26.4 Pictures And Knowledge Acquisition

26.4.1 Literature Search and Reviews

Through various on-line and manual literature searches, 2,196 primary research studies, reviews, books, conceptual papers, and magazine articles were identified, collected, and catalogued. The literature search was limited to the categories of static and dynamic pictures and knowledge acquisition. Many of the documents collected were not appropriate for the current review. For example, numerous papers reported the results of memory recognition studies including pictures. In addition, several studies were not included because of methodological flaws such as failing to include a control group or appropriate statistics. Many of the papers identified were not primary research studies or theoretical in nature. A total of 132 primary research studies were included across the two categories (static illustrations, dynamic graphics) used for the review. We first report the results of earlier literature reviews. Then an abridged guide to the literature will be presented.

26.4.2 Static Pictures and Knowledge Acquisition

In this section, we first present a summary of earlier reviews of the literature concerning the role of static pictures in the acquisition of knowledge. Second, we discuss the results of our literature search and sununary. A similar approach will be used for dynamic pictures and knowledge acquisition.

26.4.2.1. Static Pictures and Knowledge Acquisition: Literature Reviews. Spaulding (1955) reviewed 16 research studies using pictorial illustrations conducted between 1930 and 1953. Based on the findings of the 16 studies, Spaulding concluded that illustrations: (a) are effective interest-getting devices; (b) help the learner interpret and remember the content of the illustrated text; (c) are more effective in realistic color than black-and-white, but the amount of effectiveness might not always be significant; (d) will draw more attention if they are large; and (e) should conform to eye movement tendencies.

Samuels (1970) reviewed a series of 23 studies that investigated the effects of pictures on learning to read words, on reading comprehension, and on reader attitudes. Samuels's review covered a time span from 1938 to 1969. The studies reviewed included such treatments as: (a) learning to read words in isolation with and without pictures, (b) acquiring a sight-vocabulary with and without pictures, (c) using pictures as a response alternative in a reading program, and (d) using pictures as prompts. Samuels concluded that: (a) Most studies show that, for acquisition of a sight-vocabulary, pictures interfere with learning to read; (b) the majority of studies indicate that pictures used as adjuncts to printed text do not facilitate comprehension; and (c) pictures can influence attitudes. Many of the studies reviewed by Samuels were narrowly focused on the use of illustrations to learn to decode words in isolation. Illustrations used in the context of learning to read have generally not proved to facilitate learning.

An analysis of the pictorial research in science instruction has also been conducted (Holliday, 1973). The general conclusions reached by Holliday concerning the effect of pictures on science education were that: (a) Pictures used in conjunction with related verbal material can aid recall of a combination of verbal and pictorial information; (b) pictures will facilitate learning if they relate to relevant criterion test items; (c) pictorial variables such as embellishment, size, and preference are complex issues; and there are almost infinite interrelationships between picture types, presentation formats, subject content, and individual learner characteristics.

Concannon (1975) reviewed a number of studies on the effects of illustrations in children's texts (mainly basal readers). Concannon summarized the results of her review with the single conclusion that when pictures are used as motivating factors, they do not contribute significantly to helping a young reader decode the textual information.

Levin and Lesgold (1978, pp. 234-235) reviewed studies of prose learning with pictures and concluded that pictures do facilitate prose learning when five ground rules are adhered to, including:

1. Prose passages are presented orally.
2. The subjects are children.
3. The passages are fictional narratives.
4. The pictures overlap the story content.
5. Learning is demonstrated by factual recall.

While Levin and Lesgold (1978) focused on oral prose, they also suggest that pictures may benefit individuals reading for comprehension.

Schallert (1980) reviewed a number of research studies and presented the case for and against pictures in instructional materials. In the case against pictures, Shallert reviewed the work of Samuels (1967, 1970) and others. Shallert states that "the most convincing evidence against the use of illustrations in children's text has been marshaled by Samuels" (p. 505). Shallert noted that many of the early reviews completed by Samuels, Concannon, and others reported that the use of pictures serving as motivating factors do not facilitate a child's ability to decode text information. Shallert indicated that some of the reasons the pre-1970 studies did not identify picture effects were: (a) The primary emphasis in the word acquisition treatments were speed and efficiency; with the words being spoken aloud, pictures used in that context are of little value; (b) the illustrations used in many studies were not meant to convey new information and were only used as adjuncts to the text; (c) many illustrations used in basal readers vaguely relate to the contextual information in the text; and (d) the effects of illustrations on long-term memory were not measured in these earlier studies.

