The effects of presentation format for behavior modeling of interpersonal skills in online instruction.

        The purpose of this study was to identify the most effective
        model presentation format in behavior modeling to teach
        interpersonal skills in online learning environments. Four model
        presentation formats were compared; video, pictures plus audio,
        audio only, and text-script only. The effects of the model
        presentation were investigated in terms of learning outcomes,
        which were measured by learners' reactions; cognitive retention
        of learning content; and behavioral reproduction. No significant
        differences between groups were found in any measure of learning
        outcomes. The implication of the findings is that it is
        reasonable to use cost-effective model presentation formats to
        teach interpersonal skills using behavior modeling. Also, three
        possible moderating variables of the relationship between model
        presentation formats and learning outcomes are discussed.

There is no doubt that interpersonal skills are important to an individual's professional life and the success of organizations. In general, interpersonal skills are emphasized across all fields in current society, not only in service-oriented occupations. Some of the interpersonal skills, such as leadership and communication skills, are required of employees regardless of their jobs. However, according to Industry Report 2002 (Galvin, 2002), interpersonal skills, such as coaching, communication skills, customer service, leadership, and mentoring, are rarely taught online despite its growing popularity.

The reason for this might be the difficulty of implementing effective instructional methods for interpersonal skills online. According to Gilmore and Fritsch (2001), while instructor-led courses teaching interpersonal skills use various instructional methods, such as presentation, facilitated discussion, reflection, discussion and laboratory, online courses dominantly use readings and drills. Readings and drills are appropriate instructional methods for self-study in the online learning environment. However, they are not considered to be effective methods of conveying interpersonal skills training because interpersonal skills require the affective, cognitive, and psychomotor domains to work together (Argyle, 1981). The ineffective instructional methods of teaching interpersonal skills online may lower students' motivation and achievement, hence resulting in online interpersonal skills training's unpopularity. Given the increasing popularity of online learning and the significance of interpersonal skills for every aspect of organizational life, it is imperative to identify effective instructional methods for online interpersonal skills training.

INTERPERSONAL SKILLS TRAINING

Interpersonal skills, in layman's terms, refer to the skills individuals need when they interact with other people. Also, the meaning of interpersonal skills has a number of synonyms, such as social skills, soft skills, people skills, or communication skills. The distinctions among the terms are not always clear, in many cases they are used interchangeably. Despite the significance of interpersonal skills in our life, research on interpersonal skills training is difficult to find. One explanation for this might be because interpersonal skills are used unconsciously as a part of daily life; people view the interpersonal domain as inborn skills or characteristics, not skills or knowledge to be acquired consciously. Another reason for this might be the overarching characteristics of interpersonal skills, which require the combination of cognitive, psychomotor and affective components simultaneously.

Many researchers explain that interpersonal skills have complicated characteristics. Marlowe's (1986) social intelligence model shows multidimensional characteristics as consisting of: care for others, social self-efficacy, and the ability to understand others. Similarly, Hargie (1986) identifies the following six characteristics of interpersonal skills: (a) goal-directedness and intentionality, (b) inter-related situationally appropriate social behaviors, (c) qualitative differences in social behavior, (d) a series of related skills in sequence, (e) the composition of the sub-skills, and (f) the influence of cognitive control.

Similarly, Bailey and Butcher (1983) emphasized that effective interpersonal skills training should focus on developing perceptual, cognitive, and behavioral skills components together. That is, learners need to make reasonable judgements at the perceptual level to respond in a timely manner. At the cognitive level, the learners should understand the purpose of an interaction and have clear ideas of what appropriate behaviors are in specific situations. Finally, the learners should demonstrate actual behaviors at the performance level to execute their individual interpersonal skills. These multidimensional characteristics of interpersonal skills require a different instructional approach from traditional approaches that focus on developing individuals' cognitive skills. Given the complexity of interpersonal skills characteristics, combining two or more methods that have different strengths and weaknesses may promise better results. This implies the necessity of an integrated approach for interpersonal skills training.

Behavior Modeling Training

Behavior modeling has been recognized as the most appropriate instructional method for interpersonal skills training since 1970 (Baldwin, 1992; Goldstein & Sorcher, 1974; Robertson, 1990). It is primarily based on Bandura's (1977) social learning theory, which explains that people can learn by observing others. Noe (2002) defined behavior modeling as "A training method in which trainees are presented with a model who demonstrates key behaviors to replicate and provides them with the opportunity to practice those key behaviors" (p. 452). The primary advantage of behavior modeling is that it enables individuals to learn without a trial and error process. Therefore, behavior modeling serves as an efficient way to teach interpersonal skills by reducing the costs and time involved in a trial and error process. In addition to the efficiency of behavior modeling, many research studies report its effectiveness (Latham & Sarri, 1979; Nowack, 1986). From the learning theory perspective, Simon and Werner (1996) explained that behavior modeling is somewhere between direct instruction and exploration in the continuum of instructional methods because it provides learners with both cognitive instruction and the opportunities to explore learning content with practices. That is, behavior modeling enhances learning effectiveness by combining bipolar approaches.

