Did I text you? The influence of the mode of transmission on destination memory

ABSTRACT In a society where people often communicate through digital technology, it is crucial to investigate whether the transmission mode influences destination memory performance (our capacity to remember to whom we transmitted certain information). In Experiment 1, we asked young adults (N = 31) to share of a set of familiar proverbs only by typing and the rest by both typing and saying them aloud. Better destination memory was observed when the information was transmitted by the two means (aloud and typing). Did this better performance occur because participants shared the information aloud or because the information was transmitted by two means? In Experiment 2, young adults (N = 34) shared familiar proverbs aloud, by typing, or simultaneously aloud and by typing. Results showed that transmission aloud led to a better destination memory than typing, and no further improvement occurred when the transmission was both aloud and by typing. Additionally, no differences were observed regarding item memory, supporting the idea that item and destination memories are two different types of memory.

Given the vast number of interpersonal interactions in our daily lives, it is not surprising that we frequently fail to remember the person to whom we transmitted a specific piece of information.This failure causes social embarrassment because it leads us to repeatedly share the same information or history with the same recipient.Besides embarrassment, memory lapses of this type can also negatively impact interpersonal relationships.For example, a person can believe that they said something to someone (when, in fact, they did not) and later blame the other person for not listening or paying attention.Indeed, greater communication effectiveness has been associated with better destination memory (Gopie & MacLeod, 2009;Johnson & Jefferson, 2018).This is one main reason why it is so important to remember to whom we previously transmitted information.
To remember the person to whom we said something, we rely on destination memory, a type of episodic memory recollected in the context of a particular time and place concerning oneself as a participant in the event (Tulving, 1983).The opposite processthe capacity to remember which person shared information with ushas been called source memory (Johnson et al., 1993).Both types of memory are associative because they imply remembering the association between the information transmitted and the person (recipient or transmitter).
The standard paradigm used to study destination memory was developed by Gopie and MacLeod (2009) and consists of a study and a test phase.Participants say aloud 50 pieces of information to 50 faces in the study phase.For that, a fact is first presented and, after reading it silently, the participant presses a key in the keyboard, resulting in a face presentation.At this moment, the participant shares the fact aloud with the face presented.As Gopie and MacLeod (2009) proposed, celebrity faces are often used because we usually spread information to familiar people (i.e., family, friends, and colleagues).The procedure to study destination memory was similar to the procedure to study source memory, with the principal difference being the order of presentation of the stimuli (i.e., facts and faces) in the study phase.In source memory, the celebrity's face was initially shown, followed by the presentation of a fact.At this point, the participant would read the fact as though the celebrity had conveyed it.
After encoding all 50 associations (i.e., fact-face pairs), participants usually randomly perform two recognition tests: an item memory test and an associative memory test.In the item memory test (i.e., memory for the facts and the faces) a single stimulus at a time (i.e., a fact or a face) is presented, half studied (i.e., target), and half not studied (i.e., new stimulidistractor).In the destination memory and source memory test, fact-face pairs are presented to the participants, and they answer whether they said that fact to that face or whether that face had told that fact, respectively.Half of the pairs are shown precisely as they had been learned in the study phase, whereas the other half are re-matched.
Better source memory than destination memory performance was observed (Gopie & MacLeod, 2009), meaning that it was more challenging to recognise with whom we shared information than to recognise who told us something.This result was extended to older adults (Gopie et al., 2010) and was explained based on attentional resources.When participants transmit information to someone else, their attentional focus is on themselves and the processes required to share information, which leaves fewer attentional resources available to associate the information with the person to whom it was transmitted.This leads to a weak information-person association.In the source memory task, where participants received information from someone else, the attention was directed to the incoming information and the person who provided it, resulting in a better information-person association (Gopie & MacLeod, 2009).
The influence of several variables on destination memory performance has been studied, namely, information-related variables (e.g., the transmission of proverbs vs. personal facts - Gopie & MacLeod, 2009;Johnson & Jefferson, 2018;Pinto & Albuquerque, 2023) and recipient-related variables (e.g., the distinctiveness of the recipients' faces - Barros et al., 2021).In sum, it seems that better destination memory occurs when the attentional focus is driven to the recipient of the information, for example, by adding distinctive features to the faces (Barros et al., 2021), which leads to a distinctive association between the information and the face.When the attentional focus is directed to the information, for example, by asking participants to transmit personal facts, destination memory is poorer (Gopie & MacLeod, 2009;Johnson & Jefferson, 2018;Pinto & Albuquerque, 2023).
