ENGLISH VERSION

EXAM QUESTIONS FOR PSYF10STAT- INTRODUCTORY STATISTICS FOR PSYCHOLOGISTS

Full Name:
Årskortnummer:
Date:

This exam is made up of three parts: Part 1 includes 30 multiple choice questions; Part 2 includes 10 matching questions; and Part 3 includes 3 written/calculation questions. You must complete all three parts. You have three hours to complete the exam.

In this package, you have been given both an English and a Danish version of the exam questions. Use the version that you feel most comfortable with. You have also been provided with a formula sheet and critical value tables for t and χ2.

Please do not record your answers on the question sheets in this package- you have been given a separate package of answer sheets where you should record all of your answers and show all of your calculations for the exam questions. Make sure to write your name on each of the answer sheets in case the pages get separated!

When you have completed the exam, you must write your name on all of the exam materialsand hand them in. You cannot take any of the question sheets, answer sheets, or formula sheets away with you.

GOOD LUCK!

PART 1. MULTIPLE CHOICE QUESTIONS (30 POINTS)

Choose the best answer for each of the questions below. Please record your answers on the separate answer sheets provided.

  1. When designing an experimental research project, we typically aim to:
  1. Maximize systematic variation and limit unsystematic variation
  2. Maximize unsystematic variation and limit systematic variation
  3. Balance systematic and unsystematic variation
  4. Eliminate systematic and unsystematic variation
  1. Which of the following is an equal probability sampling method?
  1. Snowball sampling
  2. Quota sampling
  3. Stratified sampling
  4. Convenience sampling
  1. Which of the following terms best describes the sentence: ‘Sleep deprivation will reduce the ability to perform a complex cognitive task’?
  1. A research question
  2. A directional hypothesis
  3. A null hypothesis
  4. A non-directional hypothesis
  1. If a researcher collects information on athletes’ rankings in a race (e.g., 1st, 2nd, 3rd…), what type of data has he collected?
  1. Binary
  2. Nominal
  3. Ordinal
  4. Interval
  1. If p=.01, then the risk of making a Type 1 error is:
  1. 99%
  2. 1%
  3. .01%
  4. 5%
  1. Critical values for significance in a theoretical distribution depend on:
  1. The type of statistical test conducted
  2. The p level chosen for the test
  3. The degrees of freedom for the test
  4. All of the above
  1. If the critical t-value for significance at p<.05 is 2.09 and a researcher calculates that the t-value for his test is also 2.09, what should he do?
  1. Reject the null hypothesis
  2. Fail to reject the null hypothesis
  3. Conclude that the alternative hypothesis has been proven correct
  4. Both a and c
  1. A statistically significant test indicates:
  1. That there is a high probability of the result occurring if the null hypothesis is true
  2. That the result is important
  3. That the null hypothesis has been proven false
  4. That there is a low probability of the result occurring if the null hypothesis is true
  1. If we calculated an effect size and found that it was r = .42, which description would best describe the size of effect?
  1. Small
  2. Medium
  3. Large
  4. Significant
  1. Consider the SPSS output below. What does it indicate?
  1. That the assumption of sphericity has been violated
  2. That the homogeneity of variance assumption has been violated
  3. That the assumption of sphericity has been satisfied
  4. That the homogeneity of variance assumption has been satisfied
  1. If your sample data is normally distributed but your sample size is small (<30), you should:
  1. Automatically use a parametric test
  2. Automatically use a nonparametric test
  3. Invoke the central limit theorem and then use a parametric test
  4. Collect more data
  1. The frequency distribution below is:
  1. Positively skewed
  2. Leptokurtic
  3. Platykurtic
  4. Negatively skewed
  1. What is the median in the following distribution 8, 12, 15, 19, 19, 20, 22, 25, 32, 34?
  1. 19

b. 20

c. 19.5

d. 5.5

  1. In a standard z-distribution, what percentage of scores will fall between +2.58 and -2.58?
  1. 95%
  2. 99%

c. 1%

d. 99.9%

  1. If the scores on an exam have a mean of 26 and a standard deviation of 4, what is the z-score for an exam score of 18?
  1. –2
  2. 11
  3. 2
  4. –1.41
  1. The output below could be interpreted to mean that:

  1. As delay increased, exam score increased
  2. As delay increased, exam score decreased
  3. As delay increased, exam score stayed the same
  4. There was no relationship between delay and exam score
  1. What is the percentage of shared variance for a correlation of r= .90?
  1. 81%
  2. 18%
  3. 9%
  4. None of the above
  1. Which of the following statements about a Pearson’s correlation is not true?
  1. The values of r ranges from +1 to -1
  2. The null hypothesis is that r =0
  3. The significance of r is determined using the t distribution
  4. A significant r indicates that one variable causes the other
  1. Using the regression equation in your formula sheet, determine the predicted value of y for a regression line with an intercept of 13.26, a slope of 0.43, and an x value of 24.