Decker and Nathan (1985) provided five components of behavior modeling: (a) modeling, (b) retention process, (c) behavioral rehearsal, (d) feedback, and (e) transfer of training (Figure 1). The graphic representation of the behavior modeling training process is as follows:

Note. From Behavior modeling training: Principles and applications (p. 5.), by P.J. Decker & B.R. Nathan, 1985. New York: Praeger Publishers.

[FIGURE 1 OMITTED]

First, modeling is the most distinct feature of the behavior modeling approach. The objective of modeling is to facilitate the learners' attention towards the key behaviors, which is a set of behaviors intended for the students to learn. Decker and Nathan (1985) indicated that when a model's behavior is "distinctive, meaningful to the observer, not too complex, and observable," the effectiveness of modeling is increased (p.43). Second, the retention process facilitates learners' memorization of what they learn from the models. As Bandura (1977) mentioned, if learners do not remember what they saw during the modeling, learning will not occur. To facilitate the retention process, Decker and Nathan (1985) recommended using learning points, which are verbal instructions in words or phrases, to summarize or generalize key behaviors. Third, behavioral rehearsal provides the learners with opportunities to practice the modeled behaviors. Acquiring cognitive concepts of skills does not guarantee actual behavior reproduction, so it is necessary to provide learners with practice opportunities to enhance the likelihood of obtaining desirable performances. Dick and Carey (1996) indicated that practice and feedback are key to the success of psychomotor skills acquisition. Finally, the learners are asked to apply the new knowledge gleaned from the models to their real life. The significance of transfer of training cannot be overemphasized because it is expected to eventually lead to positive performance results, which is the ultimate goal of providing learners with training.

Online Interpersonal Skills Training

Some researchers are skeptical about the effectiveness of interpersonal skills training delivered by technology because of the limited interaction and feedback in the online environment (Holsbrink-Engels, 1998; Cohen & Rustad, 2000). Meanwhile, others are enthusiastic about conveying interpersonal skills training through technology because of the cost-effectiveness it achieves by reducing training costs and its delivery time (DeBloois, 1988; Cronin & Cronin, 1992). The former are more concerned about learning effectiveness of online interpersonal skills training whereas the latter have high expectations of online learning for the cost-effectiveness reasons.

In terms of cost-effectiveness of interpersonal skills training programs, Campbell, Lison, Brosook, Hoover, & Arnold (1995) compared training programs conveyed by way of computer/video technology and instructor-led programs and found that there is no significant difference in learning gains between the two programs. As a result, they claimed that computer/video technology can reduce costs and instructors' working hours to half that which would be spent on instructor-led programs. Despite Campbell and his colleague's study, the effects of interpersonal skills training delivered by technology are still controversial because of insufficient empirical evidence. However, in spite of the technical and pedagogical limitations of online environments for interpersonal skills training, given the proliferation of online learning and the magnitude of interpersonal skills in daily life, it is imperative to explore the most effective method to deliver interpersonal skills training online.

Based on the proven effectiveness of behavior modeling for interpersonal skills training in classrooms, this study applied it to online environments. For this, some changes should be made because of the discrepancies between online and classroom learning environments. One change is model presentation. Given that models convey learning content, key behaviors, it is necessary to design instructionally effective models for learning outcomes (Konczak & Dossett, 1986). The model presentation formats can be live, video, audio, written scripts, or a combination of two or more of these methods (Decker & Nathan, 1985; Bandura, 1986).

For enhancing the effectiveness of modeling, Decker and Nathan (1985) indicated that models' behaviors should be displayed in enough detail to convey the learning content. By contrast, from the cognitive load theory, some researchers insist that it is necessary to provide learners with the appropriate amount of information because of a learners' limited working memory (Mayer, 2001). Meanwhile, Bandura (1986) stated that the effectiveness of modeling is determined by the learners' cognitive development in interpreting key behaviors and the models' characteristics, such as its complicatedness. He also added that if the amount of information conveyed is identical, the presentation formats do not influence learning outcomes because the role of model presentations is to provide learners with information. However, when the results of information processing are defined broadly by processed amount, speed, accuracy, and cost-effectiveness, the model presentation formats may influence the learning outcomes in certain ways.

MULTIMEDIA LEARNING

Since models in behavior modeling are provided to learners through media, it is worthwhile to discuss the influence of media on learning outcomes. The purpose of using media for learning is to facilitate communication between instructors and learners by gaining students' attention, conveying clear messages, and providing students with realistic/concrete learning experiences (Heinich, Molenda, & Russell, 1993; Wileman, 1983). For a better understanding of the influence of media on learning, it is necessary to focus on factors other than the media delivery system, such as the learning experiences delivered by instructional media, learners' cognitive processes, and technological and cost issues.