In the case supporting positive picture effects, Shallert (1980) reviewed a series of studies that covered the time period from 1972 to 1977. The general conclusions reached by Schallert were that pictures can help subjects: (a) learn and comprehend text when the pictures illustrate information central to the text, (b) when they represent new content important to the overall message being presented, (c) when they help depict the structural relationships covered by the text, and (d) if the illustrated information contributes more than a simple second rehearsal of the text.

Readence and Moore (1981) conducted a meta-analytic review of the literature on the effect of experimenter-provided adjunct pictures on reading comprehension. The 16 studies reviewed included 2,227 subjects and incorporated a total of 122 measures of association between the use of adjunct pictures and reading comprehension. The overall results across all studies revealed only minimal positive effects on reading text and subsequent reading comprehension when using adjunct pictures. The magnitude of picture effects were more substantial for university subjects who read text containing adjunct pictures.

One of the most comprehensive reviews of the effects of illustrated text on learning was done by Levie and Lentz (1982). The Levie and Lentz (1982) review compared three separate areas concerning the role of illu$tration in learning: (a) learning illustrated text information, (b) learning nonillustrated text information, and (c) learning using a combination of illustrated and nonillustrated text information. Studies included in the Levie and Lentz review cover a time period from 1938 to 1981. Levie and Lentz also present a functional perspective, which could be used to explain how illustrations might function to facilitate learning. Functional frameworks will be covered in detail in a later section of this chapter.

Summarizing die results across all studies included in their review, Levie and Lentz (1982) drew three primary conclusions: (a) Learning will be facilitated when the information in the written text is depicted in the illustrations; (b) learning of text material will not be helped nor necessarily hindered with illustrations that are not related to the text; and (c) when the criterion measure of learning includes both illustrated and nonillustrated text information, a modest improvement may often result from the addition of pictures.

Using Levin's (1981) framework to classify pictures according to the function they serve in prose learning, Levin, Anglin, and Carney (1987) conducted a meta-analysis of the pictures in prose studies. The reviewers concluded that for pictures (not mental images), serving a representation, organization, interpretation, or transformation function yielded at least moderate degrees of facilitation. A substantial effect size was identified for the transformation function.

One of the most significant programs of research on visual leaming has been conducted by Dwyer and his associates (Dwyer, 1972, 1978, 1987; Levie & Lentz, 1982; Rieber, 1994)., The research program. is unique in several ways, The studies in the Dwyer series used similar stimulus materials. In particular, the stimulus materials included a 2,000-word prose passage describing the parts, locations, and functions of the human heart along with various types of visual materials including line drawings, shaded drawings, and photographs in black and white and in color. The materials were delivered in a number of formats and combinations including: written prose with illustrations, a slidetape program with audio and television, and computer-based. In addition, a rationale was provided for the inclusion of visual illustrations in the treatments. If the information tested in a particular section of the text material was not difficult for the student (did not require external visualization), visual information would not be included and tested for this section of the text. Several types of criterion measure were developed by Dwyer and his associates, including a drawing test, an identification test, a terminology test, and a comprehension test. The research has been conducted with over 48,000 students (Dwyer, 1972, 1978, 1987).

Levie and Lentz (1982) conducted a meta-analysis using the treatments developed by Dwyer and presented in a text format or programmed booklet. All studies included in the meta-analysis included a text-only condition. Based on 41 comparisons of treatments with text-plus-prose, with text only using four criterion measures (drawing test, identification test, tem-tinology test, comprehension test), Levie and Lentz (1982) report that 36 comparisons ' favor illustrated text, and 4 favor text alone (see Appendix Table 26-1).* As with other reviews of literature discussed, one conclusion that can be drawn from the work of Dwyer and his colleagues is that visuals are "effective some of the time under some conditions" (Rieber, 1994, p. 132). Space limitations do not permit a more detailed discussion of the Dwyer series (Dwyer, 1972, 1978, 1987). *All tables are gathered in an appendix at the end of the chapter.