However, Pieschl and Moll (2016) observed a contrasting result: Better destination memory occurred when participants transmitted personal facts than non-personal ones.In their research procedure, participants were asked to share information by typing it through a social network simulation.Participants were asked to share personal or non-personal information (between-participants design) to large or small audiences (within-participants design).Participants had to choose one of two categories (i.e., between two personal or two non-personal categories) and disclose a fact by typing it into an empty text field at the bottom of the computer page.For example, participants could be presented with two personal facts categories (e.g., "your opinion regarding cigarette consumption" vs. "your opinion regarding meat consumption").Then, participants had to choose which opinion they wanted to transmit (e.g., their opinion about cigarettes) and then type the personal fact into the computer.
One significant difference between Pieschl and Moll's study (2016) and other studies (e.g., Gopie & MacLeod, 2009;Johnson & Jefferson, 2018;Pinto & Albuquerque, 2023) that can explain the contradictory results was the type of personal information transmitted.Pieschl and Moll (2016) asked participants to share personal opinions about different topics, whereas, in the other studies (e.g., Gopie & MacLeod, 2009;Johnson & Jefferson, 2018;Pinto & Albuquerque, 2023), they were asked to transmit personal facts, like their favourite food or colour.The type of personal information should be a variable to be considered and further explored.Another difference between the studies was how the information was transmittedorally vs. by typing.In most of the studies on destination memory, participants have been asked to transmit the information aloud (e.g., personal facts, proverbs, etc.).In the study by Pieschl and Moll (2016), participants had to type the facts on the computer.So, it is possible that how the information is transmitted to the recipient can influence memory for the information and also destination memory.
Indeed, it has previously been observed that reading out loud (in a normal speaking voice) improves memory to a greater extent than other production modalities (e.g., reading silently and writing)a memory phenomenon designated as the production effect (MacLeod et al., 2010).Also, better memory was observed for typed words in a computer than for silently read words (Forrin et al., 2012).The authors concluded that typing involves unique encoding features compared to silently reading words, and this uniqueness in encoding provides additional and discriminative information for the typed words, leading to better memory.
It has been shown that words read aloud were remembered better than words produced by other unique responses (e.g., typed words), which were, in turn, remembered better than unproduced (i.e., silently read) words (Forrin et al., 2012).Also, it is essential to note that the production effect occurred not only with words but also with sentences (Ozubko et al., 2012).An example of a sentence included in this study was "Do not go there tonight."The production effect was also observed when texts, such as educationally relevant essays, were presented to the participants (Ozubko et al., 2012).
Therefore, in sum, previous literature (Ozubko et al., 2012) showed a production effect with different types of stimuli (words, words-pairs, sentences, and texts), various test formats (recognition test and fill-in-the-blanks test), and several retention intervals (immediate test vs. oneweek delay test).These results suggest that any mode of production that features unique, item-specific responses will result in a distinctive record in episodic memory that can then be retrieved and used during a memory test.
Besides the words and the sentences, the production effect was also noted in pair-association learning paradigms (Putnam et al., 2014) to examine whether production can enhance the memory of associative information.Participants studied word pairs by reading both words aloud, silently, or by typing the word pairs in a text box.In Experiments 1 and 2, no production benefit was observed.The authors explain the absence of the effect as being due to the inclusion of a semanticrating task (Putnam et al., 2014).Rating how related the two words seemed to promote an associative link between the words, providing a stronger mnemonic effect that overshadowed the production effect.However, when this semantic-rating task was removed in later experiments, results showed that production did enhance memory for paired associates.
This result occurred in a cue recall test, where participants needed to recall the target word when presented with its studied cue word.Additionally, the same pattern of results was observed in a pair recognition test, where participants responded recognise, answered yes or no, the target members of the pairs (Putnam et al., 2014).Therefore, the authors concluded that speaking aloud and typing are effective forms of production relative to silent reading, with speaking providing the more significant benefit.In sum, they concluded that production enhances the associative link between the two elements of a pair (Putnam et al., 2014).
Given the influence of the production effect in pairassociation learning paradigms, and considering that destination memory is a type of associative memory, our goal was to explore the influence of the mode of transmission on item and destination memory in two experiments.Since we live in a society where we often communicate through digital technology, it would be valuable to assess how we retain the destination of the emitted message.For this reason, in Experiment 1, participants were asked to share half of a set of familiar proverbs aloud and the other half by typing.To better understand the results of Experiment 1 and the influence of the mode of transmission on item and destination memory, in Experiment 2, participants were asked (1) to type and share aloud proverbs, (2) only to type proverbs, and (3) only to share aloud proverbs.

