Video Imagery and Children's Acquisition of Televised Geographic Information: Affecting More Than Visual Content.

This study examined whether televised imagery can improve children's acquisition of visual-spatial geographic information. It also investigated whether salient imagery hinders or facilitates processing of auditory information that is unrelated to the image. Fifty second graders and 40 fourth and fifth graders watched a videotaped program on US geography. For half the states shown, imagery was included to enhance elaboration and retrieval of state shapes. As expected, imagery aided in recall of state shapes but not locations; older children learned more from imagery than younger children. Recall of auditory information, when presented in conjunction with imagery, was enhanced during immediate recall, but was depressed when measured two weeks later. Support for the visual superiority and conjoint retention hypotheses are examined. Implications for instructional program production are discussed.

This study was conducted to determine whether the use of imagery aided children's learning of televised geographic information. As learning and memory research have shown, a learner who uses imagery to incorporate the material to be learned has better recall than one who does not use it (c.f. Paivio, 1971; Pressley & Levin, 1980). With television, one can go further and impose such imagery on important content rather than relying on viewers to generate their own imagery spontaneously. The technique of imposed imagery should be helpful particularly for young children who may lack the cognitive capacity to apply imagery adequately (Bender & Levin, 1976; Pressley, 1982, Rohwer, Kee, & Guy, 1975).

The second goal of this study was to determine whether imposed imagery enhances or interferes with the acquisition of auditory information. Visual imagery may provide referential links to information presented in other modalities, thus improving recall of non-visual information (Abel & Kulhavy, 1986, 1989; Kulhavy, Lee, & Caterino, 1985; Webb, Saltz, McCarthy, & Kealy, 1994). A contrasting hypothesis proposes that by focusing the learner's attention and processing capacities upon visual information, imposed imagery may detract from processing information simultaneously presented in non-visual modalities. In particular, cross-modality research provides evidence that visual information can sometimes interfere with recall of auditory information (Greenfield & Beagles-Roos, 1988; Pezdek & Stevens, 1984).

Imagery as an Imposed Organizer

Imagery provides an economical framework for simultaneously representing all aspects of perceptual/spatial information (e.g., an intact image of a map) and for integrating the relations between the visual and verbal information (Kulhavy, et al., 1985; Paivio, 1971; 1986). Fifth graders who studied a map with icons (e.g. a cross representing the location of a church) and with labels (e.g. the word church alongside the icon) were able to recall more information when cued with both icons and labels than with just the icons alone (Webb et al., 1994). In addition, children's recall of information found only in the narrative improved. In this case, verbal as well as spatial content were encoded with images.

Does Imagery Hinder Auditory Recall?

Visual superiority hypothesis. While children's recall of visual information may benefit from imposed visual elaborations, there is a second issue that the present study addresses as well. The evidence from two bodies of research leads to different predictions on the recall of verbal information in relation to visual information. According to the visual superiority hypothesis, when information is presented in multiple modalities and the individual perceives the visual input as adequate for responding, attention and processing are directed toward vision (Klein & Posner, 1974; Pick, Warren, & Hay, 1969; Rock & Victor, 1964).

The visual superiority effect is best exemplified when the visual and auditory modalities are unrelated. When the auditory and visual modes of a TV program were unrelated, 5-year-olds favored the visual mode, as evidenced by lower comprehension scores on auditory information than when the audio mode was presented alone (Pezdek & Stevens, 1984). Visual information may detract from attending to or processing irrelevant information that is simultaneously presented via other modalities.

Conjoint retention hypothesis. In the majority of cases, however, television presents us with simultaneous auditory and visual information that are related. When visual and auditory presentations are complementary, one mode can serve to enhance attention to and comprehension of the other. In contrast to the visual superiority hypothesis, the conjoint retention hypothesis suggests that multimodal presentations may enhance recall of information.

