer program for young children aiming at stimulating similar experiences with the proper name as a print-rich environment at home may offer. The program is web-based and automatically adapts to the child’s proper name, providing the child with targeted instruction using the first letter of the name. The program affords far more exposure than from the teacher alone in identifying the name, identifying the first letter in the proper name, and identifying this letter in other words. The program includes a tutor who adapts to the child’s responses. We thus tried to model all games in the program after Dorit and Iris’ findings when observing mothers.

XV

The program enabled us to design randomized control trials in a large number of schools and study the effects of the computer intervention in various age groups and samples differing in other moderators. We have by now done three large experiments with very promising results. One study, for instance, focused on five-year-old Dutch children who had no notion that letters relate to sounds. Directly after the intervention, we found gains in children’s alphabetic skills – on the order of half a standard deviation on average.

Even more interesting was that the treated group outperformed the non-treated group after two years of beginning reading instruction. The group at risk who received the remedial computer program at kindergarten age scored half a standard deviation higher on standardized tests than the non-treated group. The program narrowed the skills gap by about 8% in the first two years of reading instruction.

XVI

The study of teaching reading is rocket science, and researchers like Iris cannot be missed. It is a pity that we cannot know where her thinking would have taken her after her long career path across the arenas of early literacy research. We live in a changing society in which technology takes an increasingly important position. So far, we do not notice much of technology in preschool and kindergarten, probably due to the low investment in software for young children, but children do spend a lot of time with commercial software at home at the expense of reading and writing with adults.

Iris could only make a start with looking at the effect of technology on young children’s literacy. Familiarizing with letters happens nowadays less by sharing writing with adults and more and more through computer programs. Young children make use of typing or following the letter’s shape on a computer screen with a finger instead of handwriting. Brain researchers suspect that this may affect letter learning negatively. When you type the letter, you do not create motor activity unique for each letter because the action of typing has no intrinsic relationship with the shape of the letters – you make the same movement regardless of which key you press.

Iris made a start with exploring the promise of eBooks in collaboration with Ofra Korat. She liked computer interventions, probably in particular because she was in control of what children experienced.

XVII

After my last meeting with Iris at the end of March/beginning of April this year, she wrote on April 6: “I am sorry we could not meet again. I felt too weak. But it was really a treat to see you and talk to you. I had so many more questions to ask you ...” And I had so many more questions for Iris. I am so sorry that our discussion, ongoing for almost thirty years, has stopped, at least face-to-face. But her writings will remain a source of inspiration.

December 1, 2013