Mediated Learning Experiences

Some media, such as films or virtual reality, provide learners with very realistic learning experiences, and others, such as line drawing and audiocassettes, do not. Dale (1969) provided a visual analogy, the Cone of Experience, to categorize learning experiences along the concrete-abstract continuum. The model consists of eleven levels of learning experiences; direct purposeful experiences reside at the bottom of the model and the abstract verbal symbols are located at the top of the model.

The Cone of Experience has stimulated controversies over how effective realistic learning experiences are for learning outcomes. Some researchers support the so-called realism theory that more realistic learning experiences are more desirable for facilitating learning (Dale, 1946; Finn, 1953). For example, better learning outcomes can be expected when learners are provided with contrived experiences, such as role-play--a more realistic learning experience, than with only verbal information, such as readings or lectures--the most abstract learning experience. Similarly, Shih and Alessi (1996) presented the concept of fidelity as "the extent to which instruction accurately reproduces the real context of performance" (p.205). In direct opposition to these views however, others claim that the most realistic learning experiences are not necessarily the most effective for learning. Dwyer (1972) pointed out that excessively realistic features in visual materials may inhibit learning because the learners' limited cognitive capacities are overloaded by the large amount of information to process. Shih and Alessi's research findings summary also showed no positive linear relationship between the level of fidelity and the students' learning gains.

Limited Cognitive Load

To make sure learning occurs, learners should actively interpret the mediated information conveyed by instructional materials. According to the information processing theory, information is processed through the three stages: (a) sensory memory, (b) working memory (short-term memory), and (c) long-term memory (Driscoll, 2000). The sensory memory stage receives information from the five separate senses as input. It processes the information momentarily and transfers it to the working memory. In the working memory stage, the transformed information is manipulated to make it meaningful for long-term storage. This stage is important in learning because it determines whether or not the information can be stored in the long-term memory, which is considered permanent memory.

During the working memory stage, learners consciously devote efforts to make sense of learning materials. However, it is important to note the fact that working memory is limited in its capacity (Foshay, Silber, & Stelnicki, 2003). Due to this limited capacity, all the information transferred from the sensory memory cannot be processed. Some of it is inevitably lost. If an excessive amount of information is presented to students concurrently, the learners' cognitive load, which indicates the total amount of mental activity required to process information in the working memory, will overflow resulting in inefficient information processing. Sweller and Chandler (1994) made distinctions between intrinsic and artificial sources of cognitive load.

The intrinsic source of cognitive load has to do with the level of difficulty and the interactivity among elements in materials. An example of high intrinsic cognitive load is when the learning material has a high level of learning content difficulty and requires learners to acquire several elements concurrently. By contrast, the artificial source of cognitive load is associated with instructional design--how to arrange and present instructional elements. Regardless of the intrinsic cognitive load in the learning materials, poorly designed instructional materials, such as redundant information or ambiguous learning activities, will result in the increase in the learners' cognitive load. When learners have to interpret unnecessary instructional elements embedded in the materials, learning effectiveness is reduced. Given a learners' limited working memory, it is necessary to provide learners with the appropriate amount of information with well-designed instructional materials to reduce their intrinsic and artificial sources of cognitive load (Brunken, Plass, & Leutner, 2003).

Dual-Coding Theory 

Among the human beings' five senses, the eyes and ears are the primary sense organs to receive information (e.g., visual and auditory information) for cognitive learning. For a better understanding of how individuals receive information from the external world, Paivio (1991) makes a distinction between symbolic modality and sensory modality. He explains symbolic systems related to sensory modalities for learning as follows: "Cognition is served by two modality-specific symbolic systems that are experientially derived and differentially specialized for representing and processing information concerning nonverbal objects, events, and language" (p.16).

Symbolic modality consists of verbal (e.g., written text or spoken languages) and imaginal (e.g., pictures, full-motion pictures, and objects). Sensory modality can be categorized into visual (through eyes) and auditory (through ears). For example, both printed words and pictures are presented in the visual mode, but the former are verbal information and the latter are imaginal information. Also, reading a book and listening a lecture are both learning activities to receive verbal information, but their sensory modalities are different: visual and auditory, respectively.

To illustrate the procedure by which information is received and processed, Paivo (1991) proposed dual-coding theory, an explanation of how individuals encode information separately in the verbal and the imaginal channels (Moore, Burton, & Myers, 1996). However, Mayer (2001) explained that the independently received verbal and imaginal information can each be transformed to the other format to allow it to be processed effectively when is stored, organized, and retrieved.

Technological/Cost Issues

The advance in digital technology enables instructors to adopt multimedia for online learning. However, when multimedia are applied to the online learning environment for presenting models, learners may encounter logistical problems, such as insufficient bandwidth, the internet connection, and the required hard-disk space. Other problems include student patience while downloading files, distorted video screens, and program malfunctions. Also, Fisher (2000) warned of the limitation of technical capacity to provide multimedia to the current online learning environments.

Another factor to be considered when using multimedia for presenting models on the Web is that of cost. The costs of preparing and delivering different sorts of presentations can vary dramatically. Given that costs are an ever-present and very sensitive issue in decision-making about creating and maintaining instructional programs, the cost-effectiveness for instructional units should be taken into consideration when designing online learning programs.

