Magoosh GRE

Source monitoring involves thinking about our memories, it involves the processes in ascertaining the origins of our knowledge and making decisions about what sources the knowledge or memories

| November 20, 2012

Introduction (2425)

 

Human memory is far from perfect, often; we see somebody we ‘know’ in the street but can’t for the life of us think where we know them from. This is basically a mistake within our memory systems; more specifically it is a ‘source monitoring error’. Source monitoring involves thinking about our memories, it involves the processes in ascertaining the origins of our knowledge and making decisions about what sources the knowledge or memories are from (Johnson, Hashtroudi &Lindsay, 1993). For example, source monitoring processes help us to decide whether we saw an event or whether we were simply told about it, whether we found important facts from a reliable source or a glossy magazine. Source monitoring errors may occur when someone is familiar with something (be it a person, place, event etc.) but misplaces the source of where that familiarity is from. For example one may see a person in the street whom one would not normally see in the street, the person may seem familiar to you, but you cannot place where you know them from. This familiarity, once placed with recollection will provide recognition of why one finds a particular person, place or event familiar. As with the person in the street, this problem is likely to arise when the subject is out of context, hence the ‘butcher-on-the-bus’ phenomenon (Yovel et al 2004). Familiarity does not help us out until it is combined with information such as spatiotemporal context of an episode, in which the memory was acquired, this will then aid recollection.

 

When we are thinking about memories we undergo heuristic (automatic or unconscious) judgement processes in order to locate, effectively, where we know something from without making any conscious effort. Reality monitoring helps us know or realise what memories are internally generated (what memories are thoughts) and what memories are externally derived (what memories are of events that actually happened). Johnson et al (1993; 1997) build on Johnson and Raye’s (1981) reality-monitoring framework in order to produce a source monitoring framework that suggests that source monitoring is a combination of, different attributes or characteristics of memories and judgement processes which help to discriminate between different memories and types of memories (Johnson et al, 1993; Johnson, 1997). Source monitoring distinguishes between all the different internal thoughts; between different externally derived events and also between internal and external sources (e.g. dreaming about something, waking up and thinking it had actually happened) (Johnson et al 1997). By ‘source’ it is meant the general conditions under which a memory was acquired, i.e. the contexts of the event, how and when it was witnessed. Furthermore, source attributions are made, according to the source monitoring framework, to different degrees of confidence and specificity which depend on the quality and quantity of information available at the time of initial experience. Johnson et al also point out that among other things, information available, criteria used to determine the source of a memory and task demands of attributing a source are all important factors (Johnson 1993).

 

When we are making judgements about memories (through a judgement process) we attribute source monitoring. Source monitoring takes into account different kinds of information, or different attributes, about memories in order to complete the task of locating the source of a given memory. Different characteristics of the memory that monitoring is based on are; perceptual information (acquiring sensory information), contextual information (spatial information and information about your surroundings at the time), semantic information (information about language use), affective information (emotions) and cognitive detail (thoughts or information that may have been acquired) (Johnson et al 1993, Johnson 1997). According to the source monitoring framework there are several factors which determine the ease and accuracy of identifying a source; the first is the type and the amount of  these different characteristics of the memory that monitoring is based on. The second is how distinctive these characteristics are from source to source, two sources with similar characteristics will be harder to distinguish from each other. The last factor is the effectiveness of the decision processes; richly detailed memories have unique attributes which enable decision processes to be effective (Johnson et al 1993).

 

Although most source monitoring decisions are indeed made in this heuristic or automatic manner, sometimes, more strategic processes are required to gather the appropriate information in order to avoid detrimental effects or situations. As well as heuristic processing, systematic or more ‘controlled’ processing is sometimes used (Johnson et al 1993). This more extended version of source monitoring processes is more complex. There may be other reasons or factors for you deciding, for example, what particular event occurred and when. Sometimes other beliefs, specific memories or general knowledge may have to be accessed in order to evaluate what you believe to be the source of your target memory (the memory you are trying to find the origin of). For example you may recall an event where a friend told you a story at a certain party; however other information about that friend being away travelling that year may lead you to recall that it could not have been that friend but must have been another.

