Possible Sources of Error

Since I've just completed the phase of my research involving data collection, I'm ready to move onto data analysis and working out any conclusions I may find evidence for. However, before posting all of the graphs, raw data, and conclusions I'm starting to form, I'd like to bring up some possible sources of error.

Sample Size: 

About ten students showed up in each group. Simply put, ten students in each sample is on the small end. Ideally, I would have had fifty or so in each group, and the more the better. Having groups as small as mine were can still yield accurate results, but the chance of error is much higher than it would be for larger groups. 

Musical Preference: 

The musical preference of subjects in psychological experiments has been known to influence the effect of music on cognition. For this reason, I did include a question in the survey given to subjects on whether they enjoyed the music or found it distracting, but I'm not sure if I will be able to separate groups into further categorizations based on this question because of the already small sample size. 

Moving Rooms: 

Unfortunately, for one of my trials, I had to use a different room than I used for the other two. Although the rooms' builds were the same, one was the Mandarin room while the other was for French. While this may or may not have had any impact on the participants, it is still worth noting. Perhaps the students who regularly had a class in one classroom unconsciously felt slightly more comfortable in it than those who didn't. 

My Interactions with Subjects: 

While I'd like to think I said all of the same things and spent about the same amount of time each day explaining the instructions to each group, I must consider the possibility that my interactions with the subjects influenced their moods or behavior. For example, I might have seemed more excited and encouraging on the day when I expected subjects to do best. 

Background Noise: 

During one of the silent periods, a teacher walked into the adjacent room, making some noise for about a minute. It wasn't loud, but it might have been a bit distracting to students who were completely focused on their task beforehand. 

The Group Dynamic:

Because I tested subjects in groups rather than as individuals, I should consider what effects this might have had. For example, students might have already felt motivated and competitive about finding as many words as possible simply because they were in groups. This is one of my main concerns about my procedure. 

So, as you can see, there are many, many different possible sources of error. When making my conclusions, the best thing I can do about this now is to give each some thought and discussion. The results of my study could be an accurate reflection of the independent variables I changed for each group, but if not, these sources of error are some possible reasons why. 

Conducting My Experiment

My own experimental trials will involve the help of thirty-three helpful ninth-grade volunteers from BASIS Flagstaff! Over the span of three days, I will be conducting my study and collecting data. I've randomly split the volunteers into three groups of eleven with roughly even amounts of males and females in each group. Each group will come in during their lunch for half an hour to participate in the research. Here's what I'll be doing.

The Three Groups:

Group 1 - Priming with Music

This group will only be primed with music. I've chosen Vivaldi's Spring to be used for the purposes of this experiment. Similarly to John Bargh's landmark study (The Automated Will: Nonconscious Activation and Pursuit of Behavioral Goals) the participants will be asked to solve a word search with thirteen "neutral," non-priming words in it. (For example, "turtle," or "building.") They will complete this task while listening to Vivaldi. Next, they will be given ten minutes to find as many words possible in three following word searches with the themes "colors," "foods," and "animals." Lastly, they will be given a short survey. The last two tasks are what I will be analyzing: specifically the number of words found on the three word searches and their answers to the survey questions. 

Group 2 - Priming with Words

This group will only be primed with words. The subjects will complete a word search with thirteen words on it, several of which are aimed at priming for motivation and high-performance. Some of these words include "strive," "succeed," and "achieve." It was necessary to give Group 1 a similar word search with neutral words in order to keep as much of the experimental procedure between groups as constant as possible. Next, they will complete the same two tasks completed by Group 1 following the initial word search. 

Group 3 - Priming with Music and Words

This group will be primed with both words and music. Participants will complete the same word search done by Group 2, except in this group, subjects will listen to Vivaldi while doing it. As in Group 1 and Group 2, they will proceed to work on the three themed word searches and the survey. 

After collecting all of my data, I will have to analyze it and look for differences in results. If you are interested in taking a look at the word searches and survey I'm using, they are attached below.

Materials: Three Word Searches, Neutral WS and Survey, Prime WS and Survey

Important Testing Considerations

When running tests of any kind, scientists must try to limit the variation between different sample groups to the independent variable(s) they are hoping to test. In psychology, this is especially difficult because even the slightest differences in environment can affect results. Furthermore, inevitable differences in preferences, interests, and general mental processes between individuals bring yet another challenge.

So, what do psychologists do to ensure their tests are as controlled as possible? There are some things nearly all experiments take into consideration.

Sex:

Scientists often like to split their sample groups randomly, but with roughly equal proportions of the male to female sexes in each group. If they do not, they may consider this in their statistical calculations when analyzing their data. For the most part, gender does not have an influence on results. However, it is worthwhile to consider that the male and female brains are different in several ways. For example, males tend to be better at spatial processing and focusing on a single task while females tend to be better at communication, empathy, and multitasking. Read more here.

Age:

Age is also an important factor in psychology simply due to the way our brains change as we grow older. Scientists have theorized that cognitive decline may begin as early as our 20s or 30s! Read more here. When their experiments do not have to do with differences in age groups, psychologists try to test a single age group. For example, many studies use college students as test subjects. 

Sample Size:

Researchers must balance having a sample size large enough to obtain accurate results from with one that is manageable considering time and resource restraints. Published studies usually use anywhere from about thirty to several hundred subjects per group. 

Other Things to Control:

The simple answer? Everything you can. 

If researchers are interacting with subjects, what they say to each sample group should ideally be scripted. Psychologists should take into consideration the color of the rooms they use, whether they test subjects individually or in groups, any background noise, the time of day, even the smell of a room or their own moods. In fact, many researchers use volunteers who aren't aware of what their research is to conduct the testing and input the data. 

In my research, I will try to keep the sample size, the proportion of males and females, the room, the scripting, the age of participants, and the time of day constant.