Semantic Priming: Effects of Related and Unrelated Words

Semantic Priming: Effects of Related and Unrelated Words on Response Times

Introduction to Cognition, Biological Psychology and Quantitative Research Methods

Abstract

The aim of this experiment was to measure the response times on the string of letters that were shown to the participants on a computer screen. The task of the participants was to decide whether the string of letters were words or non-words. The hypothesis of the experiment was that the response time in the lexical decision task would be significantly faster for target words related to the prime than for target words unrelated to the prime. The experimental design was repeated measures, where participants responded to all the stimuli that were propounded. Participants were presented with a string of words related to the prime, words unrelated to the prime, non-word created from a word related to the prime and a non-word created from a word unrelated to the prime. The findings of the experiment showed that participants responded faster when the letter strings were related words rather than unrelated words. This is because related words are closely associated in implicit memory, therefore the response times of the participants will be faster in recognising the words. Explanations for the results of the experiment will be discussed further in the report.

Introduction

Priming is an effect on implicit memory in which the reaction time of a response to a stimulus is faster due to having previous experience of the stimulus. For example, individuals can recognise a word faster if it is paired with a related word. For example, “table” and “chair” and are slower at recognising words if they are unrelated. For example, “nurse” and “butter” (Refer to appendix A). In the study of Meyer and Schvaneveldt (1971) we refer to semantic priming, where semantic refers to the logic and language individuals store in their implicit memory. Meyer and Schvaneveldt (1971) suggested that the response times were faster when the string of words were related because a part of the memory is activated in which the retrieval of words from semantic memory is faster.

Associative priming and the lexical decision task is an extension of the study by Meyer and Schvaneveldt (1971). In their experiment they hypothesized that recognition of a word is faster when it is associated with its prime. In their experiment they presented 12 participants with two strings of letters, one above the other. Their task was to press the ‘yes’ key if the two strings were words and the ‘no’ key if one or both were non-words. In their results, they found that response times were faster when the target string consisted of related words such as ‘bread’ and ‘butter’ or ‘nurse’ and ‘doctor’ proving their hypothesis to be correct. Furthermore, another psychologist McNamara (1992) demonstrated results from a similar experiment of associative priming. He also found that mean response times were faster when words were related rather than unrelated. He also found that errors rates (percentage of errors) were much higher in the non-related condition than in the related condition. The present experiment is in favour of that of Meyer and Schvaneveldt’s (1971) and McNamara’s (1992) because similar results were obtained from both research studies. The difference to the present study to that of Meyer and Schvaneveldt’s (1971) was that the string of words were not presented simultaneously, they were presented one after the other and the participants were to decide if both strings were words or if one of both were non-words. From this it is unknown which letter string is the one that the participant responds to from their implicit memory. However the present experiment could allow us to investigate the effects of the time between the letters are presented on the effects on response times. Another modification that was made was that in the present experiment, was that one letter was changed in either a related or an unrelated word. This allows us to investigate the effects of the response time on the identification of words. Previous research by Meyer and Schvaneveldt (1971) and findings from the present experiment are relevant in explaining how the long term memory is organised in individuals.

The hypothesis of the experiment was that the response time in the lexical decision task would be significantly faster for target words related to the prime than for target words unrelated to the prime.

Participants

150 participants took part in the present experiment, all being part of the same educational institution. The participants consisted of mixed gender with majority being female. The age of the participants varied from 18 and above and they were also from various ethnicities. The type of sampling that was used in the experiment was opportunity sampling because the participants were available to take part in the experiment at that time.

Apparatus

As the experiment was computer based, each participant had access to a computer. The program was written in TCL for Windows, and ran on PCs using the Windows 7 operating system. There were two types of stimuli used in the experiment; words and non-words, related and unrelated words. Key findings in the experiment showed that participants responded faster when words were related to each other; for example, ‘bread’ and ‘butter.’ Thus, participants responded slower when words were unrelated to each other; for example, ‘table’ and ‘nurse.’

Design

The experiment was set up in an educational institution and the task was carried out using computer equipment. There were 150 participants involved in the experiment and consisted of first year psychology students. This was an opportunity sample, as the participants were available at the time the study was being carried out. The program was written in TCL for Windows, and ran on PCs using the Windows 7 operating system. The design of the experiment was repeated measures where participants were presented with all the stimuli. There were two independent variables for this experiment: related or unrelated words and word or non-words. The dependent variable was the response time in which participants decided if the string of letters were words or non-words. The conditions in the experiment were to press the ‘L’ key or the ‘A’ key if the string of letters were words or non-words. The experiment program showed the participant which key to press for each response. The controls that were used in the experiment were the ways in which the stimuli were presented. The target was presented 300, 600 or 900 milliseconds after the prime was presented on the computer screen. When the results were presented the times after the stimuli were presented were averaged out.

Procedure

The participants carried out the experiment simultaneously. The participant was seated in front of the computer throughout the experiment. The stimuli were presented one after the other on the screen after which the participant had to make the choice if the string of letters presented were words or non-words. The participant responded by pressing the ‘A’ or ‘L’ key on the keyboard; these were random for each participant. Response time was measured by the computer programme. The experiment lasted around 20 minutes. Participants were shown 18 trials in which the target string was a word related to the prime, 18 trials in which the target string was a non-word, 18 trials in which the target string was a non-word created from a word and 18 trials in which the target string was a non-word from a word unrelated to the prime, which means there were 72 trials in total. (Refer to appendix B).