 

Evidence that source monitoring involves memory characteristics is shown by comparing memories for perceived and imagined events. Events that are actually perceived have more perceptual and contextual detail, hence if a large amount of perceptual information is found; it is easier to make a source monitoring decision on that particular memory (Johnson et al 1993). One study that supports this framework is Johnson et al (1988), he asked participants to rate real events and dreamt or imagined events and rate them on memory characteristics, participants rated actually perceived events as having clearer temporal and spatial information and more perceptual information (Johnson et al 1993). Evidence that memories are attributed to sources by processes during the source monitoring procedure is shown by Raye, Johnson and Taylor. In their studies they used two different strategies to compare features of memories for internally generated and externally perceived events finding that given stricter instructions, participants are more selective about what information they need to preserve (Raye, Johnson &Taylor 1980). Evidence, through developmental research, that reality monitoring, external source monitoring and internal source monitoring are different classes of source monitoring problems is supported in poor performance on one but not another situation. As Hashtroudi found, older adults can be impaired in internal & external source monitoring but not in reality monitoring (Hastroudi et al, 1989).

 

Source monitoring can also be linked in with a range of other psychological and day to day phenomenon; such as; old-new recognition (Ferguson et al 1992), direct and indirect tests of memory (Richardson-Klavehn & Bjork, 1988), eyewitness testimony (Loftus, 1979) and misattribution of familiarity (Johnson & Foley 1984).

 

Although source monitoring can be applied in these different ways (automatic and controlled) and are used in every day life, there are indeed ways in which source monitoring errors can occur as seen with the example of the person in the street, in Jacoby’s False Fame effect and also through déjà vu . These inevitable source monitoring errors can incur practical, social and emotional consequences (Johnson et al 1993) and can happen in a number of ways. For example a source monitoring error may result in you telling a joke to someone who originally told you the joke, a source monitoring error may also result in accidental plagiarism that could have been easily avoided. Peters et al (2007) has shown that people who have ‘memories’ or beliefs about abnormal or implausible life events, commit an increased amount of source monitoring errors or in other words they make more source monitoring errors by ‘misclassifying familiar non famous names as famous names’ (Peters et al 2007, p162) in the false fame task than people who do not have these memories or beliefs. (Peters et al 2007). Peters et al (2007) specifically tested people who claimed to have ‘hypnotically induced previous life memories’ against a control group (Peters et al 2007, p163). Furthermore, Clancy et al (2002) found that people who claim to have experienced alien abduction or claim to have memories of such abductions, tend to falsely recall and recognise critical lure words that they have not been presented with (Clancy 2002) in the DRM paradigm (Deese 1959; Roediger & McDermott 1995).

 

One particular phenomenon related with source monitoring errors evolved in 1989. Jacoby et al compiled a paper containing studies that showed what he called the False- Fame effect (Jacoby & Woloshyn, 1989). During a false fame task participants are shown a list of non-famous names to read aloud and are told that all of the names are non famous. Later, the participants are presented with the same (old) non-famous names mixed with different (new) non-famous names as well as actual famous names and ‘fame judgements’ are made on each name. In general, findings show that old non-famous names regularly get judged as famous due to a source monitoring error wherein the participant has misattributed the familiarity of a name, the correct source of the name would be from the initial list studied however the participant has incorrectly judged the name to be famous due to familiarity and a source monitoring error combined. After being explicitly told that the names studied within the study stage are non famous names, the act then, of naming a non famous name as famous at test, must be a source monitoring error. Jacoby’s false fame effect is also evidence that the past can be used to influence present performance without intervention of conscious recollection. If conscious recollection was indeed present, the participant would remember that they saw the non famous name earlier in a study stage and locate that as the source of the memory, and not think to themselves ‘I recognise that name therefore it must be famous’. In Jacoby et al’s (1989) study, the unconscious influence of memory is due to divided attention as Jacoby tested participants on the false fame test in full vs. divided attention groups.  Jacoby & Woloshyn’s (1989) False-Fame Effect is a good way to show source monitoring errors (Jacoby & Woloshyn 1989).