Despite the high expectation of multimedia for learning outcomes, it is also important to note that the glitzy features of multimedia do not always promote learning efficiency. It means that the multimedia features available (possible) are not necessarily effective for learning outcomes (Moore et al., 1996; Mayer, 2001). Rather, they may distract learners' attention or lower their retention. The selection of multimedia features for instructional units should be made based on learning effectiveness, not their technical availability or their novelty.

Until recently, video has been the most frequently used format to present models in behavior modeling approaches in the classroom because it enables trainers to provide learners with "semi-live" models at a lower cost. However, compared with other models, such as illustrations, pictures, or verbal information, video still costs more to develop in both classroom and online learning environments. From a cost-effectiveness perspective, it is desirable to use lower cost model formats if they are able to produce the equivalent amount of positive learning outcomes. Konczak and Dossett (1986) examined the effectiveness of different presentation formats: audio only, text only, audio plus text, and video for adult learners in a supervisory skills training program. This study found no significant difference among the four model presentation formats in terms of learning recall. Similarly, Hyte (1985) compared the four different presentation formats: videotape, audiotape, text-script, and slide plus audiotape in an assertive communication training program. She found that there was no significant difference among the four modeling formats. However, the results of the two studies cannot be directly applied to online environments because of different experiment settings (classroom vs. online learning environments).

In summary, in the light of technical/costs limitations for providing realistic models in the online learning environment and some previous research findings that realistic features in models do not correspond to the learning outcomes they yield, it is necessary to investigate the efficiency and effectiveness of different model presentation formats for behavior modeling in online learning of interpersonal skills.

METHODS

The research question of this study is as follows:

In implementing the behavior modeling approach to interpersonal skills training in the online context, do different presentation formats have differential effects on learners' outcomes in terms of:

1. learner reactions (primarily affective)?

2. short-term (cognitive) retention of learning content?

3. (delayed) behavioral reproduction?

To examine the effects of the presentation format in behavior modeling approaches to online interpersonal skills training, four different presentation formats were compared: video, still pictures and audio, audio alone, and text script alone. These were chosen to represent a broad range of levels of realism (Heinich et al, 1993);

1. video: auditory plus visual information in full motion,

2. picture plus audio: auditory information plus still, realistic visuals,

3. audio only: auditory information without visuals, and

4. text script only: verbal information in written format without visuals.

Participants

A total of 86 college students who enrolled in a job search strategy class at a major Midwestern university participated in the study. Initially, the experiment started with 97 participants. However, as the experiment proceeded, it turned out that 11 participants' information was missing in measurements. Demographic data demonstrated that 59% of the participants were juniors, 29% were seniors, and 12% were of other classifications. Female students (58%) outnumbered males (42%).

Given that the training program used for the research was about job interview skills, questions about the participants' entry levels of interview skills were asked. It turned out that the participants had moderate amounts of previous interview experience from part-time jobs or internships (Mean=4.3 times). Eighty four (84%) percent of participants answered that they were "somewhat prepared" or "not well prepared" for interviewing and more than two third of the participants (78%) reported that they had little or some prior knowledge about interviewing. In addition, histograms of the participants' entry interview skills level scores were normally distributed.

Training Program

The training program used for the research was a web-based job interview preparation program. The learning content of the program consisted of two parts: behavior-based interview skills and verbal interview skills. Behavior-based interview skills included (a) making a good impression on an interviewer, (b) giving a firm handshake, and (c) maintaining good posture during the interview. The verbal interview skills consisted of the self-introduction and answering three frequently asked interview questions, such as "what is your strength?" "what is your weakness?" and "why should I hire you?" Three behavior-based and four verbal interview skills constituted the seven units of the program.

The training program was developed based on Bandura's (1977) observational learning process and Decker and Nathan's (1985) five components of behavior modeling. However, the program structure for online learning was modified to be suitable for online learning environments. The structure of the program is as follows:

1. Provide an overview of the topic, given verbally,

2. Display behavior models (both a positive and negative model),

3. Provide learning points--principles, procedures, and advice given verbally,

4. Provide self-practice opportunities, and

5. Provide self-assessment criteria

A brief explanation of each component is as follows: the overview of the topic explains why the topic is important and provides a concise summary of the content as an advanced organizer. Displaying behavior models included both positive and negative models based on Baldwin's (1992) conclusion that the combination of positive and negative model displays was effective in promoting learners' generalization of what they learn. As a result, the learners had a total of 14 model displays, a set of positive and negative model displays for each of seven topics. The topics of model displays include making a good impression on an interviewer, offering a firm handshake, maintaining good posture during the interview, self-introduction, explaining both one's greatest weakness and strength, and justifying why he/she is an excellent candidate for the position. Learning points followed after model displays to facilitate learners' retention. And then skills practice opportunities are provided. Decker and Nathan (1985) especially asserted that it is difficult to expect skills acquisition without a sufficient number of practices in interpersonal skills training. However, since the program was provided online, the learners had self-practice opportunities without interacting with other students. Finally, the learners were given self-assessment criteria to evaluate their performance as substitutes for feedback from other students or instructors.

