Gillian Long, Alicia Roddy, Bethannie Swisher, & Tiffany Williamson (2001)

Two ways of information retrieval are recognition and recall. Recognition is a process where certain studied information is identified within a group of items. One theory, the spreading activation theory, has been one accepted explanation of information retrieval. A cue word, which is used to prompt another word, can activate a node, which is a connection between the cue word in the target word. A target word is the word that the cue word prompts. The longer it takes for the target word to be produced, the weaker the cue and target pair is (Ratcliff & McKoon 1988). The retrieval theory assumes that stronger word pairs are the result of immediate images chosen over other available images in the memory, much like free association. Both spreading activation theory and the retrieval theory discuss strength of the connections between word pairs. Strong word pairs mean that there is a higher probability of a certain cue producing a certain target. Along with strength of word pairs, there are types of connections. Word pairs can have either direct or indirect connections. A direct connection is a connection between a cue word and a target that is immediate, or needs no connector words. An indirect connection is one where connector words are needed in order for the cue word to successfully cue the target.

Research by Nelson, Bennett, and Liebert (1997) researched direct and indirect connections of word pairs. They found that indirect connections between word pairs were successful in recall testing. The current research is based on Nelson, et al.s (1997) research, but expands it by looking at both indirect and direct connections with cued recall and cued recognition tests. We are specifically looking at how strength of connection and type of connection of word pairs affect performance on recognition and recall testing.

From our knowledge of memory testing and past research, we hypothesize that recognition scores will be higher than recall scores. This is because information tested with recognition testing will not be limited to seven plus or minus two items, which is expected to happen to recall information. Having a participant bring forth information studied by presenting them with a cue tests recall. Because there will be no word bank or choices for the participants, it will be more difficult for participants to recall all fifteen words studied since recall and short term memory are rather limited. A study by Nelson, McEvoy, and Schreiber (1990) found that long-term memory is activated when a familiar word is studied. This means that when words are studied, recognition would be activated since recognition pulls more from long term memory than short term memory (du Plessis 1994).We also hypothesize that strong primes, or word pairs, will lead to better performance on both recall and recognition tests.

Forty Sweet Briar College students volunteered to participate in the study. Each participant was assigned to a certain word pair group : strong direct, strong indirect, weak direct, or weak indirect. Testing occurred during a one-45 minute session. During this session, participants viewed a PowerPoint presentation that flashed target words according to their assigned group at six-second intervals. When the filler task was completed, the participants took one of the two memory tests that corresponded with the word group they studied for (recall or recognition). For recognition testing, participants were given four target word choices for each of the fifty target words. They were then asked to select the correct target word that matched with the cue word presented. After this test was completed, the participants then studied words for the second test following the same procedure. After all analyses were done on the tests, participants were fully debriefed.

Our data showed that the mean recognition test scores, M=80.80% (SD=14.83), were higher than recall test scores, M=58.63% (SD=28.01). Analysis of the interaction of strength of connection and test type presented significance, F(1.36)=13.33,p<.05. From the significant differences found in the interactions between strength and type of connection, as well as strength of connection and test type, it was not surprising to find a significant three way interaction between strength, type of connection, and test (F(1,36)=5.17,p<.05). We split the three-way interaction by making two new ANOVAs which tested strength of connection and type of connection with either recall or recognition. After this further analysis, significance was found only between recall tests and strength of connection. To verify significance, we ran both a Tukey test and a Schefe test, which both reinforced the data which showed significance only between the recall test and strength of connection.

Our results for recall are similar to results found by Nelson, et al. (1997) who found that recall test scores were higher for words with relatively strong cues. While the combination of strength of connection and type of connection produce significance, our results do not support our hypotheses which state that direct primes will outperform indirect primes regardless of strength or test. Our results also show that strength of connection and type of connection have no effect on recognition testing. This is most likely due to the fact that the recognition tests rely on a participant being able to pull the studied target word from a word bank rather than bringing it purely from memory. Although strength and type of connection have no effect on recognition testing, there is a significant difference between recall and recognition testing. Consequently, there is significance to support the recall portion of our hypothesis, which states that strong primes lead to enhanced performance on both memory tests. Because strong primes have a higher probability of being connected with its cue word, these word pairs have a higher chance of being correctly paired up.