Research on children's map retention indicates a facilitative relation between maps and related text (Webb et al., 1994). According to the conjoint retention hypothesis, geographic information learned through maps and text provide dual-modality representations, which in turn aid in recall of information presented in either modality (Kulhavy et al., 1985). The conjoint retention hypothesis implies that because the perceptual/spatial structure is created and organized separately from the linguistic/verbal structure, activation of one information store should not interfere with activation of the other and in fact should facilitate its access.

The conjoint retention hypothesis is derived from dual coding theory (Paivio, 1986). Visual stimuli--in this case, map features--are encoded separately from verbal stimuli--the narrative text--yet maintain referential links to one another. The configuration and features of a map therefore provide a mnemonic framework for associating subsequent text information. Research supports the notion that map features can serve as a secondary cue for recalling related verbal information (cf. Kulhavy, Stock, Woodard, & Haygood, 1993). However, the hypothesis is bound by a training technique in which the visual and verbal stimuli are presented sequentially rather than simultaneously. The conjoint retention hypothesis may not be supported in a situation where modalities are presented simultaneously, as is typically the case with televised instruction.

The purposes of this study are to test the hypothesis that imagery will improve children's recall of visual-spatial information and to evaluate whether imagery will enhance or interfere with acquisition of unrelated auditory information. Geographic information about the United States was taught using a videotaped program. Children's recall of visual as well as auditory information was measured. The data reported here are part of a larger study designed to measure the effects of visual presentation on children's recall of geographic information.

Method

Participants

Ninety children were recruited from an elementary school in Lawrence, Kansas. The younger cohort included 50 second graders (M = 8.06 yr.), while the older cohort comprised 19 fourth graders and 21 fifth graders (M = 10.45 yr.). There were 22 girls and 28 boys in the younger cohort; 21 girls and 19 boys were in the older cohort. One girl from the younger cohort was dropped from the study because she failed to complete the posttest questionnaire.

Stimuli and Materials

Stimuli consisted of a 10-min. instructional videotape that depicted geographic information about eight states from the contiguous US The eight states were selected from different regions in the country excluding Kansas and its adjacent states. Each state was shown for approximately two minutes. The program was created using production techniques known to elicit and maintain children's attention (e.g., female narrator, lively music, visual special effects) (Levin & Anderson, 1976).

Information presented only on the narrative included state capitals and nicknames. Instruction on half of the states included elaborative imagery designed to help recall state shapes. For example, the shape of Vermont was overlaid by an image of the letter V. The other four states were presented without such imagery. To reduce order effects, the presentation of states with and without imposed imagery were presented in random order.

Pretest and posttest materials. A series of tasks were designed to measure children's geographic knowledge. Children were asked to place an adhesive dot on a map to mark where each of the eight trained states were located (State location recall). The map consisted of only the outer border of the 48 contiguous states; no state outlines were included. In the second task, children were shown a map of the continental US that included state outlines and were told to point to the eight states as the experimenter named them (Cued location recall). Children were then shown the US map with individual states highlighted and were asked to name each state as the experimenter pointed to them (Name recall). Following name recall, children received corrective feedback then were asked to recall the state's capital and nickname.

The pretest and posttest materials were identical except the posttest included three additional tasks designed to measure the differential effects of instruction. Children were shown the contours of 18 states: 8 were included in instruction, 10 were not. Children were asked to name isolated state contours shown out of the geographical context of the US map (Contour identification). For those states that were presented with imagery, participants were then prompted, "What does the shape of that state remind you of?" (Imagery recall).

Two to three weeks following instruction, a geography test was administered to the entire second, fourth, and fifth grades, providing a retention measure. Because the test was given to the class as a group, the retention test used a variation of the dot placement task and a multiple-choice questionnaire that included the same questions found in the regular posttest, but in a different order. Questions about states not shown in the video were also included.

The state contour task involved showing children the contours of 12 states--including the eight states shown on the videotape--and a list of 16 state names. Children were asked to match the shape of each state with its correct state name. All retention test items were read aloud to the second graders. The instructions for each task were read aloud to the fourth and fifth graders; those who had difficulty reading or understanding the test items were assisted individually.