 

A second phenomenon that is often seen to be related with source monitoring errors is when something may seem so familiar that you feel as though you have experienced it before, be it a place, a conversation or an event. This phenomenon is also to do with familiarity and is the familiar feeling of déjà vu. Due to the nature of déjà vu it has proved hard to study the phenomenon, however recent neurological advances have thought that déjà vu could potentially, at any one time, be down to one of the following; a slight and brief change in speed of transmissions; a short split in perceptual information causing the (present) experience to seem as though it is two different experiences or the presence of unconscious familiarity without the conscious recollection of the source of the familiarity. Déjà vu literally means ‘already seen’, Neppe (1983) described déjà vu as “any subjectively inappropriate impression of familiarity of a present experience with an undefined past” (Brown, 2003, p394). Brown (2004) also described déjà vu as a ‘jolting confrontation between our subjective sense of familiarity and our objective evaluation of unfamiliarity’ (Brown, 2004, p256). The déjà vu experience is often seen as another example (as well as the false fame effect) where the source of ‘familiarity’ is not recollected, and so is essentially a source monitoring error. Kusumi (2006) states that déjà vu is caused by ‘comparable present experience and past experiences’ (Kusumi, 2006, p303). If one realises that they are experiencing or that they have experienced déjà vu it is because reality monitoring helps one to confirm that an experience is new even though it feels familiar, however at the time of déjà vu it is likely that one does not know they are experiencing it.

 

Jacoby 1989 shows that the past can be used to influence present performance without the intervention of conscious recollection and with source monitoring errors. This study remains in sync with this notion as participants are shown a set of stimuli and then shown again later knowing that they have previously seen a selection of the stimuli and know that those specific stimuli are non-famous. Therefore if they are to label an old non famous stimulus as famous their performance has been influenced by the past without conscious recollection. If the participants had consciously recollected where they recognised the stimulus from then they would have known they already saw it, remember correctly that they had been told it was non famous and then at test, they would have labelled the stimulus correctly as non-famous rather than incorrectly as famous. Jacoby also shows that he has produced this effect of unconscious influences through the help of dividing attention.

 

The earlier mentioned examples are a few ways in which the false fame test has been manipulated, or tested combined with different factors; the present study aims to explore into combining the false fame test with alternative stimuli that is; it will try to see if the same effects are seen when the participants look at actual images of faces on a computer screen rather than reading out words (names in word form). This study will also look into the question of whether people who experience the phenomenon of Déjà vu frequently do worse when tested on the false fame paradigm than people who do not experience it as often (frequency of déjà vu is defined by answers of the Inventory of Déjà vu Experiences Assessment (IDEA). This is believed to be an appropriate next step to take into research about source monitoring errors and who is more susceptible to making these errors. The hypothesis is ‘Are people with higher frequency déjà vu encounters more susceptible to False Fame tests?’

 

My predictions include that errors made on a false fame test will relate to frequency of déjà vu. That is, people who experience déjà vu more frequently will be more susceptible to false fame and therefore make more fame judgement errors at test, than people who experience déjà vu infrequently or less frequently. This is due to ….****************Another prediction is that old non famous faces will be judged wrongly as famous more often than new non famous faces due to unconscious influence of the past combined with source monitoring errors. A third prediction is that both the factors of frequency of déjà vu and also the type of face presented will have an overall joint affect on source monitoring errors and people will be more susceptible to false fame.

 


Method (1,397)

 

Participants

The total number of participants tested was 76; 64 females and 12 males. The participants’ ages ranged from 18 to 44, the mean age was 20.17 years with a standard deviation of 2.96 indicating that most of the ages are close to the mean. The participants were recruited from the University of Lincoln by advertisement of the study through email and recruitment via word of mouth; participants reading Psychology were enticed by a credit point which would in turn allow them to access the student pool during their own Independent study. All of the participants were either undergraduate students at the university reading various courses or otherwise involved with the university. The participants were placed into groups, after testing, based on the frequency of déjà vu experienced, established by the Inventory for Déjà vu Experiences Assessment (IDEA). Group 1 were participants who in general, encounter a low frequency of déjà vu and group 2 are participants who encounter a high frequency of déjà vu.

 

Materials

To conduct the experiment 60 images of non-famous faces and 30 famous faces were presented over three different displays for the first and third stage of the experiment combined. All the faces were cropped to a grey scale, chin to forehead, ear to ear section of each face in order to try to eliminate anything recognisable other than the face e.g. clothing, setting, body language, hair style and colour. These faces were all cropped to a width of 10cm to try to ensure that they were all roughly equal sizes and recognition would not be due to over or under sized images. The images were presented to participants using a presentation file with Microsoft PowerPoint 2010 on a plain black background via a 15.6” HD screen.