a. 328.56

b. 23.58

c. 37.69

d. 318.67

  1. What regression assumption, if any, has been violated in the scatterplot below?
  1. The assumption of sphericity
  2. The assumption of linearity
  3. The assumption of normality
  4. There is no violation
  1. A researcher is interested in the effects of group therapy on social anxiety disorders. She takes the same group of 8 patients and measures their social anxiety (out of 15) before beginning group therapy and again 2 weeks after the therapy is completed. The data are not normally-distributed. Which test should she use to compare her two conditions?
  1. Dependent t-test
  2. Mann-Whitney test
  3. Wilcoxon signed ranks test
  4. Chi-Squared test
  1. A psychologist collects data on 20 students’ scores on a quiz before and after they attended a review session. Based on the SPSS output below, what is the value of Wilcoxon’s T?

a. 32.5

b. 138.5

c. 2.31

d. 8.13

  1. If a dependent t-test has 19 degrees of freedom, how many participants are in the study?
  1. 20
  2. 21
  3. 39
  4. 40
  1. A researcher is interested in the effects of positive vs. negative reinforcement on exam performance. Before an SPSS exam, a lecturer gave half of her students positive reinforcement and the other half negative reinforcement. The students then wrote the exam and their grades were compared (higher scores= better performance). What should she report based on the SPSS output below?
  1. Students receiving positive reinforcement did significantlybetter than those receiving negative reinforcement, t(38) = 2.05, p < .05
  2. The grades for students receiving positive as compared to negative reinforcement did not significantlydiffer, t(38) = 2.05,p >.05
  3. Students receiving positive reinforcement did significantly better than those receiving negative reinforcement, t(23.12) = 2.05, p < .05
  4. The grades for students receiving positive as compared to negative reinforcement did not significantlydiffer, t(23.12) = 2.05, p>.05
  1. A researcher conducts an independent t-test with a sample size of 20 people (10 people in each group) and gets a two-tailed t-value of 1.98. Using the critical value table for t, determine the significance level of his test.
  1. p<.05
  2. p<.01
  3. p>.05
  4. p<.001
  1. If an ANOVA F-ratio is greater than 1, what can we conclude?
  1. Effect > Error
  2. Error > Effect
  3. F is significant
  4. F is not significant
  1. A researcher conducts a One-way Independent ANOVA with 4 groups and wants to carry out post-hoc Bonferroni tests which control for a family-wise error rate of 5%. If he compares all possible combinations of the 4 groups, what p value should he use for the post-hoc tests?
  1. .05
  2. .0166
  3. .0083
  4. .0125
  1. A developmental psychologist compares the rank scores of 15 children and 15 adults on a moral reasoning task. Using the SPSS output below, calculate an effect size measure for this test.

  1. r= -.85
  2. r= -1.21
  3. r= .000
  4. r= .93
  1. A researcher has conducted a Friedman’s ANOVA to explore the effect of three treatment conditions on an ordinal measure of outcome. What type of post-hoc test should he conduct to identify which conditions differ from one another?
  1. Independent t-tests
  2. Dependent t-tests
  3. Mann-Whitney tests
  4. Wilcoxon Signed-rank tests
  1. Which of the following is not an assumption of the chi-squared test?
  1. 2 categorical variables
  2. Mutually exclusive categories
  3. All expected frequencies >1
  4. Normally distributed data

1

PART 2. MATCHING QUESTIONS (10 POINTS)

Match each sentence on the left to one of the possible answers on the right. Each answer can only be used once. Please record your answers on the separate answer sheets provided.

1

Definition

  1. The variable that is manipulated or controlled by the researcher in an experimental study
  1. The experimental design best able to control for unsystematic variance
  1. Statistical procedures used to draw generalizations about a population based on findings in a sample
  1. The type of significance test used when a researcher has stated a directional hypothesis
  1. The standard deviation of the sampling distribution of means
  1. The correlation coefficient used to assess the relationship between an interval variable and a binary variable
  1. A correction that can be used to deal with heterogeneity of variance
  1. The value of y when x = 0 in a regression equation
  1. The degree to which a statistical model is an accurate representation of the observed data
  1. An objective and standardized measure of an effect that can be directly compared across studies

Possible Answers

  1. Two-tailed test
  2. Confounding variable
  3. Effect size
  4. Kendall’s Tau
  5. Statistical significance
  6. Independent variable
  7. Greenhouse Geisser
  8. Correlational design
  9. Standard error
  10. Descriptive statistics
  11. Point Biserial rpb
  12. Dependent variable
  13. Model fit
  14. Slope
  15. Repeated measures design
  16. Variance
  17. One-tailed test
  18. Spearman’s rho
  19. Inferential statistics
  20. Independent design
  21. Intercept
  22. Brown-Forsythe

1

PART 3. WRITTEN/CALCULATION QUESTIONS (30 POINTS)

Answer each of the questions below in as much detail as possible. Pleaserecord your answers and show all of your calculations on the separate answer sheets provided. You may choose to answer in either Danish or English.

