1) What is the independent variable? What is our Dependent Variable?

Transcription

1 1) What is the independent variable? What is our Dependent Variable? Independent Variable: Whether the font color and word name are the same or different. (Congruency) Dependent Variable: The amount of time it takes for the participant to indicate the color. 2) What is an appropriate set of hypotheses for this task? What kind of statistical test do you expect to perform? Justify your choices. H0 : There would be a lower mean time for the incongruent tests than the congruent tests or no noticeable difference in time duration between the congruent and incongruent tests HA: There would be a noticeable increase in time duration between the congruent and incongruent tests. The experiment should be a single-tailed type scenario as there the question does mention...does incongruency increase response times as well as the incongruent times were predominantly larger than the congruent times. µ1 = the mean of time spent for the participant group of congruent test µ2 = the mean of time spent for the participant group of incongruent test The null hypothesis (H0): µ1 - µ2 0, the time duration difference of the congruent and incongruent tests would equal 0 or be greater than 0, meaning the incongruent times would be less or equal to the congruent times for the population. The alternative hypothesis (HA): µ1 - µ2 < 0, the time duration difference of the congruent and incongruent tests would be less than 0, meaning the incongruent times would be greater than the congruent test times for the population The recommended statistical test we should use for the Stroop experiment is a T-test as this we don t know the mean of the total population however we have detailed data on a sampled group for both the congruent and incongruent tests. In addition, we only have a data set of 24 people in the sample making it more better to use T-test to calculate our statistical test, the type of T-Test selected is: Dependent: I chose to go the dependent t-test route as even though we do two tests (congruent and incongruent) it is with the same group of people that was tested on the congruent words then tested on the incongruent words. Paired: I chose to go the paired t-test because the participants were measured at two time points (congruent and incongruent word sets).

2 3) Report some descriptive statistics regarding this dataset. Include at least one measure of central tendency and at least one measure of variability. Central Tendency: Going into the data set of the Stroop experiment determining the means of the congruent recordings (designated xx c) and the incongruent (designated xx i) xx c = xx i = The median of the congruent group (designated as Cmed) and the incongruent group (Designated as Imed) Cmed = Imed = Considering all data on the congruent and incongruent test groups have unique time no same time as the data is continuous there is no mode available on both test groups. Measure of Variability: To get a better understanding of the general variability of the data sets (congruent and incongruent) we can find the IQR which accounts for 50% of the data for the perspective set of data. First arrange the data values in order of least to most and then find the values of the 1 st quartile and the 3 rd quartile: C q1 = I q1 = C q3 = I q3= After finding the values of the first and third quartile we can determine the inner quartile range (IQR) for each data set. C IQR = I IQR =5.335 Considering the variability of each test group we see when ordering the values and finding the range we get (Crange designated as range of congruent sample and Irange as range of incongruent sample): Crange = Irange = Average Deviation for the two test groups (congruent and incongruent) can be calculated with the Known mean of xx c = and xx i = using the equation:

3 We get: Cad = 2.85 and Iad= 3.40 Next measure we want to calculate is the measurement Variance with the data sets given (Congruent and Incongruent Groups) To find the Variance of the data sets however we first need to determine the Square of Sums so we can have a simpler Variance equation. Using the means of each test we can determine the Square of Sums of each the congruent test data and the incongruent. (Using: nn ii=24 (xx ii xx ) 2 ) SSc = SSi = Given the SS (Square of Sums) we can use this to determine the variance of the data set for each test (congruent and incongruent). Since we are dealing with a sample we need to use the equation with the denominator of n-1, n being the number of participants on a given test (n = 24): VVVVVV = SSSS nn 1 Var(congruent) = Var(incongruent) = Given we now know the Variance for each of the data sets we can use that knowledge to get the sample standard deviation for each of the test groups (congruent and incongruent). VVVVVVVVVVVVVVVV Sample SD (congruent) = Sample SD (incongruent) = ) Provide one or two visualizations that show the distribution of the sample data. Write one or two sentences noting what you observe about the plot or plots.

4 Since we have determined the IQR and Range for each data set (congruent and incongruent) the first visualization we can use is a box plot that shows us a graphical representation of the range of each set at a glance, summarizing and allowing us to compare the two test s data and the distribution of that data and any indications of symmetry within that data. However, since exact values are not retained we are unable to have a thorough analysis in the data, the box plot is simply a quick top down view of the data from the congruent and incongruent tests. In the data we see that there is a significant increase in duration of time taken to finish the test from the congruent test versus the incongruent test as 75% of the data in the incongruent test is above the IQR of the congruent test. That being said about 25% of the data in the congruent test looks like it falls in about 50% to 75% of the incongruent test so there is quite a bit of overlap when considering all data but there is some tendency to believe that incongruent test on average from this plot that summarizes the data will take longer to finish than the congruent test.