26.4.2.2. Guide to the Liteiature: Static Illustrations. Based on our literature search, 90 studies investigating the role of static pictures in knowledge acquisition were identified (see 12.2). The 90 studies were conducted with more than 13,528 subjects ranging from elementary school children to adults (see Appendix Table 26-3.) All of the studies static visual illustrations of various types with a prose-only included at least one comparison of leaming with prose and treatment. A number of the studies included written prose materials, while other included prose presented orally. It should be noted that many of the studies summarized included other comparisons irrelevant to this review, and they are not discussed. In the 118 experiments included in the 90 studies, 102 significant effects for treatments including text and visual illustrations vs. text only were identified. The results of the "box score" summary indicate that static visuals can have a positive effect on the acquisition of knowledge by students. The treatments used were varied, and many of the studies were not based on a particular theoretical perspective. In many of the studies, it was not possible to identify the role or ftinction of the visual illustrations in the instructional treatments. Examples of visuals and criterion measure items should more regularly be included in published studies. It was also difficult to determine what type of information was tested using die criterion measures in many of the studies. The reliability coefficients of the criterion measures were infrequently reported in the studies reviewed. In addition, few of the studies have been replicated. Notable exceptions are the research programs of Dwyer and Levin. A more detailed summary of each study is reported in Appendix Table 26-2. The studies by Dwyer and his associates that are reported in Table 26-1 are not duplicated in Table 26-3.

Based on our review of reviews of the literature and our own literature summary concerning the role of visual illustrations and knowledge acquisition, we still agree with a conclusion stated by Levie (1987) that:

It is clear that "research on pictures" is not a coherent field of inquiry. An aerial view of the picture research literature would look like a group of small topical islands with only a few connecting bridges in between. Most researchers refer to a narrow range of this literature in devising their hypotheses and in discussing their results. Similarly, authors of picture memory models, for example, take little notice of theories of picture perception (Levie, 1987, p. 26).

One of the primary reasons much of the research on the role of visual illustrations in knowledge acquisition is not easily integrated is that the role or function of the pictures and illustrations in the instructional treatments are not identified. We feel that it is critically important to determine, in advance of conducting research, the particular functions of the visual illustrations.

26.4.2.3. The Use of Functional Frameworks in Static Visual Research. In spite of the considerable amount of research concerning how static visuals facilitate learning, many empirical research studies reflect an unclear perception on the part of researchers of the manner in which illustrations ftinction in facilitating learning. A number of researchers have provided a variety of functional frameworks that may provide assistance in classifying static visuals into meaningful functional categories (Alesandrini, 1984; Brody, 1984; Duchastel. & Waller, 1979; Levie & Lentz, 1982; Levin, 198 1; Levin et al., 1987). We will provide a brief summary of several functional frameworks.

Two taxonomies have been proposed which take a morphological approach (what an illustration physically looks like) to picture classification (Fleming, 1967; Twyman, 1985). But classifying the role of pictures on the basis of "form" rather than "function" has not proved to be very useful (Duchastel & Waller, 1979). According to Duchastel and Waller, what is needed is not a taxonomy of illustrations but a grammar of illustrations that provides a functional set of principles that relate illustrations to the potential effects they may have on the learner.

Duchastel (1978) identified three general functional roles of illustrations in text: (a) an attentional role, (b) a retentional role, and (c) an explicative role. The attentional role relies on the fact that pictures naturally attract attention. The retentional role aids the learner in recalling information seen in an illustration, and the explicative role explains, in visual terms, information that would be hard to convey in verbal or written terms (Duchastel & Waller, 1979). Duchastel and Waller (1979) concluded that the explicative role of illustrations provides the most direct means with which to classify the role of illustrations in text. Seven subfunctions of explicative illustrations were identified by Duchastel and Waller (1979, pp. 21-24). These explicative subfunctions; are:

1. Descriptive. The role of the descriptive function is to show what an object looks like physically.
2. Expressive. The expressive role is to make an impact on the reader beyond a simple description.
3. Constructional. The intent of the constructional role is to show how the parts of a system form the whole.
4. Functional. The functional role allows a learner to follow visually the unfolding of a process or the organization of a system.
5. Logico-mathematical. The purpose of this role is to show mathematical concepts through curves, graphs, etc.
6. Algorithmic. The algorithmic role is used to show action possibilities.
7. Data display. The functional role of data display is to allow quick visual comparison and easy access to data such as pie charts, histograms, dot maps, or bar graphs.