The contents and scenario of the model displays were developed by the two instructors who taught the job search strategy course. To investigate the effects of the presentation format, four versions of the training program, which had the same content and structure, but different presentation formats, were prepared. The presentation formats were video, still pictures and audio, audio alone, and text script alone. For development process reasons, the models in the video format were the first group to be produced and then these video clips were converted to still pictures plus audio, audio only, and the text script format in that order. This procedure made it possible to produce four programs with different presentation formats, but identical content. The program was developed to take approximately one hour to complete and it was delivered to the participants through an online learning system.

Measurement Instruments

1. Reaction. The reaction measurement asked the participants to rate how much they enjoyed the online job interview preparation program using ten items on a 5-point Likert scale. The reaction measure was administered online and consisted of two parts: (a) general reactions regarding enjoyment, usefulness, and how much they felt they learned through the program (possible scores ranged from 3-15) and (b) reactions to individual seven learning topics specifically (Alpha coefficient: .72, possible scores ranged from 7-35).

2. Short-term retention. The short-term retention test assessed how much participants recalled the concepts and principles regarding interviewing presented in the treatment program. It was administered online and measured using 14 multiple-choice items online. The alpha coefficient of this measure was .76. and possible scores ranged from 0 to 14.

3. Behavioral reproduction. An observation checklist, which consists of the criteria for assessment of the participants' interview performance, was developed to facilitate the raters' observation. The observation was conducted for verbal-interview and behavior-based skills separately. The interrater agreement was calculated based on the assessment results of 200 observed behaviors of first seven participants. The agreement between the two observers was initially 75%, so the disagreement between two observers was resolved afterwards.

[FIGURE 2 OMITTED]

To ensure participants' access to the program without technical problems, several versions of computer operating systems, the required hard-disk space, web-browsers, and internet bandwidth (or internet connection) were tested in both on-campus and off-campus computer environments in different time slots. Based on the results, specific minimum technology requirements were described in the instruction.

Experimental Procedures

This study was a posttest only randomized experiment. Participants were randomly assigned one of the model presentation groups (ie., video, still pictures plus audio, audio only, and text scripts only) and received training programs as treatment according to the group they belonged to. They were asked to complete the programs posted on an online learning system within two weeks. To prevent contamination of the treatment, a program for each group was posted on each group space, resulting in not allowing other groups to access it.

Immediately after the completion of the training program, the participants were asked to fill out both the online reaction questionnaire and the short-term cognitive retention test. Approximately two weeks later, the participants performed their interviews at group mock interview sessions for around 45 minutes for the measuring the delayed behavior reproduction of interview skills. In order to prevent copying others' answers at group mock interview sessions, the questions were reworded or rephrased and then asked to the participants in random order. Two observers rated their interview skills with the prestructured observation checklist.

An analysis of variance (ANOVA) using SPSS 11.5 was performed to examine whether there were significant differences between the four groups in terms of reaction, short-term retention of learning content, and behavioral reproduction. An alpha level of .05 was used for all statistical tests and mean substitution was employed as an alternative to deal with missing information.

RESULTS

Effects of Different Presentation Formats on Learners' Reactions

Learners' reactions to the program were divided into two categories for analysis: general reactions to the program and reactions to each of the seven individual learning topics. The overall participants' general reaction to the program was positive and the result of the ANOVA test indicated that there was no statistically significant difference among the four groups, F(3, 82)= .43, ns (see Table 1). The reactions to individual learning topics, which are perceived educational gains for each learning topic through the program, were also positive. No statistically significant differences between groups were found in reactions to individual learning topics, F(3, 82) = .76, ns.

Participants were also encouraged to answer to open-ended questions about the usefulness of the program. It showed that they liked the program because it gave them clear ideas of effective interview strategies to help them win the job they want. In addition, the participants thought that the program was useful because it provided them with opportunities to reflect on what they did wrong in their previous interviews by comparing the behaviors of desirable and undesirable interviewees in the model presentations.

The comparison of comments from each group showed that many participants in the video group made comments on video examples, such as "video examples were well done, and appreciated," "easy to understand," "It was helpful to see some real life examples from real people to demonstrate how to answer tough questions in those situations," "It felt like I was there viewing behind a mirror," "Until I actually watched the video clips I didn't realize that I could be displaying some of the negative action," "It was helpful to see visual images of what an interviewee should look like when answering and what kind of answers he/she should say" By contrast, comments on the model presentation formats were rarely found in the other three groups' remarks.

Effects of Different Presentation Formats on Learners' Cognitive Retention

The ANOVA result demonstrated that there was no significant difference between groups in short-term cognitive retention of the learning content, F(3, 82) = .75, ns (Table 2).