Experiment 1
In a world where interpersonal communication occurs frequently using digital technology and texting (i.e., the act of composing and sending electronic messages), it is important to observe whether there are differences in destination memory performance when participants are transmitting information only by typing compared with information shared both by typing and speaking.In this way, we simulate situations where people transmit information by texting compared with situations where people text but have the opportunity to transmit the information aloud to the other person.
To accomplish this, we applied the standard paradigm used to study destination memory (Gopie & MacLeod, 2009), where the only manipulation was how participants were asked to transmit information.Based on the literature review, two possible outcomes could occur.Given that the production effect was also observed in studies examining associative memory (Putnam et al., 2014), where production of the pairs of words aloud led to a better associative link between the two elements of a pair, one hypothesis is that better destination memory would occur when participants transmitted information orally and by typing compared with only typing.However, Hourihan and Smith (2016) concluded that both parts of a studied pair need to be produced (i.e., verbalised aloud) for a memory advantage to occur.So, because participants did not produce (either by typing or by saying aloud) the celebrities' namesthey only produced the proverbsanother hypothesis is that no differences would be observed between the two conditions because only one element of the pair was produced.
Our second goal with this experiment was to observe whether better item memory would occur when information was transmitted aloud and by typing rather than only by typing.Considering the production effect (MacLeod et al., 2010), producing the proverb (i.e., speaking aloud the information besides typing) should increase distinctiveness, and this encoding uniqueness should provide additional and discriminative cues leading to better item memory than when proverbs are only typed.

Participants
The sample consisted of 31 undergraduate students (30 females) with ages ranging between 18 and 30 (M age = 20.48,SD = 3.01).The sample size was calculated a priori through G*Power (Faul et al., 2007), targeting a paired samples t-test and using an alpha (α) of .05, a medium effect size (Cohen's d = .50),and statistical power of .80.
Since most of the studies in the destination memory domain have applied a between-participants design, we opted to use a medium effect size, following the study conducted by Barros and colleagues (2021, Exp. 1) that also employed a within-participants design.
The study involved individuals fluent in European Portuguese with either normal vision or vision corrected to normal.Written consent was acquired from all participants, who were rewarded with course credits for their involvement.The local Ethics Committee approved this experiment.
From these proverbs, two lists of 30 proverbs each were created, with similar values of length (M List1 = 6.37;M List2 = 6.50), familiarity (M List1 = 4.64; M List2 = 4.67), and valence (M List1 = 3.13; M List2 = 3.12).All stimuli were counterbalanced so that each proverb was presented in each of the two conditions, and they were displayed in 14-point lowercase white font on a black background.

Faces
Sixty celebrity pictures (e.g., Barack Obama) were selected from a celebrity database previously validated for the Portuguese population using the same age groupyoung adults (Lima et al., 2021).The 60 celebrity pictures selected had over 80% recognition (M = 97.17;SD = 3.61) and 75% naming accuracy (M = 91.69;SD = 5.34).The selection of the images was balanced in terms of background, consisting of 30 Portuguese and 30 international faces, as well as gender, with 30 male and 30 female faces.These variables were proportionally represented.All images were shown with a 9×9 cm dimension and in colour against a black background.

Design
Two independent variables were manipulated: (1) Type of Transmission (spoken and typed vs.only typed) and (2) Type of Test Stimuli (proverbs vs. faces).Both variables were manipulated in a within-participants design.The dependent variables considered were the sensitivity (d') score [z(P(hit))z(P(false alarm))] and the response bias (C ) score [ -(z (Hit) + z (False Alarm))/2], calculated for both memory tests (item memory and destination memory).Hits refer to "yes" responses to the faceproverb pairs, proverbs, and faces that were presented in the study phase (correct "yes" answers).False alarms refer to "yes" responses to face-proverb pairs, proverbs, and faces that were not presented in the study phase (incorrect "yes" answers).