An alternative functional framework offered by Levie and Lentz (1982) suggests a functional framework that includes classifying illustrations in text based on how they impact a learner in attending, feeling, or thinking about the information being presented. Their framework contains four major functions: (a) attentional, (b) affective, (c) cognitive, and (d) compensatory. The attentional function attracts or directs attention to the material. The affective function enhances enjoyment or in some other way affects emotions and attitude. Illustrations serving a cognitive function facilitate learning text content through improving comprehension, improving retention, or providing additional information. The last functional role identified by Levie and Lentz is the compensatory role used to accommodate Poor readers. Levie and Lentz, after reviewing a large number of studies containing 155 experimental comparisons of learning, have found much empirical support for the utility of their functional framework. Such a framework can help researchers sort out the functions that illustrations perform and can be used to identify the ways illustrations should be designed and used for specific cases (Levie & Lentz, 1982).

A functional framework that has proved to be useful in explaining differences in research studies concerning pictures and prose is provided by (Levin, 198 1). Levin contended that different types of text-embedded pictures serve five prose learning functions. The five functions identified by Levin are: (a) decoration, (b) representation, (c) organization, (d) interpretive, and (e) transformation. The decoration function is associated with text-irrelevant pictures (e.g., pictures used to make a written text more attractive) and does not represent the actors, objects, and activities happening in the.text. Representational pictures are associated with text-relevant pictures and do represent the actors, objects, and activities happening in the text. The role of organizational pictures is to provide an organizational structure giving the text more coherence. Interpretational pictures serve to clarify passages and abstract concepts or ideas that are hard to understand. Transformational pictures are first, unconventiopal, and not often found in traditional textbooks. Transforinational pictures are designed to have a direct impact on a learner's memory (e.g., pictures used as a runemonical aid serve a transformation function).

After reviewing the frameworks offered by Duchastel, Levin, Levie and Lentz, and others, Brody (1984) suggests that many of the specific functions identified within these frameworks do not clarify how pictures function in instructional settings. First, some functions are too broad or general in nature and add little to gaining an understanding of the instructional roles served by visuals. As an example, Brody contends that a single picture can increase comprehension in multiple ways such as gaining attention, repeating information, offering new information, and providing additional examples, A broad functional role such as increasing prose comprehension does not provide an adequate explanation of how a picture is to be used to affect prose comprehension (Brody, 1984). Brody (1984) also suggests that many previously defined functional roles of pictures are often too narrow in their view. In an effort to amelionate the limitations of previously identified functional roles of pictures, Brody offers his own set of representative instructional functions served by illustrations. Brody's approach to creating a potentially more useful functional framework was to identify functions in terms of what occurs during the instructional process. Another prime objective was to make the functional framework as general as possible in scope, that is, to make the functions independent of the specific form of instruction, content area, or types of learning skills being taught. Brody identified 20 representative instructional functions served by pictures. A potential problem with Brody's classification system for determining the role of illustrations in instructional materials is that it already contains a large number of categories. To extend his classification scheme ftirther would make it less practical for identifying the role of pictures in either research or instructional design practice.

Alesandrini (1984) states that some of the previous ftinctional frameworks dealt only with representational pictures, that is, pictures that represent the actors, objects, and activities taking place in the text. Alesandrini notes that other frameworks also include arbitrary or nonrepresentational roles of pictures such as graphs and flowcharts in the functional mix. Alesandrini offers a functional framework based on how instructional pictures convey meaning. Based on previous work by Grooper and Knowlton, Alesandrini classifies the role of instructional pictures into three functions: (a) representational, (b) analogical, and (c) arbitrary. Representational pictures can convey information in a direct way through tangible objects or concepts, or indirectly by the portrayal of intangible concepts that have no physical existence. Photos and drawings, or models and manipulatives, are examples of representational illustrations. Analogical pictures convey meaning by acting as a substitute and then implying a similarity for the concept or topic being presented. Arbitrary pictures (sometimes referred to as logical pictures) are highly schematized visuals that do not look like the things they represent but are related in some conceptual or logical way. Arbitrary illustrations include schernatized charts and diagrams, flowcharts, tree diagrams, maps, and networks.