Scoring. For the dot placement task (State location recall), the width of the adhesive dot (1/4 in. diameter) was the unit of measure. Five concentric circles were drawn around each state, with the smallest circle circumscribing the state. The radius of each successive circle increased cumulatively so each radius was 1/4 in. larger than the previous radius increase. The highest score of 8 points was achieved by placing the dot entirely within the state boundary; 7 points were given if the dot landed on the state boundary; 6 points were given if the dot was outside the state boundary but within the smallest circle, and so on. Children who placed a dot outside the fifth circle but within the US received 1 point. For all other recall tasks, participants were awarded 1 point for a correct response, 0 for incorrect. For each task, scores for states trained with imposed imagery were summed, producing a maximum score of 4 points. Scores for the four states trained without imagery were also summed. The contour identification task required summing scores by imagery instruction, no imagery instruction, or no instruction. Summary scores for the "no instruction" states were divided by a constant to make their scores directly comparable to those of instructed states.

Procedure

Participants were pretested individually immediately before viewing the instructional program. Viewing was done in groups of three to five students, rather than individually, to simulate the way television is used in a classroom setting. During instruction, participants were seated together opposite a video monitor where the experimenter presented the televised stimuli. Prior to viewing, children were told to watch the program carefully because they would be tested on the material afterwards. Immediately following viewing, participants were administered the posttest individually. Two to three weeks later, the retention test was administered to each classroom as a group.

Results

Summary scores for each dependent variable were analyzed according to a 2 (Cohort) x 2 (Imagery/ No Imagery) x 2 (Pretest/Posttest) repeated measures analysis of variance with imagery as a within-subjects factor, and pretest/posttest scores as a repeated measure. All participants with valid data for an individual task were included in that set of analyses.

State Identification and Location

State location recall. The dot placement task was designed to measure children's accuracy in the geographic location of states when no visual or verbal cues were available. The means for this and the other tasks are shown in Table 1. Because information on state locations was not presented in conjunction with imposed imagery, no effect for imagery was expected. And indeed, imagery effects were nonsignificant. However, there were significant main effects for cohort, F (1,87) = 5.05,p [is less than] .05, and time of test F (1, 87) = 157.98, p [is less than] .01, and a cohort x time of test interaction, F (1,87) = 5.67,p [is less than] .05. Post hoc tests showed that the two cohorts did not differ on the pretest with the older cohort having higher posttest scores than the younger cohort, F (1, 87) = 7.64, p [is less than] .01.

Table 1 Means (and Standard Deviations) for Geography Tasks by Time of Test, Cohort, and Type of Instruction

                                          Pretest

Group                             n       Imagery

Younger Cohort
  Dot Placement: State Location   49   16.22   (6.43)
  Pointing to States on the Map   49     .90   (1.07)
  Naming Highlighted States       47    1.25   (1.30)
  Naming States Out of Context    47
  Recall of Adjacent States       47     .24   (0.26)
  Recall of State Capitals        47     .06   (0.32)
  Recall of State Nicknames       47     .06   (0.25)
Older Cohort
  Dot Placement: State Location   40   17.93   (5.43)
  Pointing to States on the Map   40    1.33   (0.97)
  Naming Highlighted States       40    2.25   (1.28)
  Naming States Out of Context    40
  Recall of Adjacent States       40     .30   (0.21)
  Recall of State Capitals        40     .30   (0.61)
  Recall of State Nicknames       40     .13   (0.34)

