Generation Effect Experiment – Data Analyses

Generation Effect Experiment – Data Analyses

This page presents all of the questions that you must answer for this assignment using Microsoft Excel. Follow each step carefully and input your answers on the excel page where indicated.

You must use Microsoft Excel for this assignment. If you submit a file that has used any other software, you will get zero.

The excel file contains the raw data from the Generation Effect experiment. This data includes each participant’s scores for the spoken, written, and typed conditions as well as some demographic information.

Descriptive statistics

1. Calculate the average age of your sample. Use the AVERAGE function and insert it in Cell C33. Use the data shown in Cells B2 to B31.

If you do not know how to do this, start by watching the video that I made shown on the “Data analyses using Microsoft Excel” page of the lab. Otherwise, you will find Youtube very helpful here. Search for “Excel function average”.

2. Calculate the number of reported gender categories for this experiment. Execute this step using the function COUNTIF. For example, to count the number of individuals who reported “Male”, enter the cells C2 to C31 in the “Range” field. Then, enter “Male” in the “Criteria” field. Repeat the analogous steps to count “Women”, “Prefer not to say”, and “transgender”. Your answers need to appear in Cells C36, C37, C38, and C39, respectively. Finally, in Cell C40, calculate the total of counted individuals belonging to these four categories. To do so, use the SUM function. If this number is not equal to 30, you will know that something is wrong with your previous calculations.
3. When the participants generated words during the experiment, certain words were expected. For example, when prompted with the word MARGARINE, a participant would be expected to generate BUTTER and not BARBER. Obviously, one cannot remember a word that they did not generate. So, you need to calculate how many such errors were committed.

Calculate the average number of generation errors per participant in your sample in the Typed condition. Use the AVERAGE function and insert it in Cell G33. Use the data shown in Cells D2 to D31.

4. Next, calculate this error as a percentage. To do so, click Cell G34 and insert a formula into the Excel formula bar. Formulas always start with the equal sign “=”. Use it to divide the content of G33 by 40, which is the total number of words that participants were asked to generate in the experiment. Does this number seem reasonable? How does it compare to the one reported in Slamecka and Graf (1978) on page 594? You do not need to write these answers anywhere but think about it.

The data shown in columns E, F, and G presents each participant’s recognition probability for the typed, written, and spoken conditions, respectively. For example, Participant 12 remembered all the words that she wrote (i.e., the probability was 1.0) whereas Participant 15 only remembered half of the words that he spoke (i.e., the probability was .5).

5. Calculate the probability of recalling words per condition. In Cell J3, use the AVERAGE function to calculate the probability of recall for the typed condition. Repeat this step for the written and spoken conditions. Insert your work in Cells K3 and L3, respectively.
6. Calculate the Standard Deviation (SD) of the probability of recalling words per condition. In Cell J4, use the STDEV.S function. Repeat this step for the written and spoken conditions. Insert your work in Cells K4 and L4, respectively.

Note. A standard deviation is a measure of data dispersion. It says how spread out data are around the mean, and more specifically, how far away from the mean each data point is on average. A small SD means that most data are close to the mean (everyone got almost the same result) and a large SD means that data can be pretty far away from the mean (everyone’s results are quite varied). You need to calculate these SDs to complete the next step.

7. Calculate the 95% confidence interval of the probability of recalling words per condition. In Cell J5, use the CONFIDENCE.NORM function. Repeat this step for the written and spoken conditions. Insert your work in Cells K5 and L5, respectively.

Hint. The alpha is the significance level used to compute the confidence level. Set it at .05. This will mean that 95% of the time, the mean of the population (if it were tested) would fall within the boundaries set by the confidence interval. For example, if a proportion for a condition is .8 and the confidence interval is .1, then there is a 19 out of 20 chance that the population mean for that condition will be between .7 and .9. Also, you have just calculated the standard deviation for each condition and “size” refers to the number of participants in the sample. Use this information to correctly fill out all the fields for the CONFIDENCE.NORM function.

Constructing an APA style column chart

8a. Select cells J2 to L2 and J3 to L3. Insert a “clustered column” chart.

8b. Resize the graph to fit the area described by the cells I8 to L8 and I23 to L23.

8c. Select the graph and click the “+” sign that you can see. It brings up “chart elements”. Select “Axis titles” and unselect “Gridlines”.

8d. Change the Chart title to Mean recognition probabilities for each condition. Change the Axis title to conditions. Change the Axis title to Recognition probability.

8e. Double-click the y-axis. In the Axis Options, set the minimum to 0.5 and the maximum to 1.0. Also click “Fill & line” and select solid line.

8f. Double-click the x-axis. click “Fill & line” and select solid line.

8g. Click one of the bars and right-click it. Set both the fill and the outline to the color “black”.

8h. Select the outline of the graph and right-click it. Click font in the menu. Set font color to “black” and size to 12.

8i. Insert error bars. To do so, double-click one of the bars in the graph.  Next, click “Add Chart Element”. This will open a menu. Now click Error Bars and More error bars options. Now click Custom and Specific Value. This will open a “Custom Error Bars” box. Click the “Positive Error Bars” text box, delete its contents, and select Cells J4 to L4. Repeat this for the “Negative Error Bars”. Finally, click “OK”. If you did all the steps correctly, you should see a short line on each bar.

8j. Select the “Series 1” box that appeared and delete it. This step removes the legend from the graph if one appears.

Great! You now have an APA style graph that is ready for the paper that you will be writing.

Inferential statistics

9. Conduct a t-test to see if there was a significant difference between recognition for the typed and written words. Insert the function T.TEST in cell L27. Arrays 1 and 2 require data. So, select all the typed data for Array 1 and all the written data for Array 2. You expect that recognition for the written condition will be superior to recognition for the typed condition. Thus, input “1” in the Tails field. Otherwise, the design of this experiment was within-subjects because all participants completed all the conditions. So select “1” for the Type field (i.e., paired).

Is the t-test that you conducted significant (i.e., is the p-value that you found smaller than .05)? You do not have to write your answer anywhere, but think about it. You will need to present it in your result section.

Now, conduct two more t-tests. One to see if there is a statistically significant difference between the typed and spoken conditions and another to see if there is a statistically significant difference between the written and spoken conditions. This work must appear in Cells L28 and L29, respectively.

When you report inferential statistics in an APA style research report, you must present three key pieces of information: The value of the tested statistic, the degrees of freedom, and the statistical significance. In a manuscript, for example, these three pieces would be presented as follows for a t-test: t(20) = 2.51, p = .17, where t identifies the test, the number in parentheses identifies the degrees of freedom, the number that follows the first equals sign shows the value of the statistic, and the number that follows p present the probability that the result of the test is statistically significant. All this information allows expert-readers to determine if the statistical analysis was conducted appropriately. While some specialized software will generate all pieces of information simultaneously (e.g., SPSS), Excel does not. Hence, you will answer the next question to obtain those t-test values.

10. Using the TINV function, you will calculate the exact value for the t-test. Start by inserting the TINV function into cell K27. Now, do the following.

• The probability field. Select the p-value in Cell L27 and type in “*2”. Hence, when you are done, the field should contain “L27*2”.
• The Deg_freedom field. Degrees of freedom for a within-subject design are equal to the sample size minus one.

Now, using analogous steps, calculate the t-values for the two other comparisons. This work must appear in Cells K28 and K29, respectively.

You now have all the descriptive and inferential statistics that you need to write your APA style paper!