PH 142 Final Exam

March 13, 2019

Please read the following statement and sign below to indicate that you understand the policies for this exam

You can use the back of each page as scratch paper, however no points will be given for answers on the back pages

Cellphones and computers must be stored and an silent during this exam.

You may use a non-graphing calculator and one double sided page of handwritten notes or notes printed at 10 point font or larger.

You must show your student ID when you submit your test.

For several of the questions we have added an answer box for your final a:nswer. This helps us to grade quickly. You may receive partial credit if you show your work, but to receive full credit your final answer must be recorded in the answer box for example if your answer were 3 it should look like this:

Your final answer here [1 point)

3

For questions with calculations, please express your answer to 2 decimal places.

Please sign below to indicate your understanding of the academic integrity statement. I u nderstand the University of California policy on academic integrity. I will not collaborate on this exam with my classmates or any other individuals. I will not use any resources to complete this exam other than those noted above. I understand that if I violate any of these policies, I will receive a O on this exam.'

Signature:

v2

1

Corinne Riddell

Skip this question — we tested this material on midterm II�

Question 5 [5.5 points]

DIRECTIONS For each of the following samples and outcomes - indicate the type of variable, an appropriate hypothesis test and the null hypothesis of that test expressed as an equation:

For example given the text below: A study testing whether the true mean weekly amount of dollars spent on coffee by university students is $20. The population standard deviation is known, the sample size is 20, and the data are normal.

The expected answers would be: The outcome variable is continuous. An appropriate hypothesis test would be a one sample Z test. For that test, Ho : p = 20

Q5.1 [1.5 points]

To understand the workings of the placebo effect on patients with Parkinson's disease, scientists measure level of activity at a key location in the brain when patients receive a placebo that they think is an active drug and also when no treatment is given. They measure the brain response activity under two conditions for each patient. The 4c7 patients are randomized to either active treatment and then placebo, or placebo and then active treatment. The data are s]gh!;ly skewed.

The outcome is a -�Q�0�V�1ti;�v�Jv�·o�uS _ _ _______ variable.

An appropriate hypothesis test would be:_+cp�()....�lj:_.,���=2:_S�ri�A�VIA'--'fdc,.,.>.oc,f�, �+--�-t_=t'=S-�f_ I

For that test Ho: difference in mu for each patient under different conditions = 0

Q5.2 (1.5 points]

You have heard that the grading scale is harsher at UC Berkeley than at other California universities. You want to test this rumor with data. You have data from a random sample of 100 transcripts from students at UCB who took PH142 and data on the letter grade distribution for undergraduate statistics courses in general from a California wide survey.

The outcome is a _ _.�"'""e<Xt�l?�Cj+O,�Y\�·-e.o...�-=9+-CVi�d�IV1�fA..,_,_,_\ _ __ variable.

An appropriate hypothesis test would be:

vc&·"\<�k�j For that test H0: ' '\ Ltl'Wl Q-P 1,

Q5.3 [2.5 points]

Ch1-Squ{},vPJ goo(ffi(�S Q,/)tii".

u�I

Your group is interested in mosquito control and resistance to insecticides used in the spraying of houses in areas where malaria is endemic. You have set up a study with test structures in an area where there are a lot of mosquitoes. You have treated e&ch structure with a different amount of insecticide and set up a mosquito trap in each structure. You then measure the n1Jmber of mosquitoes trapped In eacfistructure. _

The outcome is a --�a�,��ure��+�Q..�<-!�'U�VVi'-"U'."'-'--.... V�\-�CO-.<.a-D-- variable.

