Statistics

Please read the instructions below carefully. If anything is unclear, let me know asap (come to the
revision live session and raise it there and/or email me) – do not wait until the week before
submission!
Complete the coursework independently The purpose of this coursework is to assess your
knowledge, not the collective wisdom of your group of friends. The University uses software to
detect plagiarism. If you do not work independently, but collaborate, there is a very real risk that
your assignment will be flagged as plagiarism. The University has no way of knowing who plagiarised
whom, and so everyone in the group will suffer (even those who did the work). There are severe
penalties for plagiarism – please do not risk this!
Use APA model answers as guides – not templates (i.e., do not simply replace words and/or re-jig
sentences). If you do this, you risk your answer being flagged up for plagiarism (because it is too
similar to the model answer). Use your own words to describe the results!
Make sure you select the correct test A researcher must know what test to apply when. Recall the
selection method from our asynchronous lectures and synchronous live sessions: first we establish
the design (within or between?), then we test to see whether we can apply a parametric test (do
assumptions hold?). If the assumptions hold, we apply a parametric test. If the assumptions do not
hold, we apply a non-parametric test.
Read each question carefully and make sure you answer the set question
Submit on time The coursework is due on the 18
th of December by 3 PM. Do not leave submission
until the last minute! There are severe penalties for late submission – please do not risk this! IF you
have problems submitting before the deadline, email a copy to psych@city.ac.uk and cc’ me. That
way there is an electronic record of your submission being on time.
Formatting Report your answers on the dedicated answer sheet
PS1007_coursework_autumn_2020_answer_sheet.docx
We have a separate answer sheet to avoid the plagiarism software reporting too many false
positives. Not using the form risks your coursework being flagged as plagiarism.
APA formatting Recall that reporting of statistics should adhere to the APA format. You won’t fail if
it does not – but you will be marked down.
Finally, enjoy the challenge, have fun, and good luck!
Andreas
MCQ – 10 marks
An increase in sample size increases degrees of freedom for t-tests. What effect, if any, does this have on
our ability to detect a significant difference for a specific t-statistic (e.g., t = 3).
1) We become more likely to detect an effect (if there is one)
2) We become less likely to detect an effect (if there is one)
3) Nothing changes
Which of the following is most useful for checking our data prior to hypothesis testing?
1) Q-Q plot
2) Bar plot
3) Error-bar plot
4) 1 and 3
The standard error of the mean is the standard deviation of the sampling distribution of the mean
1) True
2) False
Too little sleep makes people grumpy. This is a:
1) Directional hypothesis
2) Non-directional hypothesis
Which of the following is a measure of dispersion?
1) Mode
2) Cohens D’
3) IQR
4) Median
The degrees of freedom for a repeated measures t-test is:
1) N-2
2) N-1
Which of the following makes it very easy to detect outliers?
1) Histogram
2) Boxplot
3) Pie chart
4) 1) and 3)
Which of the following is/are assumption/s of the non-parametric equivalents to t-tests?
1) Data is normally distributed
2) Data is at least ordinal
3) Data are independent
4) 2) and 3)
You have collected some ordinal data. By exploring it you have decided that it’s normally distributed. Which
of the following should you use to describe the data?
1) Mean and mode
2) Median and interquartile range
3) Mean and standard deviation
The effect size tells us about
1) The importance (magnitude) of an effect
2) The significance
3) Our ability to detect an effect (if there is one)
Caffeine and working memory – 25 marks
A group of researchers were interested in exploring the effect of caffeine on working memory
performance. One group of participants performed the task after taking a caffeine pill, and another
group of participants performed the task after taking placebo (a sugar pill). The greater the memory
score the better performance in this task. The data are available on Moodle as ‘caffeine.sav’.
A) Describe design (within or between?), the null and the alternative hypotheses.
B) Explore the data. Copy and paste q-q plots, boxplots and normality tests into the answer sheet
and interpret them (describe what they tell you about the data).
C) Carry out the appropriate inferential test, and in a separate section, report and interpret the
results, as if you were writing a journal article (using APA guidelines). Report descriptive statistics,
inferential statistics and effect size. Interpret the results of the test.
PPC and value representation – 25 marks
You recorded fMRI BOLD in posterior-parietal cortex (PPC) when 29 participants were doing a
decision-making task. There were two conditions, one in which participants experienced large
magnitudes and one in which the same participants experienced small magnitudes. You
hypothesized that larger magnitudes would lead to an increased activity in this region compared to
smaller magnitudes. You have already extracted the average activation in this region separately for
the two conditions. The data is available on Moodle as ‘PPC and evaluation.sav’.
A) Describe design (within or between?), the null and the alternative hypotheses.
B) Explore the data. Copy and paste q-q plots, boxplots and normality tests into the answer sheet
and interpret them (describe what they tell you about the data).
C) Carry out the appropriate inferential test, and in a separate section, report and interpret the
results, as if you were writing a journal article (using APA guidelines). Report descriptive statistics,
inferential statistics and effect size. Interpret the results of the test.
Improving sleep quality – 25 marks
A team of researchers designed an intervention to improve sleep quality. They had participants wear
smart watches that monitored participants sleep quality, and from which they were abler to extract
a sleep quality measure (the higher the score – the better the sleep quality). One group of
participants sleep quality was measured without any intervention. Another group of participants
sleep quality was measured following a relaxing regime just before bedtime. You can find the data as
‘sleepquality.sav’ on Moodle.
A) Describe design (within or between?), the null and the alternative hypotheses.
B) Explore the data. Copy and paste q-q plots, boxplots and normality tests into the answer sheet
and interpret them (describe what they tell you about the data).
C) Carry out the appropriate inferential test, and in a separate section, report and interpret the
results, as if you were writing a journal article (using APA guidelines). Report descriptive statistics,
inferential statistics and effect size. Interpret the results of the test.
Distributions and extreme data points – 5 marks
Imagine that all your parents are gathered in a room (everyone in PS1007). Now imagine that Jeff
Bezos walks in.
Which central tendency measure of the income distribution would change the most? Why?
Which measure of dispersion measure would change the most? Why?
Which descriptive statistics would you use to describe the distribution? Why?
The four inferential tests – 10 marks
Below are data from two conditions: a high stress performance rating and a low stress performance
rating. The researchers were interested in exam performance as a function of stress level and
manipulated stress in students. Note that the data has not been formatted for SPSS and that you
need to do that appropriately for each test below.
Carry out all four inferential tests we have learnt about this term. Report the result of each test
using APA conventions. You only need to report the test results and the descriptive statistics. Do not
NOT report exploratory analyses and do NOT report effect sizes.
X Y
9.4 9.2
13 14.3
8.7 15.1
11.4 16.5
10.8 11.7
11.8 15.1
14.2 19.2
12.3 13.3
16.1 15.7
17.1 24.2
11.3 18.6
6.9 6.9
2.9 5.8
13.8 15.5
5.8 2.9
Dependent t-test
Wilcoxon Signed rank
independent t-test
Mann Whitney U