  1. In order to test whether violent criminals were more likely than non-criminals to have been abused as children, a forensic psychologist conducted in-depth personal history interviews with 25 individuals serving a prison sentence for a violent crime and 25 individuals with no criminal record. For each participant he recorded whether the person had a history of physical or sexual abuse during childhood (yes or no). His data are shown below.

History of Abuse / No History of Abuse
Violent Criminal / 16 / 9
Non-Criminal / 6 / 19
  1. What test should he use to analyze the association between criminality and abuse? (1 point)
  1. Using your formula sheet, calculate the correct statistic. Remember to show all of the steps in your calculation. (5 points)
  1. Use one of the critical value tables, attached, to determine if theassociation between abuse and criminality is significant. (1 point)
  1. Calculate an appropriate effect size measure for this test. (2 points)
  1. Based on all of the above information, what should the psychologist conclude? (1 point)

2.A cognitive psychologist was interested in exploring whether different types of music had an impact on individuals’ concentration levels when performing an attention task. The psychologist recruited 20 participants for a study in which they were asked to perform the same attention task under four conditions (in silence, when listening to classical music, when listening to pop music, and when listening to rock music). The participants were given a performance score on the attention task ranging from 0-100 for each of the four test conditions. When exploring his data, the researcher found that neither the K-S test nor the Shapiro Wilks tests were significant. Based on this description and the SPSS output provided below:

  1. Report the psychologist’s findings in APA format. Remember to report all of the relevant information, including means and standard deviations. (10 points)

Music_condition / Dependent Variable
1 / Silence
2 / Classical_music
3 / Pop_music
4 / Rock_music
Descriptive Statistics
Mean / Std. Deviation / N
Silence Condition / 71,15 / 15,097 / 20
Classical Music Condition / 51,25 / 14,903 / 20
Pop Music Condition / 30,65 / 13,355 / 20
Rock Music Condition / 23,55 / 9,179 / 20
Mauchly's Test of Sphericity
Measure:MEASURE_1
Within Subjects Effect / Mauchly's W / Approx. Chi-Square / df / Sig. / Epsilon
Greenhouse-Geisser / Huynh-Feldt / Lower-bound
Music_condition / ,725 / 5,692 / 5 / ,338 / ,865 / 1,000 / ,333
Tests of Within-Subjects Effects
Measure:MEASURE_1
Source / Type III Sum of Squares / df / Mean Square / F / Sig.
Music_condition / Sphericity Assumed / 27720,400 / 3 / 9240,133 / 58,462 / ,000
Greenhouse-Geisser / 27720,400 / 2,595 / 10681,051 / 58,462 / ,000
Huynh-Feldt / 27720,400 / 3,000 / 9240,133 / 58,462 / ,000
Lower-bound / 27720,400 / 1,000 / 27720,400 / 58,462 / ,000
Error(Music_condition) / Sphericity Assumed / 9009,100 / 57 / 158,054
Greenhouse-Geisser / 9009,100 / 49,310 / 182,702
Huynh-Feldt / 9009,100 / 57,000 / 158,054
Lower-bound / 9009,100 / 19,000 / 474,163
Pairwise Comparisons
Measure:MEASURE_1
(I) Music_condition / (J) Music_condition / Mean Difference (I-J) / Std. Error / Sig.a / 95% Confidence Interval for Differencea
Lower Bound / Upper Bound
1 / d / 2 / 19,900* / 4,578 / ,002 / 6,423 / 33,377
3 / 40,500* / 4,371 / ,000 / 27,633 / 53,367
4 / 47,600* / 3,713 / ,000 / 36,669 / 58,531
2 / di / 1 / -19,900* / 4,578 / ,002 / -33,377 / -6,423
3 / 20,600* / 4,239 / ,001 / 8,122 / 33,078
4 / 27,700* / 3,892 / ,000 / 16,242 / 39,158
3 / di / 1 / -40,500* / 4,371 / ,000 / -53,367 / -27,633
2 / -20,600* / 4,239 / ,001 / -33,078 / -8,122
4 / 7,100 / 2,806 / ,122 / -1,160 / 15,360
4 / 1 / -47,600* / 3,713 / ,000 / -58,531 / -36,669
2 / -27,700* / 3,892 / ,000 / -39,158 / -16,242
3 / -7,100 / 2,806 / ,122 / -15,360 / 1,160
Based on estimated marginal means
*. The mean difference is significant at the ,05 level.
a. Adjustment for multiple comparisons: Bonferroni.
  1. A social psychologist is interested in studying whether there is a difference in the level of pro-social behaviour that individuals will display when they are in the presence of other people as compared to when they are alone. She designs a study in which participants are asked to imagine that they have just won kr. 20,000 in lottery money. The participants are then asked to state how much of this money they would be willing to donate to charity (on a scale from 0-20,000 kroner). The psychologist recruits 40 students for her study and randomly assigns 20 of them to complete this task when they are alone. The other 20 students complete the task in small groups and are asked to report their decision in front of the others. The psychologist hypothesizes that in order to portray a positive self-image the students will donate more money to charity when they are in small groups as compared to when they are alone. Based on this description and the output below, identify the following:

