Other support for the matching hypothesis also comes from Murstein’s study (1972). In the study photographs of real couples (who were either engaged or in a stable relationship) were rated as more similar in physical attractiveness than for random couples created for comparison. This finding provides support for the Matching hypothesis as it shows that individuals in a real couple are more similar in physical attractiveness than random couples. It also shows that the matching hypothesis may be applied to long term relationships as stable relationships or engagements signal a successful relationship. These findings show evident support for the matching hypothesis.
Formulation of Aims
My aim is to carry out a modification of Murstein’s (1972) study into the matching hypothesis and also to see whether it is a feasible theory when it comes to explaining interpersonal attraction. I will modify the original research piece by asking participants to rate the physical attractiveness of the opposite sex in a questionnaire. I expect to find similar findings to those found by Murstein in the real couples being similar in physical attractiveness. If the ratings are similar between real couples then this would provide support for the matching hypothesis.
Alternative Hypothesis
There will be a statistically significant, positive correlation between the ratings of physical attractiveness for real couples.
Null Hypothesis
There will not be a statistically significant, positive correlation between the ratings of physical attractiveness for real couples.
Method
Three A level students carried out a study by asking participants from an opportunity sample at a local shopping centre to fill out a questionnaire rating the physical attractiveness of 20 males (see Appendix I) and 20 females (see Appendix II). The sample used was an opportunity as participants that were available at the time in the shopping centre were used. The questionnaires that were given to the participants were created by using pictures from dating websites (see reference). Two questionnaires were created; One with 20 males (10 single males and 10 coupled males) and another with 20 females (10 single females and 10 coupled females). The participants were asked to rate each of the people on the questionnaire on their physical attractiveness with a score between 1 and 10 (1 being the lowest and 10 being the highest). With the results collected we were able to carry out a correlational analysis.
Apparatus/Materials
Using the internet and certain dating websites we gathered a collection of pictures of real life couples and also pictures of single males and females. When collecting the pictures we had a certain criteria. The individuals in the picture had to be white American or European and also be within an age range of 18-. For the questionnaire we wanted to create we needed a total of 20 females (10 single females and 10 from the real couples) and also 20 males (10 single males and 10 from the real couples that the 10 females came from). When the images were collated we firstly labelled the 10 real couples, letters from A to J (see Appendix III). We then split the image of the couples so we ended up with 10 males and 10 females. Then we added the 10 males to a questionnaire and also the 10 females to a separate questionnaire. Following this using certain dating websites we collated images of 10 single men and 10 single women. With the pictures of the single males and single females turned upside down and mixed up we selected one male and then one female, thus creating a random couple. We done this until all the pictures were finished and were left with 10 randomly created couples from the 20 single males and females. These couples were then lettered K to T (see Appendix IV). Once the random couples had been numbered the single men were added to the questionnaire with the real coupled men and the single women were added to the questionnaire with the coupled women, thus producing 2 questionnaire/photo sheets; One with 10 single males and 10 males from a couple and another with 10 single females and 10 females from a couple.
Procedure
We approached a potential participant and followed standardised instructions (see Appendix V) by asking them whether they would like to take part in the study we are carrying out. If they consented they were handed a questionnaire with pictures of the opposite sex in them. Before filling it out they were reminded not to write their name on the questionnaire and also their right to withdraw at any time. Upon completion of the questionnaire they handed it to us. We then fully debriefed the participant and gave them the opportunity to withdraw their data if they wished to.
Controls
Race: Race will be controlled in the selection of pictures. By ensuring that all the pictures chosen are from one race it eliminates the possibility of race bias. The race that we will be choosing will be white American/European. As there will only be one race to rate, in the event of an individual being racist then all of the pictures will be rated the same as they are all white. Also of the white people in the pictures the most attractive people may still be rated higher than others.
Researcher bias: If the questions were asked in person then as a researcher I could influence the rating or score given by the participants. To overcome this problem a closed questionnaire will be given to the participants which will be filled out in privacy.
Gender bias: Some males may feel uncomfortable rating other males’ attractiveness and so might either rate them untruthfully or choose not to rate them altogether. To overcome this we will be asking only males to rate the female pictures. A similar problem could occur with females as they have been shown to be very critical and so females will only be rating the male pictures.
Ethics
There are several ethical issues which will arise in the study I am conducting. They are listed below with ways in which they will be dealt with.