An appropriate hypothesis test would be a : h�O-.Y lf-J'\1(f'.)";10,,--,

the prPd1H<y� 1\-of'rv'[}q1(t11ttSfrom a cpV� aYY1cX.t'rl: oP lv-i\f c,t·1(�1d4__,,

For that test HO : � \ C)\';e_, (� ve01'.'�J(1 On I \V),t;-0

6

�

test on ��

4t l

Distribution in UCB = Distribution in CA

Corinne Riddell

Skip this question — we tested this material on midterm II�

Question 8 [8 points total)

Abstract (Adapted from Oldham-Cooper et al. "Playing a computer game during lm1ch affects fullness, memory for lunch, and later snack intake," Am J Clin Nutr. 2011 February; 93(2): 308-313)

Background The presence of distracting stimuli during eating increases the meal size and could thereby contribute to overeating and obesity. However, the effects of within-meal distraction on later food intake are less clear.

Objectives We sought to test the hypothesis that distraction inhibits memory encoding for a meal, which, in turn, increases later food intake.

Design The current study assessed the effects of playing solitaire (a computerized card-sorting game) during a fixed lunch, which was eaten at a fixed rate, on memory for lunch and food intake in a taste test 30 min later. A between-subjects design was used with 44 participants (22 per group). Participants in the no-distraction group ate the same lunch in the absence of any distracting stimuli. �

Results. Distracted individuals were less full after lunch, and they ate significantly more biscuits in the taste test than did nondistracted participants (mean intake: 52.1 SE:44�6 compared with �7.,_!_��J6 ). Furthermore, serial-order memory for the presentation of the 9 lunch items was less accurate in participants who had been distracted during lunch.

Conclusions. These findings provide further evidence that distraction during one meal has the capacity to influence subsequent eating. They may also help to explain the well-documented association between sedentary screen-time activities and overweight.

Q8.1 [1 point]

Early in the course we discussed three main problem types. The study described by this abstract is attempting to answer a problem that fits into which of the three categories?

Q8.2 [3 points]

The DAG (directed acyclic graph, or causal graph) below can be used to depict the main finding of this study (the one with the p-value listed in the abstract). You need to do three things. First, please label which variable (in this study) is the exposure and second, which is the outcome. Third, please add a confounder to the graph. In order to do this you must think of a potential confounding variable (there are many possible) and add it to the graph with the appropriate arrow(s).

Exposure: P\Gz<j\V-'0 sd rlv.._\v-C.::.,, wfdil�--+ Outcome:Qa.-h '/\ j IL-v1c1h

8

Q8.3 .[2 points]

Considering distributions of 'Grams of biscuit eaten'.

a. You want to create a visualization to assess the distribution of the variable 'Grams of biscuit eaten' inthe two exposure groups. What "geom" do you use in ggplot to make the appropriate graph for thisvariable? (1 point]

geom_ hwfu<?)\11\.IN\ ( b. ,i\1hat additional function could you add to the the ggplot call in order to plot the two different groups

on side by side plo�J [1 point] ··

Q8.4 [2 points]

Calculate the appropriate t statistic.- -·-···· --�l�?.. .. . -·

+=fc5d-. I - 7-7. l ( '- �--- -::1.,,1,,,l''

r;:�• t �! (o'u

� �2,, 1,:1,

Question 9 [3 points]

Cuvv.-Yti \ '. )I"' J-1. I ) s--=- '){o}

Y)� Tv

&yn:N· rJ; i :::; 1'"2.-, \ / s.::; Lff h I vi .c J:b

Your final answer here [1 point]

t= ),L---71

You are conducting a study on the number of emergency room (Efil phone calls received at Highland Hospital in Oakland on the night shift (9pm-5am), and you are told that the hospital averages 4 calls per hour during that time period.

· -

Q9.1 [2 points]

What is the probability of observing exactly 9 calls in a given hour during the night shift at Highland?

Your final answer here

Q9.2 [1 point]

Fill in the blanks to create the R code that you would use to calculate this probability:

o\,pVis ( q-=- q

, \ c,_ vvd2J,?.; t..) )

9

facet_wrap( ~ group)

Corinne Riddell

Skip this qu

Corinne Riddell

Skip this qu