26.4.2.4. Static Visuals and Knowledge Acquisition: Conclusions. Based on the conclusions of our review of earlier literature reviews and die studies we summarized in Tables 26-1 and 26-3, we conclude that static visual illustrations can facilitate the acquisition of knowledge when they are presented with text materials. However, the facilitative effects of illustrations are not present across all learning situations. It is very difficult to integrate the results across all studies due to the lack of connections (theoretical or functional) between many of diem. We do offer the following broad conclusions regarding the effects of illustrated visuals on learning: (a) Illustrated visuals used in the context of learning to read are not very helpful; (b) illustrated visuals that contain text-redundant information can facilitate learning; (c) illustrated visuals that are not text-redundant neither help nor hinder learning; (d) illustration variables (cueing) such as size, page position, style, color, and degree of realism may direct attention but may not act as a significant aid in learning; and (e) there is a curvilinear relationship between the degree of realism in illustrations and the subsequent learning that takes place.

There has been substantial progress in understanding how static illustrations affect the learning process. However, much remains to be done. Validations for many of the functional frameworks summarized in this chapter need to be completed. Theory-based studies that are informed by both memory research and theories of picture perception are lacking. Specific studies incorporating a particular theory of picture perception and a particular memory model need to be conducted- Theory-based research will provide us with a deeper understanding of the mechanisms that contribute to the effectiveness or ineffectiveness of static illustrations in instructional materials. It is also not clear how students use illustrations in instructional materials or that they even know how to use diem. A number of methods, including eye movement measurements, student surveys, and simply questioning students while they are using visual illustrations, win provide useful data on how students use or do not use illustrations. These data will be complimentary to the results of the recall and comprehension studies already completed. In addition, studies are needed that attempt to identify effective strategies for using illustrations included in instructional materials. Assuming that strategies for effectively using illustrations are identified, studies will then be needed that consider effective ways to train students to use these strategies. The issue of what constitutes "realisrW' in illustrations also needs to be reconsidered in light of the theories of picture perception discussed. in this chapter. Many of the criterion measures (recall or comprehension tests) are administered immediately after the presentation of the instructional treatments. It is also important to determine if the illustration effects identified in many of the studies reviewed in this chapter are durable over time. Finally, few of the studies reviewed systematically controlled for the type of text or picture included. Perhaps the effects of illustrations on learning will vary according to the type of prose passage or picture used.

26.4.3 Dynamic Pictures and Knowledge Acquisition

In this section, we first review the early research on the effect of dynamic visuals on learning (see also 1.5, 16.11). Next we will summarize more recent reviews of the literature concerning the role of dynamic visual displays and knowledge acquisition. Finally, we present the results of our literature search and analysis.

26.4.3.1. Dynamic Pictures and Knowledge Acquisition: Literature Reviews. Early studies examining the effects of dynamic visuals on learning can be found in instructional film research. Freeman (1924) summarized 13 research studies that compared the effectiveness of various forms of visual instruction. The treatment formats used in the 13 studies included film, slides, lectures, still pictures, prints, live demonstrations, and sterepgraphs. The motion treatments in these studies included the use of action pictures, animated drawings, and maps or cartoons. Based on the results of the 13 studies, it was concluded that motion or animated sequences in film are effective when (a) motion is a critical attribute of the concept being presented, and (b) motion is used to cue or draw the viewer's attention to the material being presented. It should be noted that die methodologies used in the 13 studies do not meet current standards for conducting comparative experimental research. A number of other investigators have conducted instructional fihn researchtW examined the effect of dynamic visuals on learning (Lumsdaine, Sultzer & Kopstein, 1961; May & Lumsdaine, 1958; Weber, 1926). Several conclusions can be drawn based on the early research on the role of dynamic visuals in instructional materials, including that (a) animation (motion) can lead to positive learning effects if it is a critical attribute of the concept(s) being presented, (b) animation (motion) can increase teaming of a complex procedural task, and (c) motion or action used primarily to enhance the realism of the presentation does not appear to have a significant effect on teaming. It should be noted that the conclusions drawn are based on a limited number of studies in which the motion variables were not usually tightly controlled.