 

For the second part of the experiment the participants were given the first part (part A) of the Inventory of Déjà vu Experiences Assessment (IDEA) to fill in. The IDEA is a 24 part questionnaire designed to capture qualitative information about déjà vu experiences. The questionnaire is comprised of 2 sections; A and B consisting of 9 and 14 items respectively. Participants completed the first half of the questionnaire as a distracter task in between study and test, this then indicated whether participants experienced déjà vu often or not very often. The second part (part B) of the IDEA was given to participants who revealed a high frequency of déjà vu experienced in the first part and more specifically the first question.

 

 

Procedure

Each participant was greeted and given a consent form to read, sign and date. Once this was completed the participant was prompted to ask any questions he or she may have had.

 

The participants were then given instruction to carry out the first, study stage of the experiment. This involved a timed presentation of 30 faces to which the participant was asked to make an age judgement on each face by stating ‘over’ if the participant believed the face to be over 25 or ‘under’ if the participant believed the face to be under 25. This age judgement ensured deeper encoding of the faces by the participants and each face appeared on the screen in front of the participant for 3 seconds. During the study stage, one of two study presentations, containing 30 different non famous faces each, were used and assigned to participants where, for one study presentation, old non famous faces in the test presentation would be new non famous faces when the second study presentation was used. This first section of the experiment lasted approximately 90 seconds. ***CHECK***After the presentation the participants were explicitly told that all of the faces they had just seen were non famous.

 

The participants were then handed the first part of a 2-part, 23-item questionnaire to be completed and given back for the second stage of the experiment. This part of the questionnaire was to be filled out by everyone partly as a distracter task between study and test stage but also for the information about déjà vu experiences, the first half of the questionnaire lasted about 3 minutes.***CHECK***

 

The third part of the experiment was the test stage in which participants were again shown a timed presentation (each face appearing once again for 3 seconds) of the 30 previously seen faces from the study stage, a new set of 30 non famous faces and a further 30 famous faces randomly mixed all together to make a presentation 90 faces big. An effort was made in attempt to find faces that participants would hopefully recognise as famous but would not know the accomplishment that had led them to fame straight away without systematic processing. The participants would not have time to be able to think about this however as the presentation would very quickly move onto the next face for them to judge. Instructions read to participants at this point were very careful and precise. The participants were asked to state famous for a face that was ‘even vaguely familiar’ to them, participants were told that none of the famous faces that they would see would be as famous as someone like David Beckham but that a famous face would be famous on some level even if it was to a really low degree. The participants were asked to state out loud ‘famous’ if they recognised the faces as famous or ‘non famous’ if they did not recognise the face as famous. The experimenter noted the answers on a previously made checklist for all participants. Due to the counterbalance of faces in the first study stage, half the participants were seeing new non famous faces that the other half would see as old non famous faces.

 

After completion of the last stage of the experiment, participants were debriefed, thanked, awarded with their credit point and could then leave.

 

Design

The design consists of one unrelated (or between) subjects factor, frequency of déjà vu with two levels; low frequency of déjà vu and high frequency of déjà vu. The design also consists of one within (or related) subjects factor, type of face; with two levels; old non famous faces (that the participant had previously seen) and new non famous faces (that the participants had never seen before). Due to this a two-way mixed repeated measures ANOVA was computed to determine whether people who experience a higher frequency of déjà vu are more susceptible to making source monitoring errors with in a false fame task. A related t-test was also carried out on the scores obtained from famous faces, i.e. the number of famous faces that were actually judged (correctly) as famous to determine any differences of famous faces correctly being judged as famous between the two groups of high and low frequency of déjà vu.

 

Ethics

During testing all relevant ethical issues surrounding the study, delineated by the BPS guidelines were regarded.

 

The participants were all given a consent form to read, sign and date to prove that they had given their informed consent in to taking part. Informed consent was given to make sure participants were not tricked into anything they were not aware of; false or misleading information was not given to participants simply to gain consent; withholding slight information for purpose of study that will not negatively affect the participant in any way is different to deceit. Within the consent form they were informed that they had the right to withdraw from the study at any time and that their partaking was completely voluntary.

 

Throughout testing and afterwards all information about participants was kept completely private and confidential as the Data Protection Act (1998) requires. All data gained was kept anonymous and no results attached to identities were exposed in any way. Participants were assigned codes in cases of withdrawal or if the participant needed to contact the experimenter for other reasons.

