Abstract
One of the limitations of keyword mnemonics concerns its efficacy in delayed recall. The objective of this study was to analyse the impact of keyword mnemonics on delayed recall (at a one-week interval). A sample of 794 first and second-year Compulsory Secondary Education students of both genders were presented a list of 14 Basque words, the meanings of these words were to be learnt using keyword mnemonics (keyword selected by classmates, researchers, or by the participants themselves), or the rote learning method. Half of the sample were allowed immediate and delayed recall, and the remaining 50% of participants underwent delayed recall. Activation of immediate recall influenced delayed recall. The rote learning method was more effective than the keyword method when the keyword was supplied by the researcher, or when participants generated their own keywords themselves, but not as effective as the keyword method when the keyword was generated by the classmates of the participants in the study.
Keyword mnemonics is considered to be one of the most effective foreign language learning techniques (for a review, see González et al., 2003). The technique was proposed by Atkinson (1975) for native English speakers learning Russian as a foreign language. This technique involves two steps: In the first step a concrete keyword is generated in the mother tongue that sounds similar to the word in the foreign language to be learnt. The second step consists of creating imagery linking the meaning of the keyword with the meaning of the word in the foreign language (Atkinson, 1975; Atkinson & Raugh, 1975; Campos et al., 2002).
The most common keyword mnemonics procedures are to allow students to generate their own keywords as they learn the foreign language, or for researchers to supply the keywords to the participants in a study (González et al., 2003). Mnemonics is more efficacious in individual tasks (Levin et al., 1979), and in children (Levin et al., 1982), with less efficacy in adolescents and adults (Campos et al., 2003); nevertheless, better results are obtained with mnemonics than with other learning techniques (Campos et al., 2013, 2014).
According to whether the participants generate their own keywords or the researcher supplies the keywords in time-controlled learning entails both advantages and drawbacks. If the researcher provides the keyword to participants at the moment of learning, the advantage is that participants only have to concentrate on learning, but the drawback is that keywords may not be meaningful for participants, and hinder learning. On the other hand, if students generate their own keywords while learning, more time and effort is spent on generating keywords than on learning itself (Campos, González, et al., 2004). To overcome these shortcomings, a new keyword selection method has been proposed (Campos et al., 2002; Campos, Amor, et al., 2004a, 2004b; Campos, González, et al. 2004), where keywords were generated by the classmates of the participants in the study, who were of similar sociodemographic characteristics and academic status. The advantage of this method is that classmate-generated keywords are very akin to participant-generated keywords. A further learning advantage is that participants can spend their time studying, without worrying about having to generate keywords. Thus, classmate-generated keywords obtained the best results in comparison to other keyword selection techniques (Campos & Ameijide, 2014; Campos et al., 2002; Campos, Amor, et al., 2004a, 2004b, Campos, González, et al., 2004; Campos et al., 2010).
A keyword mnemonic is an efficacious mnemonic method for learning second language vocabulary, with attested efficacy particularly in the short-term, and in the learning of a variety of languages such as English, Spanish, French, Italian, Russian, and Latin, among others (Atkinson & Raugh, 1975; Baleghizadeh & Ashoori, 2010; Campos, González, et al., 2004; Davoudi & Yousefi, 2016; Hall et al., 1981; Lawson & Hogben, 1998; Merry, 1980; Shapiro & Waters, 2005).
In comparison, the results of studies on the efficacy of keyword mnemonics in the long-term are inconsistent. Several studies have found keyword mnemonics was more efficacious than other techniques for the learning of foreign language vocabulary in the long-term (Campos et al., 2010; Rodríguez et al., 2014; Sagarra & Alba, 2006; Siriganjanavong, 2013; Soleimani et al., 2012; Tavakoli & Gerami, 2013; Zhang & Schumm, 2000). In contrast, other studies found no long-term advantages in keyword mnemonics over other learning techniques (Campos, Amor, et al., 2004a; Carney & Levin, 1998; Pressley et al., 1988), or even reported rote-learning outperformed keyword mnemonics in the long-term (Campos et al., 2003; Thomas & Wang, 1996; Wang et al., 1992, 1993).
In most of the studies analysing the efficacy of keyword mnemonics in the long-term, participants first performed intentional immediate recall prior to incidental delayed recall, normally after a one-week interval (Campos et al., 2003, 2010; Carney & Levin, 2008; Soleimani et al., 2012). Alternatively, in other studies half of the participants remembered immediately and in the long-term, and the other half remembered only in the long-term in order to analyse the effect of immediate recall on delayed recall. In all of the studies, immediate recall influenced delayed recall (Carney & Levin, 1998, 2008; Wang & Thomas, 1995; Wang et al., 1992, 1993). There were no significant differences between rote learning and keyword mnemonics, and rote learning even outperformed keyword mnemonics when there was delayed recall without immediate recall, (Carney & Levin, 1998, 2008; Wang & Thomas, 1995; Wang et al., 1992, 1993).