Rieber (1990) summarized the results of 13 empirical studies investigating the role of animated graphics in computer-based instruction. Significant effects for animated treatments were found in 5 of the primary research studies reviewed. Based on the results of the 13 studies reviewed, Rieber presented three design recommendations for the use of animated visuals in instructional materials, including that: (a) "animation should be incorporated only when its attributes are congruent to the teaming task" (Rieber, 1990, p. 79); (b) "evidence suggests that when learners are novices in the content area, they may not know how to attend to relevant cues or details provided by animation" (Rieber, 1990, p. 82); and (c) "animation's greatest contributions to CBI may lie in interactive graphic applications (e. g., interactive dynamics)" (Rieber, 1990, p. 82).

As discussed in the review of static visuals, a number of frameworks have been provided to classify static visual material. A similar functional approach would be appropriate for dynamic visual research. Rieber (1990) suggests that "generally, animation has been used in instruction to fulfill or assist one of three functions: attention-gaining, presentation, and practice" (p. 77).

More recently, Park and Hopkins (1993, p. 19) identified five important instructional roles of animated visuals:

1. As an attention guide, the animated visual can serve to guide and direct the subject's attention.
2. As an aid for illustration, dynamic visuals can be used as an effective aid to represent the structural and functional relations among components in a domain of knowledge.
3. As a representation of domain knowledge, movement and action can be used to effectively represent certain domain knowledge.
4. As a device model for forming a mental image, graphical animation can be used to represent system structures and functions that are not directly observable (e.g., blood flowing through the heart).
5. As a visual analogy or reasoning anchor for understanding abstract and symbolic concepts or processes, animation can make abstract and symbolic concepts (e.g., velocity) become more concrete and directly observable.

When both the characteristics of the domain knowledge and the characteristics of the subjects require one or more of the five instructional roles described above to be used, then animated visuals will most likely be effective (Park & Hopkins, 1993).

Using their functional framework, Park and Hopkins produced a research summary of 25 studies investigating the effects of dynamic versus static visual displays. The delivery medium for 17 of the studies was computer-based instruction, while the delivery medium for the remaining 8 studies was film or television. Fourteen of the studies yielded significant effects for dynamic visual displays. However:

The research findings do not consistently support the superior effect of dynamic visual displays. The conflicting findings seem to be related to the different theoretical rationales and methodological approaches used in various studies (Park & Hopkins, 1993, p. 427).

One of the most interesting and rigorous programs of research on the effect of animation on teaming has been conducted by Rieber (Rieber, 1989, 1994). The animation research conducted by Rieber included students across age groups, with realistic instructional content (Newton's laws of motion) and higher-level teaming outcomes. As with the static visual research of Dwyer and his associates, the Rieber series of studies used animated graphics only when there was a need for external visualization. Results from the Rieber series are mixed and do not support the use of animated graphics across the board.

In summary, conclusions drawn from early reviews of the animation research literature are mixed. Rieber (1990) states that "die few serious attempts to study the instructional attributes of animation have reported inconsistent results.... CBI designers ... must resist incorporating special effects, like animation, when no rationale exists" (p. 84).

26.4.3.2. Guide to the Literature. Forty-two studies were located which included at least one animation treatment. Information concerning the author, treatments, subjects, and results is reported in Table 26-5. Initially, we attempted to classify the animated treatments according to the function they performed (Park & Hopkins, 1993). However, we later abandoned the approach due to lack of specific information concerning the treatments. It was also difficult to classify many of the animated treatments as performing a single role using the classification system.

From the group of 42 studies, a total 45 comparisons were identified which included at least one animation treatment. Significant animation effects were identified in 21 of the ' se comparisons. Animated treatments used by investigators have included various visual content such as animated illustrations, diagrams and visuals, real-time motion graphics, dynamic spatial visualization graphics, and animated interactive maps with blinking dots. General content areas covered by these studies include general science, physics, geometry, mathematics, statistics, and electronics. Subjects for these experiments ranged from mature adults to primary school children in the first, second, and third grade. A variety of tests were used to measure teaming outcomes, including (a) teaming of facts, concepts, and procedures; (b) problem solving and visual thinking; and (c) acquisition of cognitive skills that are primarily spatial or perceptual in nature.