Procedure
The stimulus presentation and response recording were controlled with the software E-Prime 3.0 (Psychology Software Tools, Inc., 2016).Participants were first asked to sign an informed consent and to complete a sociodemographic questionnaire.The main procedure included two phases: study and test phases.In the study phase, participants typed and spoke 25 proverbs and only typed (without reading aloud) the other 25 proverbs.The proverbs were randomly paired with the 50 celebrity faces.Participants were not told that their memory would be tested later.Each study trial started with a white fixation cross on a black background for 1000 ms, and then a proverb was presented.After reading silently and memorising the proverb, participants were instructed to press the keyboard spacebar.This resulted in a blank screen for 250 ms, followed by a colour picture of a celebrity face.For 25 celebrity faces, a keyboard symbol appeared above, meaning that the participant only had to type the proverb using the keyboard.A megaphone symbol was assigned for the remaining 25 faces, meaning the participant had to type and speak the proverb.To ensure that the participants attended to the face in both conditions, they were asked to say out loud at the end of each trial whether the celebrity was male or female.This procedure was repeated in a random way until the participant had typed and told 25 proverbs and had only typed 25 proverbs to the 50 faces.
After the study phase, all participants completed two counterbalanced memory tests: the item memory test and the destination memory test.These two tests used entirely nonoverlapping sets of stimuli to prevent crosstest contamination.The item memory test randomly presented 20 faces and 20 proverbs (half studiedtargetsand half not studieddistractors).For the targets (i.e., proverbs presented before), half had been only typed, and half had been spoken and typed.The participant had to indicate whether that item had appeared during the study phase.The participant responded "yes" by pressing the "c" key on the computer keyboard or "no" by pressing the "m" key.Once a response was made, a blank screen was displayed for 250 ms, and the subsequent test trial followed.
On the destination memory test, 40 face-proverb pairs were in random order shown: Twenty pairs had been presented during the study phase, and the other 20 were presented with random re-pairings of previously studied proverbs and faces (half had been presented in the only typed condition, and half had been presented in the spoken and typed condition).Participants reported whether they had told that fact to that face."Yes" and "no" responses were made by pressing the same keys as in the item memory test.Each pair remained visible until the participant responded.After each answer, a blank screen was displayed for 250 ms, and the subsequent test trial was presented.The entire procedure took approximately 20-30 min.

Results
The mean proportion of hits, false alarms, d', and C values were analyzed for the item and destination memory tests and are shown in Table 1.The software used for the data analysis was JASP 0.11.1 (JASP Team, 2021), and a significance level of .05 was applied for all statistical tests.
For item memory, we performed a 2 (Type of Transmission: spoken and typed vs.only typed) × 2 (Type of Test Stimuli: proverbs vs. faces) repeated measures ANOVA, performed on the d' data, with the Type of Transmission and the Type of Test Stimuli as the withinparticipants factors.The non-significant main effect of the Type of Test Stimuli, F(1, 30) = 0.25, p = .62,h 2 p = .008,showed similar item memory for the proverbs (M = 2.19, SD = .55)and for the faces (M = 2.12, SD = .61).Also, the non-significant main effect of Type of Transmission, F(1, 30) = 0.16, p = .69,h 2 p = .005,showed similar item memory when the proverbs were spoken and typed (M = 2.14, SD = .42)compared to when participants had only typed the proverbs (M = 2.17, SD = .57).No interaction was observed between the two variables, F(1, 30) = 0.002, p = .96,h 2 p = .0007.