                                      Pretest          Posttest

Group                                No Imagery        Imagery

Younger Cohort
  Dot Placement: State Location   17.47   (7.18)   21.63   (7.50)
  Pointing to States on the Map    1.39   (1.22)    2.35   (1.48)
  Naming Highlighted States        1.53   (1.44)    2.43   (1.53)
  Naming States Out of Context                      2.26   (1.51)
  Recall of Adjacent States         .38   (0.25)     .27   (0.28)
  Recall of State Capitals          .21   (0.59)     .19   (0.45)
  Recall of State Nicknames         .11   (0.31)     .79   (0.72)
Older Cohort
  Dot Placement: State Location   19.05   (6.25)   24.33   (6.25)
  Pointing to States on the Map    1.48   (1.28)    2.95   (1.20)
  Naming Highlighted States        2.23   (1.27)    3.53   (0.99)
  Naming States Out of Context                      3.23   (1.10)
  Recall of Adjacent States         .45   (0.21)     .41   (0.24)
  Recall of State Capitals          .68   (0.89)    1.20   (0.97)
  Recall of State Nicknames         .40   (0.55)    2.18   (1.08)

                                      Posttest

Group                                No Imagery    No instruction

Younger Cohort
  Dot Placement: State Location   21.04   (7.96)
  Pointing to States on the Map    2.25   (1.45)
  Naming Highlighted States        2.32   (1.53)
  Naming States Out of Context     1.87   (1.31)   1.04   (1.05)
  Recall of Adjacent States         .42   (0.26)
  Recall of State Capitals          .47   (0.72)
  Recall of State Nicknames         .62   (0.64)
Older Cohort
  Dot Placement: State Location   25.83   (5.56)
  Pointing to States on the Map    3.00   (1.20)
  Naming Highlighted States        3.38   (1.03)
  Naming States Out of Context     2.50   (1.28)   1.27   (1.06)
  Recall of Adjacent States         .55   (0.25)
  Recall of State Capitals         1.60   (1.28)
  Recall of State Nicknames        1.50   (0.96)

Note. The values for all but the Dot Placement task represent the total number of correct responses out of 4. The maximum score possible for the Dot Placement task is 32.

Cued location recall. Performance on tasks that relied on shape and location cues were expected to improve with imagery instruction. When asked to point to individual states on the map, the added imagery for half of the states improved children's ability to locate those states, as indicated by the imagery x time of test interaction, F (1,87) = 5.98, p [is less than] .05, and a main effect for time of test, F (1,87) = 202.61, p [is less than] .001. Post hoc analyses indicated the effect was a result of differences in imagery performance at pretest, F (1,87) = 10.26, p [is less than] .01, but not at posttest, F (1,87) = .06, n.s. As Figure 1 shows, the pre-post gain was not as great for those states instructed without imagery.

[Figure 1 ILLUSTRATION OMITTED]

The time of test main effect, along with a cohort main effect, F (1,87) = 4.33,p [is less than] .05, and a cohort x time of test interaction, F (1, 87) = 4.85, p [is less than] .05, showed that older children's performance improved more than younger children's. Post hoc analyses indicated the older children outperformed the younger children at posttest, F (1,87) = 6.48, p [is less than] .05. There were no higher-order interactions between cohort, imagery, and time of test, F (1, 87) = 3.02, p = .08.

Name recall. While the previous task required a motor response, naming states that are highlighted on the map required verbal recall of those states' names. Contrary to predictions based on conjoint retention hypothesis, imagery did not benefit performance. A cohort main effect was present, F (1, 83) = 15.90, p [is less than] .001, showing that older children could name more states than younger children. Also, both cohorts gained from pretest to posttest, F (1, 85) = 141.28, p [is less than] .001. No significant interactions or imagery effects were noted.

State contour identification. The test of state shape knowledge required identifying states when location cues were removed. The contours of ten states that were not shown in the video instruction were included as controls. Thus, the imagery variable had three levels: imagery, no imagery, and no instruction. This task was administered only at posttest.

The two cohorts were differentially affected by imagery instruction, as indicated by the cohort x imagery interaction, F (2, 170) = 5.46, p [is less than] .01, the cohort main effect, F (1,85) = 6.88, p [is less than] .05, and imagery main effect, F (2, 170) = 103.76, p [is less than] .001. Older children performed better than younger children when tested on instructed states, F (1,85) = 9.39, p [is less than] .01, and children did better on states trained with imagery than on states trained without it, F (1,85) = 24.76, p [is less than] .001. Figure 2 shows that the older cohort performed as poorly as the younger cohort on noninstructed states.