In the only study reporting significant long-term differences without short-term recall, where participants recalled half of the words immediately, and all of the words in the long-term to expose only half of the words to immediate recall but not the other half, significantly more words were recalled both in the short and long-term with keyword mnemonics than through rote learning (Carney & Levin, 2008).
The results of keyword mnemonics are normally interpreted (Ashcraft, 1998; Campos et al., 2010) from the perspective of Tulving’s (1972, 1989, 1993) theory of Episodic Memory that asserts immediate recall is good, but with a rapid decline over time. Visual encoding is particularly susceptible to interference over time and tends to quickly fade, which would explain the higher delayed recall observed in rote learning than in keyword mnemonics.
The aim of the present study was to assess the impact of immediate recall and keyword mnemonics (researcher supplied, classmate generated, or participant-generated, keywords) and the rote learning method on delayed learning (at one week) of the Spanish meaning of Basque words. The independent variables were immediate recall (with or without immediate recall) and keyword mnemonics (researcher supplied, classmate generated, or participant-generated, keywords) and the rote method. The dependent variable was delayed recall (at one week) of the Spanish meaning of Basque words.
This study is innovative as few studies analyse the effects of immediate recall on delayed recall, and most studies use both types of recall, but fail to include a no immediate recall control group. Moreover, this is the first study analysing the long-term effects (at a one week interval) of keyword mnemonics on the learning of Basque words, the only European non-Indo-European language, which is primarily spoken in the Basque Country and Navarra (northern Spain).
Method
Participants
The sample in this study consisted of 794 first and second-year Compulsory Secondary Education (CSE) students from state CSE schools from the province of A Coruña (northwestern Spain); 416 girls and 378 boys; age range 11 to 16 years, mean age 12.90 years, standard deviation 0.02.
Instruments
In this study a list of 14 Basque words (7 high-imagery and 7 low-imagery words) were used, each with their corresponding keyword, the Spanish translation, and 14 sentences relating the keywords with the Spanish translation of the Basque words.
Procedure
The first step was to select a list of 135 nouns, 79 high-imagery, and 56 low-imagery words (Valle, 1998). A word is considered to be high-imagery if the imagery value is above 5.20, and low-imagery if the imagery value is below 4.21 (Valle, 1998). Next, the words were translated from Spanish into Basque by a native Basque speaker, and all homographs and homophones in both languages were eliminated. The researchers then generated a concrete keyword in Spanish for each Basque word, which resembled the sound of the word in Basque, and began with the same syllable whenever possible. Moreover, words with keywords similar to other words were eliminated from the list.
From the resultant list, a total of 24 high-imagery and 24 low-imagery words were randomly selected. The list of 48 words was included to allow classmates to reject words during the generation of keyword. In addition, 48 interactive sentences were generated linking each keyword with the Spanish meaning of each Basque word.
The 48 words were presented to a group of 48 first and second-year Compulsory Secondary Education (CSE) students from state schools of similar sociodemographic characteristics and academic status to the participants in the study, each student generated 48 keywords and 48 interactive sentences linking the keyword with the Spanish meaning of the Basque words. Students were allowed 50 minutes to complete the task. Then, the most frequently repeated keywords and interactive sentences were selected. Finally, from the list of 48 words and sentences generated by the group of classmates, a final list of 14 words (7 high-imagery and 7 low-imagery) was selected for use in this study according to the criterion of the highest frequency repeated by students.
The classmate-generated keyword group were presented each word on a separate slide that included the following: the Basque word, the keyword, the Spanish translation, and the interactive sentence linking the keyword and the Spanish meaning of the Basque word, with each slide being read aloud twice by the researchers. The same procedure was applied in the researcher-supplied keyword group. The group of participants generating their own keywords while learning were presented the Basque words and the Spanish translation, and were instructed to generate their own keywords themselves. The same procedure was applied in the group using the rote learning method as in the group generating their own keywords, but the group was instructed to learn the meaning of the words through the word repetition method.
The learning of the list of 14 Basque words was undertaken in the students’ usual classrooms during school hours. First, students were given instructions and provided 3 training examples prior to proceeding to the 14 slides on a projector set at an interval of 15-second per slide. Following the word presentation, participants completed a questionnaire on an item evaluating the efficacy of the learning strategy they had employed. Having completed the questionnaire/item, a group of 383 participants were allowed 1 minute and 30 seconds to write the Spanish meaning next to each Basque word on the list. The remaining 411 participants only completed the survey designed to justify learning. Subsequently, after a one-week interval, and without any prior notice, participants were allowed 1 minute and 30 seconds to once again write the Spanish meaning next to each Basque word on the list.
The study complied with the Declaration of Helsinki 2013, and was approved by the Ethics Committee of Santiago de Compostela University. Moreover, informed consent was obtained from schools, parents or legal guardians, and all students freely volunteered to participate in the study, and were assured their data would remain anonymous and confidential.