Discussion
Since interpersonal communication occurs frequently using technology and texting, our goal was to observe differences in destination memory performance when participants were transmitting information by typing or simultaneously orally and by typing.No study so far has compared the different types of information transmission and their influence on destination memory performance.
Therefore, in this experiment, we opted to simulate situations where people transmit information by texting compared with situations where people text but also transmit the information aloud to the other person.
We hypothesised that participants had better item and destination memory performance when they typed the proverb and said the information aloud than when they only typed it.This hypothesis was based on the idea that speaking aloud and typing the information should increase the distinctiveness of the event and, consequently, lead to better memory.Indeed, it was previously shown that words read aloud were remembered better than words produced by other unique responses, like typing (Forrin et al., 2012).
This advantage for reading aloud was also observed for sentences (Ozubko et al., 2012), which suggests that any mode of production that features unique, item-specific responses will create a distinct episodic memory that can then be used during a memory test.Similarly, the production effect has also been observed in a paired-associate learning paradigm (Putnam et al., 2014), where production enhances the associative link between the two elements of a pair.
We did not confirm our hypothesis regarding item memory because we did not observe a significant difference between memory for proverbs that were typed and transmitted aloud versus memory for proverbs that were only typed.This result may be explained by the low number of stimuli usually used as targets and fillers in the item memory recognition test.As mentioned, the item memory task presented 20 proverbs, 10 studied (hits), and 10 not studied (distractors).Five of the ten proverbs presented before were typed and spoken, whereas the remaining five were only typed.
However, regarding destination memory, as expected, better destination memory was observed when participants typed and spoke the proverbs than when participants only typed the proverbs.Nevertheless, we cannot conclude that better destination memory when participants said and typed the proverbs occurred because of the production effect (where transmission aloud results in better memory) or because participants transmitted the information through two means of transmission (aloud and by typing), thereby deepening the information processing.

Experiment 2
In Experiment 1, better destination memory was observed when the proverbs were transmitted aloud and by typing than when the proverbs were only typed.As mentioned, this result can be explained by the production effect, where better memory occurred when the information's production (i.e., read aloud) occurred compared with typing (Forrin et al., 2012).To explain this result, the authors (Forrin et al., 2012;MacLeod et al., 2010) proposed that a word produced aloud during the study phase has additional sources of discrimination relative to a word that was not produced, such as an auditory cue and an articulatory cue, thereby becoming more distinctive.So, it was concluded that any distinct, item-specific response should result in a production effect, leading to better memory.Also, it was previously observed that although mouthing, writing, and whispering words made them more memorable than silent reading, all three productions were worse than reading aloud (Forrin et al., 2012).Reading aloud seems to benefit from additional processing on two distinct dimensions relative to silent reading: both articulatory and auditory processing.However, another explanation for better destination memory for information shared aloud and typed is that communication by two transmission means rather than just one can result in more profound stimulus processing and better memory.
Therefore, in Experiment 2, we wanted to compare three conditions: aloud and typing, typing only, and aloud only.This will allow us to determine whether better destination memory occurred in Experiment 1 because the proverbs were shared aloud (i.e., where the transmission aloud leads to a better destination memory) or because participants used two modes of transmission (speaking aloud and typing) instead of just one (typing alone) for conveying the information.

Participants
The sample consisted of 34 undergraduate students (30 females) with ages ranging between 18 and 32 (M age = 20.62,SD = 4.29).The sample size was calculated a priori through G*Power (Faul et al., 2007), targeting a repeated measures ANOVA and using an alpha (α) of .05 and a medium effect size (Cohen's d = .50).This sample size was chosen considering Experiment 1, where we obtained a Cohen's d = .54.To achieve a statistical power of 0.80, 28 participants were determined to be sufficient.
As in Experiment 1, the study involved participants fluent in European Portuguese with either normal vision or vision corrected to normal.All participants provided written consent and received course credits for their participation.Approval for this experiment was obtained from the local Ethics Committee.

Proverbs
Another six familiar proverbs were selected from the previous study developed by Barros and colleagues (2021) and added to the stimuli list of Experiment 1, making a total of sixty-six Portuguese proverbs.From these proverbs, three lists of 22 proverbs each were created, with similar values of length (M List1 = 6.41;M List2 = 6.36;M List3 = 6.36), familiarity (M List1 = 4.65; M List2 = 4.68; M List3 = 4.68), and valence (M List1 = 3.24; M List2 = 3.22; M List3 = 3.24).The stimuli were counterbalanced, so each proverb was presented in the three conditions.The proverbs were presented in 14-point lowercase white font against a black background.

Faces
In addition to the sixty celebrity pictures (e.g., Barack Obama) from Experiment 1, another six celebrity faces were selected from the celebrity database previously validated for the Portuguese population (Lima et al., 2021).The same parameters were kept, with the 66 celebrity faces having over 80% recognition (M = 97.07;SD = 3.59) and 75% naming accuracy (M = 90.81;SD = 5.96).A copy of the material used in the two experiments (proverbs and faces) is available on the OSF Platform (https://osf.io/xsy3h/?view_only=769cefd303c445af857764028afeea34).