Right to withdraw: Ensuring that all the participants know that they can choose to withdraw from the study at any time if they wish is very important. They will be told at the start they are free to stop participating and also withdraw their data at any time.
Psychological harm: This could be an issue if the participants feel that they are ridiculed or judged on their rating of the pictures. When carrying out the study we will be making sure the risk of harm is no greater than the day to day risk. We will ensure that they cannot feel as if they are being judged on the ratings they give by giving them privacy and also asking them not to place their names on the questionnaire so that everything remains anonymous. Debriefing the participants after the study will also minimise any issues that arise.
Lack of Informed consent: As the participants are being deceived it is not possible to gage fully informed consent as this would mean that their agreement should be based on the full knowledge of the nature and purpose of the study. Instead of fully informed consent, partially informed consent will be gained as the participants will receive some information about the study and the field we are investigating upon being approached. As the participants are being deceived, debriefing the participants on the true nature and purpose of our study after they have participated and also offering them the opportunity to remove their data from the results are both very important. These will help deal with the issue of informed consent
Privacy: The privacy of the individuals in the photos could be an ethical issue as they do not know that they are being rated on their physical attractiveness. As the images were downloaded from a public domain (internet) meaning that when uploaded by the individual they knew that they would be open for anybody to look at and make their own judgements. As this is the case further action to deal with the issue will not be taken.
Results
Below are 2 tables that I was able to formulate from the results of my questionnaires (see Appendices VI & VII). I have an average score for every male in each of the couples and a corresponding average score for the females. In order for me to be able to work out the average scores I added all the scores for one individual from a couple and divided the number by 20 (the number of participants who rated the male or female). For example in order for me to get the average score for the male in the real couple A, I added all of the scores he received and divided that number by 20 (the number of participants who rated him). I repeated this for the males and females in the real and fake couples. By looking at the results for my real couples the average difference between 6 of the couples was less than 1.0 and the difference between 3 of them was less than 0.5. When compared to the random couples only 2 of them have an average difference below 1.0. This shows some support for the matching hypothesis and the hypothesis I made. I can further investigate the correlation by creating scatter graphs that will show the trend line of the relationships.
Scatter Graphs
Using the average scores I was able to draw two graphs using Microsoft Excel to see whether a visible correlation between the scores exists. My graph showing the relationship for real couples clearly shows a positive correlation as the trend line is angled from the bottom left of the graph to the top right and the majority of the points lie in the 2nd and 4th quadrant. This positive correlation provides support for the matching hypothesis and for my hypothesis. The extent to which it supports my hypothesis will be examined by running a statistical test on the data. I have circled 2 rogue results in the graph as I believe without them the trend line would be more positive.
I have also created a similar graph for the random couples. This graph shows there is a slightly negative correlation between the 2 sets of average scores. Again I have circled 2 rogue results as I believe without them the correlation would be more positive.
Inferential Statistics
In order for me to work out the correlation of the relationship between the average scores of the males and females in the real couples I used Spearman’s rank correlation test to calculate the correlation (see Appendix VIII). This test will show whether the correlation is significant enough for me to be able to accept my hypothesis. I decided to use it as firstly it is appropriate to use if a hypothesis predicts a correlation between 2 variables which mine does. Secondly it is suitable for my hypothesis as the males and females both came from one couple and therefore were related. For example the 2 sets of scores come from 1 couple (thing). A third reason for choosing Spearman’s rank correlation test was because the data I have is ordinal and therefore able to be analysed using Spearman’s.
The critical value of rho at a 5% level for a one tailed hypothesis with 10 pieces of data is 0.564. The observed value of rho for my data from the real couples equalled 0.327. Although my graph does show a positive correlation and the rho also shows one, the number does not equal or exceed 0.564 therefore is not significant enough to prove the matching hypothesis and also prove my hypothesis. For that reason the correlation is insignificant and I have decided to reject my alternative hypothesis and accept the null.
For comparison I also calculated the rho for the data from the random couples (see Appendix VIII). The rho was -0.02424. This figure does provide some support for the matching hypothesis as it shows that real couples are more positively correlated than random couples and could suggest it’s because people are attracted to and form relationships with people who are similar in physical attractiveness.
Discussion
Explanation of Results
By analysing my results it is clear that a positive correlation exists between the average ratings of the males and females in my real couples. However the correlation is not strong enough to be considered significant. I believe a correlation of 0.564 was not achieved due to a number of reasons, the main one being rogue results. I have circled 2 results on the graph which I believe to be rogue. I circled them as they lie quite far from the trend line and also because when looking at the scores there is a considerable difference between them. For example the point circled in red represents couple G where the male received an average rating of 2.4 and the female received an average rating of 5.0. The point circled in purple represents couple J where the male received an average rating of 2.4 and the female received an average rating of 4.7.