[Figure 2 ILLUSTRATION OMITTED]

Recall of Auditory, Information

Information presented through the auditory track (i.e., state capitals and nicknames) were not directly related to imagery instruction. The conjoint retention hypothesis predicts that information presented in conjunction with a salient visual image would be processed more deeply and thus recalled more readily than if no imagery were present. The visual superiority hypothesis predicts that auditory information presented simultaneously with a salient visual image would be attended to and processed less deeply than if no salient image were present.

State capitals. An imagery main effect, F (1, 85) = 14.05, p [is less than] .001, indicated that capitals were recalled better for those states that did not include imagery instruction (M = .71, SD = .90) than those that did (M = .41, SD = .61). However, because there was no interaction between imagery and time of test, the imagery effect was not a function of training. A cohort x time of test interaction, F (1,85) = 39.42, p [is less than] .001, along with main effects for cohort, F (1,85) = 32.91,p [is less than] .001, and time of test, F (1, 85) = 92.43, p [is less than] .001, indicated that older children scored higher and improved more than younger children. When the cohorts were analyzed separately, both showed significant pre-post gains: F(1, 46) = 20.99, p [is less than] .001, for the younger cohort, and F (1,39) = 63.54, p [is less than] .001, for the older cohort. Neither the conjoint retention nor the visual superiority hypothesis was supported.

Recall of nicknames. Because state nicknames were presented immediately preceding or following any imagery, recall of nicknames was more likely to be affected by visual content than other narrative information.

A triple interaction of cohort x imagery x time of test, F (1,85) = 10.65,p [is less than] .01, along with interactions of cohort x time of test, F (1,85) = 39.31, p [is less than] .001, and imagery x time of test, F (1, 85) = 26.48, p [is less than] .001, emerged, requiring post hoc analyses. Main effects for cohort, F (1, 85) = 59.20, p [is less than] .001, imagery, F (1, 85) = 4.53, p [is less than] .05, and time of test, F (1, 85) = 205.82, p [is less than] .001, were present.

An examination of the data by cohort showed that younger children's recall of nicknames increased from pretest to posttest, F (1,46) = 58.35, p [is less than] .001. Second graders' recall of state nicknames was not differentially affected by imagery, F (1,46) = 1.29, n.s. No interactions were present.

In contrast, the older cohort benefited from imagery instruction. While the imagery main effect was not significant, there was an imagery x time of test interaction, F (1, 39) = 23.50, p [is less than] .001, and a time of test main effect, F (1, 39) = 134.48, p [is less than] .001. As shown in Figure 3, gains from pretest to posttest were greatest for nicknames of states taught with imagery. For the older cohort, the conjoint retention hypothesis was supported.

[Figure 3 ILLUSTRATION OMITTED]

Delayed Posttest: State Locations and Shapes

A 2 (Cohort) x 3 (Instruction) MANOVA with type of instruction (imagery present, no imagery, no instruction) as a within-subjects factor was conducted oil the delayed posttest data.

State locations: The dot placement task. Older children were more accurate than younger children at dot placement, F (1,82) = 8.21, p [is less than] .01. There was a main effect for instruction, F (2, 164) = 3.31, p [is less than] .05. Post hoc comparisons indicated that recall for state locations shown with imagery was depressed; states taught without imagery and states not included in instruction were located with more accuracy than those trained with images, t (82) = 2.66, p [is less than] .05. The means for these items are in Table 2. There were no performance differences between states taught without imagery and those not included in instruction.