 

During the study participants were not exposed to any kind of risk or anything that would make them feel uneasy, stressed or anxious. After the whole study was complete, participants were debriefed and any information not thoroughly explained before the testing due to purpose of study was explained. Debriefing also ensured that participants were still happy to be included as research and that they could now withdraw if they were not. Participants took away the consent and debrief forms with information about how to withdraw is they so wished after they had left the room.

Results (516)

 

For each subject the proportion of famously judged faces (the number of old non famous faces judged as famous, the number of new non famous faces judged as famous and the number of actual famous faces judged as famous) was calculated as a function of déjà vu frequency. These data are shown in Table 1.  Response times were not recorded.

Table 1. Mean proportion of scores (incorrectly judged fame) of all three types of face in both high and low groups of frequency of déjà vu.

From the table it is obvious that more errors were made on judgements given for old non famous faces than new non famous faces in both groups of low and high frequency of déjà vu experienced. This result on its own is in continuation with other false fame tasks that have revealed non famous but already seen faces (names in other cases) to be judged as famous regardless of participants being told explicitly that the faces are non famous (Jacoby et al 1989, Bartlett et al 1991, Peters et al 2007>>more>>>?) On average, both groups of frequency of déjà vu judged approximately half the amount of actually famous faces, (Mean =0.52 for both groups) as famous. The mean proportions of scores also show that the number of errors for old non famous faces were indeed higher in the high frequency déjà vu group than the low, suggesting my hypothesis may be true, however further analysis has shown that difference to be non significant.

 

The results of the Mixed Repeated Measures ANOVA treating the frequency of déjà vu (Group) as the between subjects factor and the type of face shown (Type) as the within subjects factor, showed the interaction between Group and Type to be not significant [F (1, 74) = 7.624; p = 0.007] indicating that the frequency of déjà vu experience had no direct effect on incorrectly judging old or new non famous faces as famous. The interaction was calculated in order to find out the effect of both factors together on fame judgements. This being a main prediction of the research question leads to a null hypothesis to be taken. The analysis of variance failed to reveal an overall effect of Group [F (1, 74) = 0.179; p = 0.673], indicating that the overall proportion of incorrect fame judgements did not differ between low and high frequencies of déjà vu experience, (one tailed). A main effect of test (Type) was found [F (1, 74) = 45.393; p< 0.001(p=0.00)] showing that incorrect fame judgements were greater in judgement of old non famous faces than new non famous faces (one tailed); this is still in continuation with other, previous research.

 

The famous faces judged as famous had the same mean proportion across both Groups, those who experience déjà vu at a low (SD=0.20) and high (SD= 0.17) frequency indicating that there is no difference in correct fame judgements of famous faces between the two Groups. An unrelated t-test showed that any difference there may be between scores was not significant [t (74) =0.000; p=1.000] as 1.000 is greater than 0.05.

 


Discussion

 

The amount a face was judged as ‘famous’ was measured (see table 1.) for each of the three types of face; famous, old non famous and new non famous and in each of the groups; high frequency of déjà vu experience and low frequency of déjà vu experience. This was in order to find out if people who experience a high frequency of déjà vu are more susceptible to false fame, i.e. will they make more incorrect judgements or source monitoring errors during the task. The proportions of means suggested at first that people who experience a higher frequency of déjà vu are indeed more susceptible to a false fame task. Further analysis was carried out and an interaction calculation between group and type in order to find out whether people do indeed do worse on a false fame test if they experience higher levels of déjà vu showed that they do not. Therefore a null hypothesis rejecting my hypothesis is to be taken showing that people with higher levels or frequencies of déjà vu experience do not necessarily do worse on a false fame test that uses faces as stimuli as opposed to names in word form.

A calculation on the effect of group showed that the factor of whether people experience high or low frequencies of déjà vu does not affect how many errors are made. In other words no significant difference was found in the number of incorrect fame judgements, or source monitoring errors, between the two groups. This is contrary to the prediction that people in the higher frequency of déjà vu experience will do worse at the false fame test and obtain more incorrect fame judgements on old non famous faces than people in the low frequency déjà vu groups. However people in this high frequency of déjà vu experience group did, on average, make slightly more incorrect fame judgements (Mean=0.30) than people in the lower frequency of déjà vu group (Mean=0.27 ).