How can the mixed results of the animation research be interpreted? Based on these "box score" results only, one could conclude that the use of animated graphics does not facilitate learning. However, methodological issues need to be considered. For example, in many of the studies it was not indicated if it was determined that there was a need for external visuals, static or dynamic. Perhaps reading text alone is adequate. In addition, many of the investigators did not provide a rationale for why motion is needed to either indicate changes over time or changes in direction. Text or text-plus-static graphics may be the optimal treatment if motion is not required. Many of the research reports reviewed did not specifically indicate that the animated sequences were text relevant or at least congruent with the text information presented. Also, both the information tested and the test type are critical considerations when investigating the learning effects for both. static and animated graphic displays. It was not always possible to determine if the information tested was presented only in the animation, only in the animated sequence, or presented in both. It was also difficult to determine the function of the animated sequences. Using the lessons learned from static graphic research, more attention needs to be given to the functional role of animated sequences in research studies.

Such methodological problems call into question the results of these studies reporting insignificant animation effects. We believe that the comments of Rieber, and of Park and Hopkins, are still timely and appropriate. Rieber (1990) stated that "while speculative explanations for these studies which did not produce effects have been offered, many rival hypotheses linger rooted in general procedural flaws such as poor conceptualization of the research problem or inappropriate implementation of methods" (p. 84).

In a later review of the literature, Park and Hopkins (1993, p. 439) suggested that:

Probably the most profound discrepancy separating the research is theoretical in nature. One important difference between studies which found significant effects of DVDs [dynamic visuals] and studies which found no such effects is that the former were guided by theoretical rationales which derived the appropriate uses for dynamic and static features of visual displays and their presumed effect. Accordingly, leamer variables, the learning requirements in the task, and/or the medium characteristics were appropriately coordinated in most of the studies that found significant effects.

As is the case for static graphics, it is clear that facilitative effects are not present for animated treatments across all learning situations.

26.4.3.3. Dynamic Visuals and Knowledge Acquisition: Conclusions. Unlike research pertaining to static visuals, which encompass many additional studies and dozens of treatment conditions, research on the effects of dynamic visuals is very limited. The early research lacked appropriate controls, so that the specific effects of animation on learning cannot be deten-nined. Results from the limited number of completed studies of the effect of dynamic visuals on learning are mixed. As discussed earlier, a number of the studies are methodologically flawed. Thus, the verdict is still out on the effect of animated treatments on student learning.

More work needs to be completed concerning the functions of animated visuals in learning materials. Contributions of Rieber, and of Park and Hopkins, have provided a starting point for further work. Refinement and validation of the functional frameworks suggested by Rieber, and by Park and Hopkins, are needed. In addition, it has not been demonstrated if or how learners use an animated sequence in the learning process. The effect of experience, prior knowledge, and aptitude patterns on the effective use of dynamic visual displays needs to be considered. Also, will students who are naive to specific instructional content be able to determine that an animated sequence indicates changes over time or changes in direction, and relate these changes to the specific content they are learning? Perhaps students need specific training on how to use animated sequences for learning. In almost all of the animation studies we reviewed, students in an animated treatment condition received visualized instruction (an animated sequence) and then were tested verbally. It is an open question whether a verbal test covering content displayed in a visual animated sequence measures the learning that has occurred. Also, many animated sequences particularly in simulations include a significant amount of information incidental to the particular purpose of the instructional package. Studies investigating the effect of such animated treatments on incidental learning are needed. Few of the animation studies we reviewed considered the effects of developmental level on learning. Animated treatments may differentially affect older vs. younger students. Finally, as discussed earlier, Rieber has suggested that animation may be most effective in computer-based instruction when used in interactive graphic applications. Much work needs to be done in this promising area of inquiry. In any case, future research investigating the effect of dynamic visual displays on learning should: (a) be based on a functional framework (i.e., Rieber, Park & Hopkins), (b) include content for which external visual information is needed and which requires the illustration of motion or the trajectory of an object, and (c) control for the effect of static graphics.

While some progress had been made, it is apparent that we know very little about the effect of dynamic visual displays on student learning. Givep the proliferation of visual information in instructional material, it is imperative that the most effective strategies for using animated visuals be determined. Relative to the production of static visuals and text materials, the cost of producing animated sequences is high. CaraballoRios (1985) stated that "insisting on the use of computer animation in cases where it is not absolutely necessary should be considered an extravagance" (p. 4). Many additional theorybased studies, including a range of content areas, audiences, treatment conditions, and learner characteristics, are needed.