Table 2. Means (and Standard Deviations) for Geography Tasks at Delayed Posttest by Cohort and Type of Instruction

Group                             n         Imagery

Younger Cohort
  Dot Placement: State Location   46   18.46   (7.08)
  Matching Shapes with Names      45    2.02   (1.52)
  Recall of State Capitals        46    1.15   (0.99)
  Recall of State Nicknames       47    1.55   (0.95)
Older Cohort
  Dot Placement: State Location   38   21.74   (6.83)
  Matching Shapes with Names      35    2.83   (1.12)
  Recall of State Capitals        39    1.95   (1.05)
  Recall of State Nicknames       37    2.22   (0.98)

Group                                No Imagery      No Training

Younger Cohort
  Dot Placement: State Location   19.48   (8.10)   19.41   (6.41)
  Matching Shapes with Names       2.09   (1.51)    2.04   (1.24)
  Recall of State Capitals         1.26   (1.10)    1.37   (0.88)
  Recall of State Nicknames        1.75   (1.22)    1.45   (1.02)
Older Cohort
  Dot Placement: State Location   23.58   (6.39)   23.32   (5.35)
  Matching Shapes with Names       2.23   (1.42)    2.06   (0.87)
  Recall of State Capitals         2.67   (1.08)    1.95   (1.00)
  Recall of State Nicknames        3.14   (1.08)    1.46   (0.87)

Note. The values for all but the Dot Placement task represent the total number of correct responses out of 4. The maximum score possible for the Dot Placement task is 32.

Matching state contours with state names. Imagery was expected to aid in recall of state names and shapes. There was no cohort main effect, although an instruction main effect, F (2, 156) = 4.91, p [is less than] .01, and a cohort x instruction interaction, F (2, 156) = 6.00, p [is less than] .01, were present. Post hoc analyses showed no difference in the scores of younger children as a function of imagery instruction, but older children's scores differed markedly. For them, knowledge of states taught with imagery was better than for states taught without imagery, t (34) = 2.81, p [is less than] .01. Also, instructed states were recalled better than states not instructed, t (34) = 3.25, p [is less than] .01.

Retention of Auditory Information

Capitals. Retention of state capitals was not expected to be influenced by imagery. However, analyses indicated a cohort x imagery interaction, F (2, 166) = 5.27, p [is less than] .01. Main effects were also present for cohort, F (1, 83) = 33.27, p [is less than] .001, and imagery, F (2, 166) = 5.27) p [is less than] .001. Post hoc analyses showed that while the younger children's performance did not differ as a function of imagery instruction the older children's retention of state capitals was negatively affected by the presence of imagery, F (2, 76) = 8.79, p [is less than] .001. Further analyses revealed that the older cohort was better able to recognize the capitals of those states that were trained without imagery than those states with it, t (38) = 3.84, p [is less than] .001. As the means in Table 2 show, performance on states taught with imagery did not differ from states not instructed.

Nicknames. There were main effects for cohort, F (1,82) = 21.92, p [is less than] .001, and imagery, F (2, 164) = 21.90, p [is less than] .001, and a significant cohort x imagery interaction, F (2, 164) = 10.63, p [is less than] .001. The younger cohort's performance did not differ as a function of imagery; only the older cohort was affected, F (2, 72) = 33.56, p [is less than] .001. Older children knew more nicknames of states trained without imagery than those of states trained with imagery, t (36) = 4.38, p [is less than] .001, and performed better on trained states than untrained states, t (36) = 7.04, p [is less than] .001. The means for the older cohort are 2.22 (SD = .98) for imagery instruction, 3.14 (SD = 1.08) for no-imagery instruction, and 1.46 (SD =.87) for no instruction.

Discussion

In general, children's geographic knowledge improved as a result of instruction. All pretest-posttest measures showed significant gains. The improvement in performance cannot be explained by simple practice effects: Performance on a new task (e.g., matching state contours) was greater for states included in instruction than for states not included. While even the second grade children benefited from instruction, the fourth and fifth graders learned more and retained more from instruction than the younger children.

Visual-Spatial Recall as a Function of Imagery

Imagery facilitated some types of learning, in particular information related to state shapes. For both cohorts, imagery aided their ability to recognize and point to state locations on the map as well as identify individual state contours. In addition, the older cohort gained more from imagery than the younger cohort. When older children were asked to match state contours with state names, recall was higher for those states taught with imagery than without it.