Effects of Different Presentation Formats on Learners' Behavioral Reproduction

The participants' behavioral reproductions of interview skills were measured in two categories: verbal interview and behavior-based interview skills. Since no significant correlation between verbal and behavior-based interview skills was found (r= .12), these two measures were separately analyzed. The ANOVA result indicates that there was no statistically significant difference between the four groups in behavioral reproduction of verbal interview skills, F (3, 82) = .40, ns. Also, no significant difference between groups in behavior-based interview skills reproduction was found, F (3, 82) = .54, ns (Table 3).

DISCUSSION AND IMPLICATIONS

Based on the proven effectiveness of the behavior modeling approach for interpersonal skills training, this study made an attempt to apply behavior modeling to the design of interpersonal skills training for online learning environments. To identify the most effective method to present models in behavior modeling-based programs teaching interpersonal skills, four different model presentation formats were compared to assess what differences there might be in learning outcomes. The four model presentation formats were video, pictures plus audio, audio only, and text-script only.

Effects of Model Presentation Formats on Learning Outcomes

No significant differences between groups who had different model presentation formats were found for learners' reaction, short-term cognitive retention of learning content, or behavioral reproduction. These findings imply that learners do not favor a particular model presentation format. However, the findings should be interpreted with caution. Since the participants were exposed to only one of the four model presentations, the findings do not indicate their equal preferences between each model presentation.

This study confirmed the findings of Konczak and Dossett's (1986) and Hyte's (1985) research, which also found there was no significant differences between groups who had different model presentations. Another explanation for this might be the influence of the participants' entry level of interviewing. Upon analyzing the demographic information, it was found that most participants had some knowledge about interviewing from their experiences seeking part-time jobs or internships before they participated in this study. Random assignment was used to ensure that participants' level of interview skills were distributed across the treatment groups. However, their entry level of interviewing might reduce the range of learning effects from the program.

At the same time, it is important to note that the four different model presentation formats did not produce the same results given the descriptive difference and the different pattern of the scores. In other words, the findings of this study, no statistically significant differences between groups, are valid within the statistical power of the study set. Since the study quantitatively examined the collective learning outcomes without qualitative consideration, it is dangerous to conclude that the four different model presentations are equally effective in producing the desired learning outcomes. The results of this study should be interpreted with caution.

The implication of the findings is that it is reasonable to use cost-effective model presentation formats given that there was no statistically significant difference between groups. Clark (1994) believed that because there is no particular benefit to be expected from using a certain medium for learning, it is reasonable to use a cost-effective medium (i.e., model presentation formats). The cost-effective model presentation format can be defined as the presentation formats that best conveys the learning content with reduced development/delivery costs or increased delivery speed of instruction. Despite the findings of this study, there are three possible moderating variables of the relationship between model presentation formats and learning outcomes that instructional designers should keep in mind. They are learning contents, situations, and prior knowledge.

First, learning contents moderate the influence of model presentation formats on learning outcomes. If learning content is obvious to learners or it can be easily understood with abstract model presentation formats, such as text-scripts or audio models, there is no need to use a realistic model presentation, such as video. However, some knowledge, such as tacit knowledge need a more detailed model presentation format in order to facilitate learners' understanding of the learning content. It is because tacit knowledge by its nature cannot be straightforwardly conveyed to learners with abstract verbal information.

As previously mentioned, if learners have clear mental images of situations described in modeling or they are already familiar with the situations, abstract model presentation formats without details may be enough for teaching interpersonal skills. However, if a learning situation is novel, or complex, or never experienced before, learners need more realistic model presentations to have clear ideas about situations. In this study, if learners had a number of interview experiences, they could imagine interview situations described in the models even with an abstract model presentation format, such as text-scripts. However, if the learners had little experiences or did not have any previous interview experiences, they might have difficulty understanding the situations described in abstract models.

Similarly, if learners have no prior knowledge about the learning content to be taught, they need more detailed model presentation formats (Park & Hopkins, 1993; Reiber, 1990). However, if the learners have sufficient prior knowledge or were already exposed to the learning content, they do not need to have a detailed model presentation in order to figure out what the model teaches. In spite of the findings of this study, no significant differences between groups, these three variables should be considered for choosing an appropriate model presentation format for achieving better learning outcomes.

Limitations of the Study

This study has several limitations. First, since the web-based training program was used as a part of a class, the difficulty of the short-term cognitive retention measure had to be adjusted to the class expectation. This might reduce the amount of variation on the cognitive retention measure. Second, this study lacks a pilot study. According to Isaac and Michael (1997), a pilot test is necessary for ensuring the appropriateness of hypotheses and measurement plans. This study did not conduct a pilot study for logistical reasons, such as time constraints. Finally, a randomized pretest-posttest design is ideal for measuring accurate treatment effects. However, this study did not administer the pretest because of the possible interaction of pretest and treatment threat and time conflicts between the data collection schedule and the schedule of the class in which the study was conducted.