In the appendix is a graph I have drawn excluding the rogue results (see Appendix IX). The exclusion of the points from couples G and J produces a more positively correlated trend line which I believe will change the rho. This shows that the rogue results skewed the trend line negatively and I believe this is the main reason for not achieving a statistically significant correlation.
I calculated the rho again by excluding the average scores of couple G and J (see Appendix X) to see whether a statistically significant, positive correlation would be observed without the anomalies. The observed value of rho without couple G and J is 0.685. As noted earlier in my inferential statistics the critical value for a correlation to be deemed significant is 0.564. The value of rho without couple G and J increased substantially from 0.327 (the original rho) to 0.685 – more than double. This shows that couples G and J were definite anomalies and also that without these anomalies a statistically significant, positive correlation would be achieved and that would allow me to reject my null and accept my alternative hypothesis (also providing evidence for the matching hypothesis). However this is not possible as I used 10 couples and not 8 so therefore I cannot change my rejection to an acceptance.
Furthermore, as the graph for my randomly allocated couples shows a strong negative correlation and also has a rho of -0.0242, support may be provided for the matching hypothesis as it shows that real couples have a positive correlation between their attractiveness, whereas randomly allocated couples vary in physical attractiveness and therefore don’t have a positive correlation. I have circled the rogue results on the graph (see page 8). These 2 rogue results are due to the random allocation of partners in my method,
Relationship to background research
The findings of my study are similar and dissimilar to my background research. My findings are similar to the first ‘computer dance’ study carried out by Walster et al. (1966) as neither there’s nor mine supported the matching hypothesis. However my findings are different from the repeat of the study by Walster et al. (1969) as their research provided support for the matching hypothesis. I believe this could be down to the fact that their sample was of only university students whereas my sample was the general public.
Murstein (1972) found support for the matching hypothesis, however my modification of his study did not yield the same findings as my correlation was not significant enough. This could be for a number of reasons. The main reason I believe to have caused this difference in findings is time. Mursteins research was carried out 36 years ago which may mean the matching hypothesis is not of great relevance today. Values in the past regarding the formulation of relationships may have been placed solely on similar physical attractiveness however in today’s society values are more placed on similar personalities, attitudes and mentalities. However the calculation of rho for the real couples was 0.327 whereas the rho for the fake couples was -0.242. This difference in correlation therefore does provide some support for Murstein’s findings and in turn the matching hypothesis as it shows that real couples are more similar in physical attractiveness than fake couples are. This could show that similar physical attractiveness is still a deciding factor in forming relationships but not as much as it used to be in the past.
Limitations & modification
A limitation of my research could be the time and/or the location of where it was carried out. When handing out my questionnaires I may have been more prone to approaching a certain type of person as there were more of them in the area I was working within, thus producing biased results meaning that my results were not typical of the population due to the sample. For example in the day time on weekdays certain types of people are more likely to be at a shopping centre (mothers, pensioners) however on weekends a wider range of people may be at the shopping centre. To ensure this wouldn’t happen again if were to repeat the research, I would carry it out at more than one shopping centre over the course of a week and at different times of the day thus collecting a larger and more varied sample
Another limitation of my research and a possible reason why I did not achieve a statistically significant correlation could be a result of the images used. Whilst finding the pictures of real couples we used many websites and their ‘success stories’ section. However no actual proof or real data was displayed. We had to believe that the couples were genuine. However there is a possibility that the dating agency/company could have hired 2 random people or models to pose as a real couple in order to attract potential customers as after all it is a business. This would mean that there was not an actual romantic relationship formed between some of the couples and this could offset some results (possibly couple G and J). When repeating the study ensuring that all actual couples were genuinely in a romantic relationship would be vital to rule out the possibility of any doctored couples. This could be done by taking pictures of real married couples that I may know and gaining there fully informed consent and release of their pictures so that it is ethically correct.
Implications and suggestions for future research
As my findings did not support the matching hypothesis due to the correlation not being high enough to be deemed significant I will carry out research into another reason why people may attract, Proximity. I could test proximity in 2 ways, the distance individuals in a real couple live from each other and how that played a role in the forming of a relationship or I could test relationships where the individuals in a couple work in the same company or organisation. I would test this by using an open questionnaire or an unstructured interview so that the answers I receive are revealing and rich.