The benefits of imagery remained for the older cohort. Two weeks following instruction, imagery continued to produce an advantage for the older children when asked to match state contours with their names. In accord with Webb et al.'s (1994) findings, the imagery used in the current study apparently produced a meaningful context for encoding visual information. Thus, individual images of state shapes were linked to the map via imagery; imagery served as a referent for shape retrieval.

Imagery did not aid in recall for certain tasks. Children were expected to benefit from imagery in recalling state names, but did not. The task was similar to identifying state contours with the exception that the name recall task showed states in the context of the US map. It was expected that state contour identification would be a more difficult task, yet imagery effects were present for this task but not for name recall. The outcome was unexpected and cannot be explained adequately at this time.

Imagery was not expected to aid children's performance on tasks that did not contain shape cues. For the location-only dot placement task, imagery did not contribute to learning. Thus, the initial hypothesis that imagery would aid in children's learning visual-spatial information was supported to some extent: imagery enhanced learning of visual information directly related to the imagery itself (e.g. shape). However, because the types of imagery used in this study did not include a spatial referent (e.g. an image that suggested state location), imposed imagery did not aid in location tasks.

Imagery and Recall of Auditory Information

The second purpose of this study was to determine if imagery assisted in retention of auditory information. According to the visual superiority hypothesis, visual information--including imagery--would be retained at the expense of competing auditory content. The conjoint retention hypothesis, on the other hand, predicted that the type of imagery included in this study would provide an additional link to auditory and visual information, aiding in the retrieval of material from both storage systems.

The outcomes provided moderate if mixed support for the two hypotheses. While younger children's knowledge of state nicknames and capitals improved after watching the instructional program, gains were not a function of imagery. Because the imagery main effect--present in the recall of capitals--was not a function of geographic instruction, interpreting the imagery effect at the delayed posttest is problematic. For the most part, second graders' recall of auditory content was low and was not differentially affected by imagery either immediately following instruction or two weeks later.

Narrative information presented concurrently with salient imagery was thought most likely to be affected by imagery. Therefore, because nicknames were presented either immediately before or after any imagery, the outcomes related to the recall of nicknames appeared to be critical in providing support for one or the other hypothesis. Older children learned more nicknames that were included as part of the narrative when imagery was present than when there was no imagery. Thus, the conjoint retention hypothesis was supported.

Older children's recall was still influenced by imagery even two weeks after instruction. There was, however, a unique reversal in recalling state nicknames as a function of imagery: nicknames of states trained without imagery were recalled better than those trained with imagery. Apparently, imagery produced enhanced processing of both auditory and visual information, as evidenced by improvement from pretest to posttest, but did not result in further elaboration of unrelated audio information (i.e., state nicknames). In fact, on the delayed posttest, recall of nicknames for states taught with imagery was worse than that for states not included in instruction, an indication of interference in learning. Mental elaborations provided by imagery may have aided learning of some audiovisual information (e.g. linking the image of the shape with the spoken state name) at the expense of auditory information that did not complement the visuals. In the case of state nicknames and capitals, the visuals did not depict the location of the capitals or elaborate upon state nicknames. Thus, for delayed recall, the visual superiority hypothesis was supported.

The design of the present study limited the interpretation of some results. The imagery main effect found for recall of state capitals shows that those states chosen for imagery instruction were simply more difficult to recall at pretest and posttest than those of the four states not chosen for imagery instruction. The results indicate that capitals were not learned as a function of imagery. The selection of states may have made for a more conservative test of the hypotheses; multiple training conditions that vary by the states chosen for training may produce an even stronger training effect due to imagery.

The results have important implications: Imposed imagery in televised programs can be used to aid in children's encoding and retrieval of certain types of information. However, while imagery may facilitate short-term recall of auditory content, the salient aspect of imagery may detract from further processing of auditory information, particularly when a verbal task is irrelevant to its visual emphasis. Thus, visual elaboration may actually interfere with long-term recall and performance in tasks where imagery is irrelevant to the content.