Design
The independent variables manipulated were: (1) Type of Transmission (spoken and typed vs.only typed vs.only spoken) and (2) Type of Test Stimuli (proverbs vs. faces).Both variables were manipulated in a within-participants design.The only difference between experiments is that the first independent variable (i.e., the Type of Transmission) now had three conditions, not just two.The dependent variables were the same, the sensitivity (d') score and response bias (C ) score, calculated for both memory tests (item memory and destination memory).

Procedure
The stimulus presentation and response recording were again controlled with the software E-Prime 3.0 (Psychology Software Tools, Inc., 2016), and the procedure was very similar to the procedure of Experiment 1.The only difference was that in the study phase, participants (1) typed and spoke 18 proverbs, (2) only typed (without reading aloud) 18 proverbs, and (3) only spoke (without typing) 18 proverbs.So, instead of sharing 50 proverbs, participants were asked to share 54 proverbs randomly paired with the 54 celebrity faces.Again, participants were not told that their memory would be tested later.
In the condition where proverbs were spoken, for 18 celebrity faces, a megaphone symbol appeared above, meaning that the participant only had to speak the proverb.In the condition where proverbs were typed, a keyboard symbol was assigned for 18 faces, meaning the participant only had to type the proverb using the keyboard.Finally, when proverbs were to be both spoken and typed, both symbols (the keyboard and the megaphone) were presented above the remaining 18 faces.To ensure that the participants attended to the face in all the conditions (specifically when they needed only to type the proverb), they were asked to say out loud whether the celebrity was male or female at the end of each trial, as in Experiment 1.
After the study phase, all participants completed the two memory tests (the item and the destination memory tests) in a random order.In the test phase, the only difference between Experiments 1 and 2 was the number of stimuli used since we added one condition to Experiment 2. So, in the item memory test, instead of 20 faces and 20 proverbs, 24 faces and 24 proverbs (half presented before -12 targetsand the rest not presented before -12 distractors) were presented.Four of the twelve proverbs presented before were typed and spoken, four were spoken only, and the remaining four were only typed.
On the destination memory test, instead of 40 faceproverb pairs, there were randomly shown 42 pairs: Twenty-one pairs had been presented during the study phase, and the other 21 were presented with random repairings of previously studied proverbs and faces (seven presented in the only typed condition, seven in the spoken condition and the remaining seven presented in the spoken and typed condition).The entire procedure took approximately 20-30 min.

Results
The mean proportions of hits, false alarms, d', and C values were analyzed for the item and destination memory tests and are shown in Table 2. Data analysis was conducted using the software JASP 0.11.1 (JASP Team, 2021), with a significance level of .05applied to all statistical tests.
We performed a 3 (Type of Transmission: spoken and typed vs.only typed vs.only spoken) × 2 (Type of Test Stimuli: proverbs vs. faces) repeated measures ANOVA with Type of Transmission and Type of Test Stimuli, performed on the d' data.as within-participants factors.A main effect of Type of Test Stimuli was observed, F(1, 33) = 21.79,p < .001,h 2 p = .40,showing better item memory for the proverbs (M = 2.13, SD = .36)than for the faces (M = 1.65,SD = .51).
Regarding the Type of Transmission and the interaction between the Type of Test Stimuli and the Type of Transmission, sphericity was violated.Therefore, the Greenhouse-Geisser correction was applied.However, there was a non-significant main effect of Type of Transmission, F(1.54, 50.67) = 2.17, p = .14,h 2 p = .06,and no interaction was observed between the two variables (p > .05),F(1.65, 54.33) = 0.38, p = .65,h 2 p = .01.Furthermore, we conducted the same analysis (i.e., a repeated measures ANOVA) employing the same factors with C criteria as the dependent variable.Similarly, we did not observe any significant difference in response bias.
Regarding destination memory performance, a repeated measures ANOVA was applied to d' data, in which we compared the typing and speaking of the proverbs, the typing of the proverbs without speaking, and the speaking of the proverbs without typing.The analysis revealed a main effect of Type of Transmission, F(2,66) = 3.75, p = .03,h 2 p = .10.Post hoc tests with Bonferroni correction revealed better destination memory when participants spoke (M = 0.92, SD = .78)than when they typed (M = 0.55, SD = .86).As in Experiment 1, we observed better destination memory when participants both spoke and typed (M = 0.86, SD = .85)than when they shared the information only by typing (M = 0.55, SD = .86).However, in this experiment, the difference was not statistically significant (p = .08).
Most relevantly, no difference occurred in destination memory when proverbs were only spoken versus when they were both spoken and typed (p = .66).Considering the limitations of null hypothesis significance testing (Dienes, 2011), Bayesian analyses were conducted to examine the evidence in favour of the null hypothesis (H0: no differences in destination memory when proverbs were transmitted aloud compared with when proverbs were transmitted aloud and by typing) with the d' data as the dependent variable.We performed this analysis on the JASP software (JASP Team, 2021, Version 0.11.1),where we obtained the Bayes factors (denoted as BF10).A BF10 < 1 indicates support for the null hypothesis (H0).In contrast, a BF10 > 1 supports the alternative hypothesis (H1).The Bayes factor of BF10 = .20indicates evidence for the H0, that is, no differences in the destination memory performance when proverbs were transmitted aloud compared with when proverbs were transmitted aloud and by typing.