5 Another means to visualize the data can be done through a histogram and normalized curve combination which categorizes the data into frequencies to see how the data falls proportionally within ranges and gives us an approximation of the shape of the data s distribution. The congruent histogram with normal curve shows a mild right skew (mainly because there are a few data points that are beyond 20 seconds) in the recorded times which is more visible through the normalized curve as sizes of the bins can misrepresent the shape of the distribution if using histogram alone. The congruent histogram if ignoring the tail beyond 20 seconds resembles closely to a normal distribution curve. The congruent histogram shows mainly the majority of the recorded times for the congruent test falls between 10 to 20 seconds. The incongruent histogram with normal curve shows a firmer right-skewed distribution with connected data showing the majority of the times recorded falls between 15 to 30 seconds. 5) Now, perform the statistical test and report your results. What is your confidence level and your critical statistic value? Do you reject the null hypothesis or fail to reject it? Come to a conclusion in terms of the experiment task. Did the results match up with your expectations?

6 As determined in question 2 we will calculate our data using the T-Test statistical test for the Stroop Effect experiment. Considering the group took both the congruent and incongruent tests this T-Test will be considered dependent and paired. Knowing the means of each test (congruent and incongruent, from question 3): xx c = xx i = We can calculate quickly a point estimate using: xx c - xx i xx c - xx i = = A point estimate is handy to know as it gives us our Best guess on the difference of the population between the two tests (population difference estimate of congruent versus incongruent times). Calculating the difference of the time points for each person (congruent minus incongruent) and averaging the differences should lead to the same value as the point estimate we calculated earlier. Using the average of the differences which should match the point estimate of we subtract that value from the differences of each person and then square it: (Difference of Congruent and Incongruent Average of differences) 2 Having this set of squared deviations from the mean we can sum them up into Square of Sums Using ( (xx ii xx ) 22 nn ii=2222 ) Square of Sums of difference of sample means = Given the SS (Square of Sums) we can use this to determine the variance of the difference. Since we are dealing with a sample we need to use the equation with the denominator of n-1, n being the number of participants on the given tests (n = 24): VVVVVVVVVVVVVVVV oooo dddddddddddddddddddd = SSSSSSSSSSSS oooo SSSSSSSS oooo dddddddddddddddddddd oooo ssssssssssss mmmmmmmmmm nn 1 Variance of difference = Given we now know the Variance for the difference we can use that knowledge to get the Standard Deviation of Differences. VVVVVVVVVVVVVVVV oooo dddddddddddddddddddd SD of Differences = 4.865

7 If we designate SD of Differences as s (SD of Differences = s) we can now calculate the t-statistic for the difference using the equation: T-statistic = xx CC xx ii ss/ nn Given the numerator is the point estimate or average of the differences (both equal the same value), s was determined a step earlier and n is the number of participants in the Stroop experiment we can calculate the T-Statistic: / 2222 = Since we have the T-Statistic we can compare it to the T-Critical value to determine if the null hypothesis should be accepted or rejected, using what we know (24 participants, which means n=24 we can determine the Degrees of Freedom) Knowing n=24, the degrees of freedom for this t-test = n-1 = 23. Since we are aiming to find with 99% Confidence level for a one tail our α = Given df = 23 and α = 99% Confidence Level: need to be between Given that the T-Statistic found was it leads us to reject the null hypothesis due to T-Statistic being in the critical area of tested at 99% Confidence Level. The t-test confirms what was expected that incongruent tests would normally take longer than congruent tests as congruent test linked both visual colors with the correct printed name of the color while incongruent required to try to separate the visual color and the name of the color which required a bit more time to do. 6) Optional: What do you think is responsible for the effects observed? Can you think of an alternative or similar task that would result in a similar effect? Some research about the problem will be helpful for thinking about these two questions! When finding about the Stroop effect I found a background on the topic from the Rochester Institute of Technology which goes into detail of the Stroop Effect mentioning, Subjects averaged 74% longer to name ink colors of incongruent words. 1 which matches up with the observation with the boxplots. However more importantly we see the article mentions that the two studies led Stroop to conclude that since people are more practiced at word reading than naming colors, there is less 1

8 interference with word reading than with color naming. 2 Which basically comes down to is that we have less effort in making meaning to words (reading) in our brains than we do naming colors. Another similar task that could result in a similar result is the idea that is mentioned by Raymond De Young and Associate Professor of Environmental Psychology and Planning at the University of Michigan, School of Natural resources and Environment that states, Most humans are so proficient at reading, at perceiving whole words, that they do not easily notice the individual letters. This is why proofreading is so hard to do. This tendency to quickly perceive words is used in testing for the Stroop effect 3 The task of proofreading is similar task as the tests we did as it presents a similar Stroop Effect