The results also bear important implications for the production and evaluation of educational media for children. When using imposed imagery, instructional program producers must be selective in its use and placement. Otherwise, these images might draw attention and processing away from critical information. In addition, if important information can be presented only auditorally (e.g., when a young viewer is unable to read words displayed on the screen), that information should either be integrated with a visual component (e.g., linking the state name with its shape) or else temporally separated from salient visuals that may detract from processing auditory information. Program evaluators should be cautious when examining the effects of imagery. Tests of immediate as well as delayed recall are necessary to assess the outcome of imagery on learning.

With the advent of computerized instructional programs for children, complete with vivid audiovisual presentations, imagery can be incorporated readily into instructional designs to facilitate learning. When used judiciously, imagery can be an effective tool in audiovisual modes of instruction.

References

Abel, R. R., & Kulhavy, R. W. (1986). Maps, mode of text presentation, and children's prose learning. American Educational Research Journal, 23, 263-274.

Abel, R. R., & Kulhavy, R. W. (1989). Associating map features and related prose in memory. Contemporary Educational Psychology, 14, 33-48.

Bender, B. G., & Levin, J. R. (1976). Motor activity, anticipated motor activity, and young children's associative learning. Child Development, 47, 560-562.

Greenfield, P., & Beagles-Roos, J. (1988). Radio versus television: Their cognitive impact on children of different socioeconomic groups. Journal of Communication, 38, 71-92.

Klein. R. M., & Posner, M. I. (1974). Attention to visual and kinesthetic components of skills. Brain Research, 71, 401-411.

Kulhavy, R. W., Lee, J. B., & Caterino, L. C. (1985). Conjoint retention of maps and related discourse. Contemporary Educational Psychology. 10, 28-37.

Kulhavy, R. W., Stock, W. A., Woodard, K. A., & Haygood, R. C. (1993). Comparing elaboration and dual coding theories: The case of maps and text. American Journal of Psychology, 106 (4), 483-498.

Levin, J. R. (1976). What have we learned about maximizing what children learn? In J. R. Levin & V. L. Allen (Eds.), Cognitive learning in children: Theories and strategies (pp. 105-134). New York: Academic Press.

Paivio, A. (1971). Imagery and verbal processes. New York: Holt.

Paivio, A. (1986). Mental representations: A dual coding approach. New York: Oxford University Press.

Pezdek, K., & Stevens, E. (1984). Children's memory for auditory and visual information on television. Developmental Psychology, 20, 212-218.

Pick, H. L., Warren, D. H., & Hay, J. C. (1969). Sensory conflict in judgments of spatial direction. Perception & Psychophysics. 6, 203-205.

Pressley, M. (1982). Elaboration and memory development. Child Development. 53, 296-309.

Pressley, M., & Levin, J. R. (1980). The development of mental imagery retrieval. Child Development, 51, 558-560.

Rock, I., & Victor, J. (1964). Vision and touch: An experimentally created conflict between the two senses. Science, 143, 594-596.

Rohwer, W. D., Jr., Kee, D. W., & Guy, K. C. (1975). Developmental changes in the effects of presentation media on noun-pair learning. Journal of Experimental Child Psychology, 19, 137-152.

Webb, J. M., Saltz, E. D., McCarthy, M. T., & Kealy, W. A. (1994). Conjoint influence of maps and auded prose on children's retrieval of instruction. Journal of Experimental Education, 62 (3), 195-208.

The author would like to thank the graduate students and staff at the Center for Research on the Influences of Television on Children (CRITC) at the University of Kansas as well as the students and teachers at Deerfield Elementary School for their assistance in the conduct of this research. Special thanks go to John C. Wright and Aletha C. Huston, directors of CRITC, and to Dennis Kerkman, for their assistance in the development of this manuscript.

Dr. Marites F. Pinon, Department of Psychology, Southwest Texas State University.

Correspondence concerning this article should be addressed to Dr. Marites F. Pinon, Department of Psychology, Southwest Texas State University, San Marcos, TX 78666. Email: [e-mail]