Results

The inferential test chosen for the present experiment was a repeated measure t-test. Using a repeated measures t-test, with alpha set at 0.05, response times for the related words were significantly faster than the unrelated words (t=-3.762, df=149, p=0.000, two tailed). Similarly, using a repeated measures t-test, with alpha set at 0.05, response times were significantly faster for related non-words than for unrelated non-words (t=0.974, df=149, p=0.332, two tailed). The difference between the mean response times in each condition varied, the mean response time was faster when the target string was a word related to the prime and was slower when the target string was a pseudo word unrelated to the prime. McNamara (1992) reports that lexical decisions are made faster because semantic memory consists of interconnected nodes (Refer to appendix C). When a stimulus is presented, a part of the semantic memory is activated and the spreads across the network activating related nodes. The closer the nodes, the closer the words are associated in semantic memory.

Mean Reaction Times and Standard Deviations in Lexical Decision Task

Related Words

Unrelated Words

Related Non-Words

Unrelated Non-Words

Mean Response Time (ms)

838.0

902.4

1108.4

1088.5

Standard Deviation (ms)

313.4

331.3

464.6

433.6

Discussion

The hypothesis of the experiment was accepted because response time in the lexical decision task was significantly faster for target words related to the prime than for target words unrelated to the prime. This is in favour of Meyer and Schvaneveldt’s (1971) experiment as previously mentioned because in their experiment, their mean response time was fastest when the target string was a word related to the prime; similarly with the present experiment. The results of the present experiment demonstrates that Meyer and Schvaneveldt’s (1971) and McNamara’s (1992) theory of semantic priming is both valid and reliable.

An opportunity sample was used in the experiment and may be considered as a weak research method by many researchers. Therefore this kind of sample may produce a biased sample as it is easy for the researcher to gather participants from their social group; in this case, a sample of students from the same educational institution. Hence the results of the experiment are not generalizable to a whole population. Since majority of the participants involved in the experiment consisted of students, the results can be affected because a much educated, mature individual will have more knowledge of the world and whose memory may be better than that of students. Similarly, a younger, uneducated individual may not have a reasonable good memory so their results will differ from the student’s results. A more appropriate type of sample may be a volunteer sample, where the experiment will recruit all types of individuals from the society. This way the findings of the experiment may be generalizable to a total population.

References

Friedenberg, J. & Silverman, G. (2006). Cognitive Science: An Introduction to the Study of Mind. London: Sage Publications Ltd.

Goldstein, E. (2010). Cognitive Psychology: Connecting Mind, Research and Everyday Experience. (3^rd ed.). Belmont: Wadsworth Cengage Learning.

McNamara, T. (1992). Learning. Memory, and Cognition. Journal of Experimental Psychology. 18 (6) 1173-1190.

Meyer, D. & Schvaneveldt, R. (1971). Facilitation in Recognizing Pairs of Words. Journal of Experimental Psychology. 90 (2) 227-234.

Proctor, R. & Healy, A. (Eds.) (2003). Experimental Psychology, Volume 4 (2nd ed.). New Jersey: John Wiley & Sons, Inc.

Appendix
Appendix A

Stimuli used in the experiment. The words in the brackets indicate where the pseudo word originated from.

Related

Unrelated

Words

Non-words

Words

Non-words

Prime

Target

Prime

Target

Prime

Target

Prime

Target

1

cheddar

cheese

east

wost (west)

spill

trees

castle

entra (extra)

2

lobe

ear

rungs

ladler (ladder)

pride

chair

horse

sparole (sparkle)

3

arson

fire

adhere

stirk (stick)

finger

story

plate

offige (office)

4

rant

rave

bits

paeces (pieces)

going

gossip

desert

bortle (bottle)

5

aunt

uncle

meek

mald (mild)

twice

bread

eject

heage (hedge)

6

husband

wife

shove

pesh (push)

begin

reptile

flood

funrus (fungus)

7

loaf

bread

purse

maney (money)

never

mental

peace

skirny (skinny)

8

boys

girls

wrong

roght (right)

record

yellow

stock

anrow (arrow)

9

halt

stop

neat

tily (tidy)

number

staff

energy

scrool (school)

10

nape

neck

hong

keng (kong)

sleep

pocket

vacant

yingle (mingle)

11

stark

naked

assist

herp (help)

above

curry

wisdom

lutter (litter)

12

trove

treasure

rich

poom (poor)

juice

angry

pencil

rishes (dishes)

13

vehicle

car

glove

havd (hand)

light

crime

ghost

molern (modern)

14

core

apple

before

axter (after)

effort

cigar

doctor

chease (cheese)

15

nook

cranny

donor

bloot (blood)

apples

silent

roses

povice (police)

16

needles

pins

morse

cofe (code)

clever

window

bricks

wanler (wander)

17

hatter

mad

ache

paim (pain)

forget

modest

garden

steck (stuck)

18

flock

sheep

odour

smerl (smell)

health

garage

matter

cheld (child)

Appendix B

Number of trials per condition.

Related

Unrelated

Word

18

18

Non-word

18

18

Appendix C

A diagram of how knowledge in semantic memory may be organised.