I could also carry out research into the sociobilogical theory. I could do this by asking participants to list characteristics/assets they would want an ideal partner to have. This could be done in an open questionnaire or in an unstructured interview. The results of what a female desires in a man and what a male desires in a female will be compared with each other and the traits that are deemed attractive by the sociobilogical theory.
References
Cardwell, M., Clark, L. and Meldrum, C. (2004), Psychology for A2 level, Harper Collins, pp. 36-41
Cardwell, M. and Flanagan, C. (2004), Psychology A2- The Complete Companion, Nelson Thornes, pp. 2-9
Flanagan, C. (2005), Research Methods for AQA 'A' Psychology, Nelson Thornes, pp. 79-98
Murstein. (1972) in Cardwell, M., Clark, L. and Meldrum C. (2004), Psychology for A2 level, Harper Collins, p. 41
Silverman, I. (1971) in Silverman, I. (1971), Physical attractiveness and courtship - Sexual Behaviour, pp. 22-25
Singh, D. (1993) in Carver, S., Dovidio, J. and Judd, C. (1993), Journal of Personality and Social Psychology: Vol. 65, pp. 283-288
Walster et al. (1966) in Cardwell, M., Clark, L. and Meldrum, C. (2004), Psychology for A2 level, Harper Collins, pp. 38, 41
Walster and Walster. (1969) in Cardwell, M., Clark, L. and Meldrum, C. (2004), Psychology for A2 level, Harper Collins, p, 41
Website References
www.match.com, 23/06/08, 14.30
www.blogs.kansascity.com, 23/06/08, 14.50
www.dallas.iblp.org, 23/06/08, 15.15
www.wrsphotography.com, 23/06/08, 15.45
www.meetyoulater.com, 23/06/08, 18.00-22.00
Appendix Contents
Appendix I – Male Questionnaire
Appendix II – Female Questionnaire
Appendix III – Original Couples
Appendix IV – Random Couples
Appendix V – Standardised Instructions
Appendices VI & VII – Results Tables
Appendix VIII – Calculation of Rho
Appendix XI – Scatter graph without couples G & J
Appendix X – Calculation of Rho without Couples G & J
Original Couples
A F
B G
C H
D I
E J
Random Couples
K P
L Q
M R
N S
O T
Standardised Instructions
- With clipboard and pen in hand approach a potential participant.
- Attempt to gain potential participants attention by saying, “excuse me”.
- Interviewer introduces themselves, “Hello, my name is “state name”, I am a student from Preston Manor High School carrying out research for A2 coursework”.
- Ask the participant if they are over the age of sixteen by saying, “may I ask if you are over the age of sixteen”?
- If participant answer is “no” or any other similar word then explain that they are too young to participate in this research. “I’m truly sorry but as you are under the age of sixteen you are unable to participate due to a lack of parental consent”. If the participant replies “yes” or any other similar word then inform the participant of what they will need to do by saying, “could you please take a moment of your time to complete this questionnaire”?
- If the participant refuses to complete the questionnaire then interviewer says, “thank you for your time”. If the participant agrees to complete the questionnaire then the interviewer says, “thank you” and hands the questionnaire to the participant along with a pen.
- Once the participant is holding the questionnaire and pen, the interviewer should reiterate, “Could you please rate these males/females according to their levels of attractiveness on a scale of one to ten.” “One represents the least attractive and ten represents the most attractive”. “Remember that stating your nationality is optional and you may wish to stop participating at any time”.
- Once the questionnaire is complete the interviewer should remove the questionnaire from the participant and say, “thank you very much”.
- The interviewer should then inform the participant of the true nature of the research. “I would like to inform you that the true nature of the research is to investigate whether genuine couples have similar ratings of physical attractiveness in comparison to random couples”.
- The interviewer should then offer the participant the option to withdraw their data. “Now that you are aware of the true nature of the research, would you like to withdraw your data”?
- If the participant would like to withdraw their data then the interviewer should find the nearest bin and place the questionnaire firmly inside it. If the participant does not wish to withdraw their data then the interviewer should keep the questionnaire at the top of the pile of completed questionnaires.
- The interviewer should then ensure the participant that, “your questionnaire will not be used for any other purpose outside of this research and you have my assurance that these findings will not be published”.
- The interviewer should then say, “goodbye, and thanks again for your time”.