For Future Research

Potential follow-up studies should be conducted to replicate this study for other interpersonal skills training subjects such as sales, counseling, coaching, and leadership. Since each interpersonal skill has unique aspects, it is necessary to investigate different subject areas of interpersonal skills for successful results. At the same time, it will be necessary to identify the common denominators of interpersonal skills training to establish a solid guideline for model presentations in online interpersonal skills training.

The effects of model presentation formats on learning should be investigated in a variety of perspectives. Despite the finding of this study "no significant difference between groups," it is necessary to investigate the effects of model presentations formats on various learning aspects, not just quantitative learning outcomes.

To accurately measure the effect of model presentation formats on learning outcomes, it is necessary to use a more challenging subject as the learning content. In this study, since the participants had a certain amount of knowledge of interview skills prior to the experiment, it might influence on the participants' attitudes toward the treatment and the amount of learning outcomes. In addition, it is necessary to empirically examine the effects of three possible moderating variables of the relationship between model presentation formats and learning outcomes: learning contents, situations, and prior knowledge.

Table 1

The Results of the Reaction Measurement Means and Standard Deviations of
Reactions

                          Model Presentation Formats

                 Video    Pictures plus   Audio only   Text-script
                              audio                       only

                M     SD     M      SD      M     SD     M     SD

General       11.44  2.06  10.81   1.86   11.33  2.14  11.42  2.26
reactions

Reaction to   26.52  4.41  24.76   4.06   26.22  4.66  25.92  3.26
individual
learning
topics

Note: Video group (N= 23), Pictures plus audio group (N= 21), Audio only
group (N= 18), Text script only group (N= 24).

Results of Analysis of Variance on Reactions

                      Source of variance   SS      df   MS      F

General reactions     Between groups         5.65    3   1.88   .43
                      Within groups        358.72   82   4.38
                      Total                364.37   85

Reaction to seven     Between groups        37.83    3   12.61   .76
individual learning   Within groups       1370.49   82   16.71
topics                Total               1408.33   85

Table 2

The Results of Cognitive Retention of Learning Content Means and
Standard Deviations for Short-term Cognitive Retention of Learning
Content

                                   Model Presentation Formats

                         Video      Pictures    Audio only   Text-script
                                   plus audio                   only

                        M     SD     M     SD     M     SD     M    SD

Cognitive retention   12.70  1.49  12.76  1.51  12.89  1.08  13.21  .88
of learning content

Note: Video group (N= 23), Pictures plus audio group (N= 21), Audio only
group (N=18), Text script only group (N= 24).

Results of Analysis of Variance on Short-term Cognitive Retention

             Source of variance  SS      df  MS    F

Short-term   Between groups        3.64   3  1.21  .75
cognitive    Within groups       132.42  82  1.62
retention    Total               136.06  85

Table 3

The Results of Behavioral Reproduction Measurements Means and Standard
Deviations for Behavioral Reproduction

                                Model Presentation Formats

                      Video       Pictures      Audio only  Text-script
                                 plus audio                     only

                     M     SD     M       SD      M     SD     M    SD

Verbal             12.69  1.81  12.57   1.91   12.67  1.61  12.75  1.68
interview
skills

Behavior           13.63  1.09  13.28   1.48   13.21  1.33  13.13  1.73
interview
skills
(non-verbal)

Note: Video group (N= 23), Pictures plus audio group (N= 21), Audio only
group (N= 18), Text script only group (N= 24).

Results of Analysis of Variance on Behavioral Reproduction

                     Source of variance  SS      df  MS    F

Verbal interview     Between groups        3.60   3  1.2   .40
skills               Within groups       254.17  82  3.10
                     Total               254.53  85

Behavior interview   Between groups        3.32   3  1.11  .54
skills (non-verbal)  Within groups       168.65  82  2.06
                     Total               171.96  85

Acknowledgement

I thank Michael Molenda, Barbara Bichelmeyer, Curtis Bonk and Timothy Baldwin at Indiana University Bloomington for their critical comments on a previous version of this article.

References

Argyle, M. (1981). Social skills and work. New York: Methuen & Co.

Bailey, C.T., & Butcher, D.J. (1983). Interpersonal skills training I: The nature of skill acquisition and its implications for training design and management. Management Education and Development, 14(1), 48-54.

Baldwin, T.T. (1992). Effects of alternative modeling strategies on outcomes of interpersonal skills training. Journal of Applied Psychology, 77(2), 147-154.

Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice Hall.

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall.

Brunken, R., Plass, J.L., & Leutner, D. (2003). Direct measurement of cognitive load in multimedia learning. Educational Psychologist, 38(1), 53-61.

Campbell, J.O., Lison, C.A., Brosook, T.K., Hoover, J.A., & Arnold, P.H. (1995).

Using computer and video technologies to develop interpersonal skills. Computers in Human Behavior, 11(2), 223-239.

Clark, R.E. (1994). Media will never influence learning. Educational Technology Research & Development 42(2), 21-29.