Discussion
To better understand Experiment 1, we applied the same paradigm in this experiment but added another condition speaking without typing the proverbs.In this way, we could compare only spoken, typing, and doing both.Our goal was, therefore, to better understand whether the results of Experiment 1 (i.e., better destination memory when proverbs were spoken and typed compared to when they were only typed) occurred because of the production effect (i.e., because the participants spoke the information) or because of the double transmission of the proverbs that might have promoted deeper processing of the stimuli, resulting in better destination memory.
The results showed better destination memory when participants transmitted the information aloud than when they typed.This result supports the production effect, which holds that the production of the information results in an additional source of discrimination relative to a word that was not produced, making the event more distinctive (MacLeod et al., 2010).When participants shared information aloud, they had more cues (articulatory and auditory cues) than when they only typed (motor cues).Indeed, it was previously observed that typing made the words more memorable than silent reading, but this type of production resulted in a smaller memory improvement than did reading aloud (Forrin et al., 2012).
Also, results pointed to the same pattern observed in Experiment 1, where better destination memory occurred when participants transmitted the information by two means (i.e., aloud and typing) than when they only typed.However, it is essential to underline that this difference was not statistically significant.Nevertheless, more important to the purpose of this second experiment, we did not find a statistically significant differences between transmission aloud and transmission by two means (aloud and typing).So, we can conclude that transmitting information aloud leads to better destination memory mainly due to the production effect.Adding another way of transmission (like typing) did not further improve this type of memory.
Regarding item memory, better item memory was observed for the proverbs than the faces, which had not occurred in Experiment 1.One possible explanation is that the additional condition, the information shared in different ways, resulted in a higher attentional focus on the information, leading to better memory of the proverbs than the faces.However, better item memory for proverbs shared aloud than for proverbs typed or transmitted by the two means was not observed, not replicating the production effect in the item memory task, maybe due to the low number of targets and fillers used on the recognition task.