Data Analysis
In order to analyse the influence of immediate recall and keyword mnemonics (classmate-generated keywords, researcher-supplied keywords, participant-generated keywords), and the rote learning method on delayed learning (at one week) of the Spanish meaning of Basque words, a two-way ANOVA (with and without immediate recall) x 4 (3 keyword word mnemonics, 1 rote) was performed. To examine difference in delayed recall between the learning strategies (classmate-generated keywords, researcher-supplied keywords, participant-generated keywords, and the rote learning method), a Least Significant Difference (LSD) analysis was performed.
Results
The means and standards deviations of each group: with or without immediate recall, and keyword mnemonics (classmate-generated keywords, researcher-supplied keywords, participant-generated keywords), and the rote learning method are shown in Table 1. The results of the ANOVA indicate immediate recall significantly influenced delayed recall, F(1, 785) = 227.54, p <.001, ή2p = .23, power = 1. Students who were allowed to write the Spanish meaning of Basque words immediately after learning recalled more meanings of words at a one-week interval (M = 3.35; SD = 2.69), than participants who were not allowed to write the meaning immediately after learning (M = 1.14; SD = 1.32).
Means and Standard Deviations of Delayed Recall (One Week) of the Spanish Translation of Basque Words Obtained by the Three Keyword Selection Groups and Rote Learning Group, and for Participants With or Without Immediate Recall.
Significant difference was observed in the long-term recall (one week) of the Spanish meaning of Basque words between the three different keyword generating procedures (classmate-generated, researcher-supplied, and participant-generated keywords) and the rote learning group, F(3, 785) = 9.33, p <.001, ή2p = .03, power = 1. To determine significant differences between groups, a Least Significant Difference (LSD) analysis was performed that revealed the classmate-generated keyword group (M = 2.47; SD = 2.87) recalled the meaning of more words than the researcher-supplied keyword group (M = 1.70; SD = 2.09) (p < .001), or the participant-generated keyword group (M = 2.00; SD = 2.22) (p = .02). The rote learning method group (M = 2.65; SD = 2.08) recalled more words than the researcher-supplied keyword group (p < .001), or the participant-generated keyword group (p = .002). No significant differences were found between the classmate-generated keyword group and the rote learning group, or between the researcher-supplied keyword group and the participant-generated keyword group. The interaction between both variables (immediate recall and keyword mnemonics) was not significant, F(3, 785) = 1.07, p = .36, ή2p = .01, power = .29.
Discussion
The results of this study have revealed that with immediate recall, participants recalled a significantly higher number of Spanish meanings of Basque words at one week, than participants with no immediate recall. This result is in agreement with studies where half of the participants had delayed recall, and the other half both immediate and delayed recall (Carney & Levin, 1998, 2008; Wang & Thomas, 1995; Wang et al., 1992, 1993).
Moreover, the classmate-generated keyword group recalled more meanings of Basque words at one week than the researcher-supplied keyword group, or the participant-generated keyword group. This result coincides with other studies underscoring the efficacy of the classmate-generated keyword selection method (Campos & Ameijide, 2014; Campos et al., 2002; Campos, Amor, et al., 2004a, 2004b, Campos, González, et al., 2004; Campos et al., 2010).
The rote learning group recalled more words at one week, than the researcher-supplied, or the participant-generated keyword group. This result substantiated the findings of several studies where the rote learning method outperformed at one week the keyword mnemonics groups (Campos et al., 2003; Thomas y Wang, 1996; Wang et al., 1992, 1993). However, this result was inconsistent with work asserting the superiority of keyword mnemonics over all other learning methods (Campos et al., 2010; Rodríguez et al., 2014; Sagarra & Alba, 2006; Siriganjanavong, 2013; Soleimani et al. 2012; Tavakoli & Gerami, 2013; Zhang & Schumm, 2000).
No significant differences were observed in delayed recall between the classmate-generated keyword group and the rote learning group, a result that corroborated the findings of other studies reporting no significant differences at one week between the keyword method and the rote method (Carney & Levin, 1998, 2008; Wang et al., 1992, 1993; Wang & Thomas, 1995), or between the researcher-supplied keyword group and the participant-generated keyword group. This result coincides with several studies that found no significant differences between these two keyword learning techniques (Campos et al., (2004a; Carney & Levin, 1998; King-Sears et al., 1992; Shapiro & Waters, 2005; Thomas & Wang, 1996).
This study has a number of limitations such as the type of learning list used was a short vocabulary list, with a specific task completion time in only one age group, and these are variables influencing the efficacy of keyword mnemonics (Campos et al., 2003). Further research is required to analyse the optimum conditions for enhancing the efficacy of keyword mnemonics in the learning of Basque. Moreover, studies are required to assess the efficacy of keyword mnemonics in autonomous learning in comparison to other learning methods.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