Cohen, S.L., & Rustad, J.M. (2000). High-tech, high-touch, high time: A realistic look at technology's potential for teaching soft skills. In K. Mantyla (Ed.), The 2000/2001 ASTD Distance Learning Yearbook (pp.186-193). New York: McGraw-Hill.

Cronin, M., & Cronin, K. (1992). Recent empirical studies of the pedagogical effects of interactive video instruction in "soft skill" areas. Journal of Computing in Higher Education, 3(2), 53-85.

Dale, E. (1946). Audiovisual methods in teaching. New York: Dryden Press.

Dale, E. (1069). Audiovisual methods in teaching (3rd ed.). New York: Dryde-Press.

DeBloois, M. (1988). Use and effectiveness of videodisc training: A status report. Falls Church, VA: Future Systems, Inc.

Decker, P.J., & Nathan, B.R. (1985). Behavior modeling training: Principles and applications. New York: Praeger.

Dick, W., & Carey, L. (1996). The systematic design of instruction (4th ed.). New York: HarperCollins.

Driscoll, M. P. (2000). Psychology of learning for instruction. Boston: Allyn & Bacon.

Dwyer, F.M. (1972). A guide for improving visualized instruction. State College, PA: Learning Services.

Finn, J. D. (1953). Professionalizing the audio-visual field. AV Communication Review, 1, 6-17.

Fisher, S.G. (2000). Web-based training: One size does not fit all. In K. Mantyla (Ed.).

The 2000/2001 ASTD distance learning year book (pp. 63-75). New York: McGraw-Hill.

Foshay, W.R., Silber, K.H., & Stelnicki, M.B. (2003). Writing training materials that work. San Francisco, CA: Jossey-Bass/Pfeiffer.

Galvin, T. (2000). Industry report 2002: Training magazine's 21st annual comprehensive analysis of employer-sponsored training in the United States. Training, 39(10), 24- 73.

Gilmore, E.R., & Fritsch, P.J. (2001). Online versus instructor-led courses. Performance Improvement, 40(10), 30-38.

Goldstein, A.P., & Sorcher, M. (1974). Changing supervisor behavior. New York: Pergamon.

Hargie, O. (Ed.). (1986). A handbook of communication skills. New York: New York University Press.

Heinich, R., Molenda, M., & Russell, J.D. (1993). Instructional media and the new technologies for instruction (4th ed). New York: Macmillan.

Holsbrink-Engels, G.A. (1998). Computer-based role playing for interpersonal skills training.

Unpublished doctoral dissertation, University of Twente, Enschede, The Netherlands.

Hyte, L.M. (1985). A comparison of modeling modes in the behavior modeling process. Unpublished master's thesis, North Carolina State University, Raleigh.

Isaac, S., & Michael, W.B. (1997). Handbook in research and evaluation (3rd,), San Diego, CA: Educational and Industrial Testing Services.

Konczak, L.J., & Dossett, D.L. (1986). Media effects in behavior modeling training. Presented at the Annual Meeting of the American Psychological Association. (ERIC Document Reproduction Service No. ED287464)

Latham, G.P., & Sarri, L.M. (1979). Application of social-learning theory to training supervisors through behavioral modeling. Journal of Applied Psychology, 64(3), 239-246.

Marlowe, H.A. (1986). Social intelligence: Evidence for multidimensionality and construct independence. Journal of Educational Psychology, 78(1), 52-58.

Mayer, R. E. (2001). Multimedia learning. New York: Cambridge University Press.

Moore, D., Burton, J., & Myers, R. (1996) Multiple-channel communication: The theoretical and research foundations of multimedia. In D.H. Jonassen (Ed.), Handbook of research for educational communications and technology (pp.851-878). New York: Macmillan.

Noe, R.A. (2002). Employee training and development. New York: McGraw Hill.

Nowack, K.M. (1986). Pre-post-then evaluation of a behavioral modeling approach to supervisory skills training. Performance & Evaluation, 25, 14-16.

Paivio, A. (1991). Images in mind. Exeter, UK: Harvester Wheatsheaf.

Park, O., & Hopkins, R. (1993). Instructional conditions for using dynamic visual displays: A review. Instructional Science 22, 1-24.

Reiber, L.P. (1990). Animation in computer-based instruction. Educational Technology Research and Development, 38(1), 77-86.

Robertson, I.T. (1990). Behaviour modelling: Its record and potential in training and development. British Journal of Management, 1(2), 117-125.

Shih, Y., & Alessi, S.M. (1996). Effects of text versus voice on learning in multimedia courseware. Journal of Educational Multimedia and Hypermedia, 5(2), 203-218.

Simon, S.J., & Werner, J.M. (1996). Computer training through behavior modeling, self paced, and instructional approach: A field experiment. Journal of Applied Psychology, 81(6), 648-658.

Sweller, J., & Chandler, P. (1994). Why is some material difficult to learn? Cognition and Instruction, 12, 185-233.

Wileman, R.E. (1993). Visual communicating. Englewood Cliffs, NJ: Educational Technology Publications.

MIN YOUNG DOO

James Madison University 

USA

[e-mail]