General discussion
With these two experiments, we aimed to examine how transmission mode affects both item and destination memory.In Experiment 1, participants were instructed to type and speak half of the familiar proverbs while exclusively typing the remaining ones.With this manipulation, we observed better destination memory when participants shared through the two means (aloud and typing) than only the one (typing).
However, a question arose from these results.Did better destination memory occur because participants spoke the proverbs, inducing the production effect, or because the information was shared by two means, deepening the processing of the stimuli?To answer this question, Experiment 2 involved three modes of transmission: typing and speaking proverbs, solely typing proverbs, and only speaking the proverbs.
Results showed that the transmission of proverbs aloud led to better destination memory than typing, and no improvement occurred when another way of transmission was added (no differences were observed when information was spoken or was both spoken and typed).Therefore, we can conclude that the results of Experiment 1 occurred because participants shared aloud and not because they transmitted the information by two means.
So, when participants were asked to share the proverbs aloud, better destination memory occurred, regardless of whether this transmission occurred by one means (only speaking) or by two means (speaking and typing).A more significant memory advantage for speaking words than writing or typing words has already been observed, with the authors explaining that speaking benefits from distinct auditory processing that does not overlap with non-vocal productions or silent reading (Forrin et al., 2012).So, in the destination memory task, the results were congruent with the production effect investigation (Forrin et al., 2012;MacLeod et al., 2010) that showed that reading aloud has consistently led to the best memory for words, holding an advantage over nonauditory (writing, mouthing, and typing) and even over other types of auditory (whispering) production.
However, contrary to what was expected, transmitting information aloud did not lead to better item memory than did the transmission by typing or by the two means, as the destination memory test showed.In other words, no differences were observed for the proverbs when proverbs were only spoken aloud, only typed, or spoken and typed.This result could be due to the lower number of stimuli used in the item memory test since, in our experiments, the stimuli needed to be divided between the item and destination memory tasks.Another possible explanation for the absence of significant differences is the information's predictability, a characteristic present in sayings or proverbs.In a study developed by Rommers and colleagues (2020), participants were asked to read sentence-final words aloud or silently.These words could be predictable or unpredictable.For the word "broom," an example of a predictable condition was "John swept the floor with a broom."It was anticipated that the final word was broom.An example of an unpredictable condition was "He could not find the red broom."In this case, multiple words could have been used to complete the sentence.In both conditions, participants read the word broom aloud or silently.It was observed that the production effect was smaller for predictable words.Therefore, it was concluded that production improved memory to a lesser extent for highly predictable words (Rommers et al., 2020).Since our experiment used familiar proverbs (i.e., predictable information), this could explain obtaining the same pattern of results as was found by Rommers and colleagues (2020).The authors explained that a sort of covert mental "production" occurs when reading the predictable information silently.This internal speech production during reading predictable information seems to influence participants' memory in the test phase.So, even though auditory and motor cues were not presented, other cues were available for remembering these words (Rommers et al., 2020).
It is essential to note that our procedure was the standard one applied in destination memory research (Gopie & MacLeod, 2009), where participants memorised and read the information silently, and only when the face appeared did they transmit it aloud.Therefore, when participants transmitted the information aloud, they also saw the celebrity faces.In production effect research, typically, the participants read aloud the information when it is presented.The words were presented, and at this moment, participants read them aloud (Forrin et al., 2012;MacLeod et al., 2010).So, maybe the difference in the paradigms used can also help to explain the results.
Regarding the faces, in the item memory test, the only difference was observed in Experiment 2, where better memory occurred for the proverbs than for the faces.One possible explanation for this result is that transmitting information in different ways (i.e., three different ways) leads to an increased attentional focus on the information and in the process of sharing, leading to better memory of the proverbs than faces.
In future studies, when participants are asked if they transmitted that information to that person, they also could be asked to indicate whether they told the information aloud, by typing, or aloud and by typing.These data could generate more information about the influence of the mode of transmission of the information in destination memory.Participants could better remember to whom they said something when they shared the information aloud.However, it will be interesting to understand whether they also remember that the information was shared aloud.Remembering the transmission mode could help to remember to whom the information was shared, serving as a clue.Also, the same could be applied in the item memory task (participants must answer whether they transmitted the proverb before and if they transmitted it aloud, only by typing, or by the two means).
Additionally, given the results obtained in this study, it would be interesting to see what happens when the face is not available but only the name, simulating when we are texting someone.We increasingly communicate through written or vocal messages in both personal contexts (e.g., messages and social networks) and in professional contexts (e.g., emails and WhatsApp groups), where the destination faces are unavailable.Therefore, in a future study using a similar design, participants could be invited to send messages without looking at faces, with only the receiver's name displayed.
In sum, with this experiment, we can conclude that transmitting information aloud leads to better destination memory and that sharing information simultaneously in two ways (by typing and aloud) does not further improve memory.So, to better remember the person to whom we said something, it is probably best to share the information aloud (and not by typing) if we can choose how to transmit it.However, no differences at the item memory level were observed.Although more studies are required to understand why the production effect did not occur at the item memory level here, it was an interesting result because it supports the idea that item and destination memories are two different types of memory.A better destination memory does not imply a better item memory, and viceversa, confirming what has been previously reported (Gopie & MacLeod, 2009).

Table 1 .
Experiment 1: Mean proportion of hits, false alarms, d', and C to item and destination memory as a function of condition.

Table 2 .
Experiment 2: Mean proportion of hits, false alarms, d', and C to item and destination memory as a function of condition.
Note: Standard deviations of the mean are reported in parentheses.