 

A calculation on the main effect of test was found, showing that people made more source errors or incorrect fame judgements on old non famous faces than new non famous faces which is in keep with Jacoby et al’s false fame experiment (1989) and also confirms my prediction that old non famous faces will be judged wrongly as famous more often than new non famous faces due to unconscious influence of the past combined with source monitoring errors. In other words people were more likely to judge a face as famous if they had already seen it regardless of the fact that they had been explicitly told that the faces they had seen in the study stage were all non famous faces.

 

>>> talk about – past influence, familiarity no recollection and source monitoring error combined.

 

 

As already stated my hypothesis did not turn out to be true and people who experience high frequencies of déjà vu were not more susceptible to false fame in this experiment, however if the same experiment was to be carried out with the original stimuli, i.e. using words instead of faces plus similar groups the outcome may be desirable. This may be due to the fact that the original test had been used and been successful in the past.

People who are more ****well travelled etc***** are more susceptible to the false fame test therefore people who experience déjà vu may be too. (is it people who experience déjà vu are oftern wel travelled etc or is it people who do bad on false fame task— I think its déjà vu??? In which case that is wrong

Overall my results do support other false fame tests such as Jacoby et al (1989), Peters et al (2007)… Bartlett et al 1991 find all examples I can where false fame effect are shown; intro?

http://en.wikipedia.org/wiki/Source-monitoring_error

http://psiexp.ss.uci.edu/research/papers/memory/KelleyJacoby.pdf

http://scholar.google.co.uk/scholar?q=false+fame+studies&hl=en&as_sdt=0&as_vis=1&oi=scholart

 

Findings of my experiment

What it shows

What it supports

Possible reasons for results

          e.g. different instructions

relationships of anything

use of distractor tasks

links to current experiment

further research in area

evidence of false fame effect

 

Older adults more likely to call old nonfamous names famous in fame judgement task, however déjà vu is observed less frequently in older adults which adds a reason to the point that people who experience déjà vu more oftenwould not be more susceptible to the false fame task.

 


Bibliography

 

Brown, A. S. (2003). A Review of the Déjà vu experience. Psychological Bulletin, 129, 394-413.

Brown, A. S. (2004). The Déjà vu Illusion. Current directions in Psychological Science, 13(6), 256-259

Ferguson, S., Hashtroudi, S. & Johnson, M. K. (1992). Age differences in using source-relevant cues. Psychology and Aging, 7, 443-452.

Hashtroudi, S., Johnson, M. K., & Chrosniak, L. D. (1989). Aging and source monitoring. Psychology and Aging, 4, 106-112.

Jacoby, L. L., & Woloshyn, V. (1989). Becoming Famous Without Being Recognised:Unconscious Influences of Memory Produced by Dividing Attention. Journal of Experimental Psychology, 118, 115-125.

Johnson, M. K., Foley, M. A., (1984). Differentiating fact from fantasy: The reliability of children’s memory. Journal of Social Issues, 40(2), 33-50.

Johnson, M. K., Hashtroudi, S., & D. Lindsay, S. (1993). Source Monitoring. Psychological Bulletin, 114, 3-28.

Johnson, M.K. (1997). Source Monitoring and Memory Distortion. Philosophical Transactions: Biological Sciences, 352 (1362) 1733-1745.

Kusumi, T. (2006). Human metacognition and the déjà vu phenomenon. In K. Fujita & S. Itakura- (Eds.) Diversity of Cognition: Evolution, Development, Domestication and Pathology, (pp 302-314). Kyoto University Press.

Loftus, E. F. (1979). Eyewitness Testimony. Cambridge, MA: Harvard University Press

Peters, M. J. V., Horselenberg, R., Jelicic, M., Merckelbach, H. (2007). The false fame illusion in people with memories about a previous life. Consciousness and Cognition 16, 162-169.

Richardson-Klavehn, A. & Bjork, R. A. (1988) Measures of memory. Annual Review of Psychology, 39, 475-543.

Yovel, G., Paller, K. A. (2004). The neural basis of the butcher-on-the-bus phenomenon: when a face seems familiar but is not remembered. NeuroImage, 21, 789- 800.

http://www.psywww.com/resource/apacrib.htm

Tags: ,

Category: Free Essays, Health