Tests of Normality
Kolmogorov-Smirnova / Shapiro-Wilk
Statistic / df / Sig. / Statistic / df / Sig.
Donation_Amount / ,220 / 40 / ,012 / ,747 / 20 / ,000
  1. The predictor and outcome variables for this study and their levels of measurement. (2 points)
  1. A null hypothesis, non-directional alternative hypothesis, and directional alternative hypothesis. (3 points)
  1. A step-by-step decision process leading to a statement of what type of statistical test to run in this situation (you do not need to calculate the test). Be as detailed as possible! (5 points)

DANSK VERSION
EKSAMENSPØRGSMÅL FOR PSYF10STAT- INTRODUCTORY STATISTICS FOR PSYCHOLOGISTS

Navn:
Årskortnummer:
Dato:

Denne eksamen består af tre dele: Del 1 indeholder 30 multiple-choice spørgsmål; Del 2 indeholder 10 matchende spørgsmål; og Del 3 indeholder 3 skriftlige- /beregningsspørgsmål. Du skal udfylde alle tre dele. Du har tre timer til eksamen.
I dette opgavesæt, har du fået udleveret både en engelsk og en dansk version af eksamensspørgsmålene. Brug den version, du føler dig mest tryg ved.Du har også fået udleveret et ark med formler og tabeller for kritiske værdier for t og χ2.
Undlad venligst at registrere dine svar på dette opgavesæt. Du har fået udleveret et særskilte svararksæt, hvor du skal registrere alle dine svar og beregninger på eksamensspørgsmålene. Husk at skrive dit navn på alle svararkene i tilfælde af, at siderne bliver adskilt!
Når du er færdig, skal du skrive dit navn på alt materiale og indlevere alle arkene. Du må hverken tage spørgsmåls,- formel- eller svarark med dig!

HELD OG LYKKE!

DEL 1. MULTIPLE CHOICE SPØRGSMÅL (30 POINTS)

Vælg det bedste svar for hver af nedenstående spørgsmål. Registrer venligst dine svar på det særskiltesvarark.

1.Når vi designer et eksperimentelt forskningsprojekt, har vi typisk har til formål at:
a. Maksimere systematisk variation og begrænse usystematisk variation
b. Maksimere usystematisk variation og begrænse systematisk variation
c. Balancere systematisk og usystematisk variation
d. Eliminere systematisk og usystematisk variation

2.Hvilke af følgende metoder forstås som en ”equal probability” metode?
a. ’Snebold’ metoden
b. ’Quota’ metoden

c.’ Stratificeret’ metoden
d. ’Convenience’ metoden

3.Hvilke af følgende udtryk beskriver bedst sætningen: 'Søvnmangel reducerer evnen til at udføre en kompleks kognitiv opgave"?
a. Et forskningsspørgsmål
b. En retningsbestemt hypotese
c. En nulhypotese
d. En ikke-retningsbestemt hypotese

4.Hvis en forsker indsamler oplysninger om atleters placeringer i et løb (f.eks 1., 2., 3. ...), hvilken type data har han så indsamlet?
a. Binær
b. Nominel
c. Ordinal
d. Interval

5.Hvis p=.01, så er risikoen for at foretage en Type 1 fejl:
a. 99%
b. 1%
c. 0.01%
d. 5%

6.Kritiske værdier for signifikans i en teoretisk fordeling afhænger af:
a. Typen af den udførte statistiske test
b. Det valgte p niveau
c. Testens frihedsgrader
d. Alle de ovennævnte

7.Hvis den kritiske t-værdi for signifikans ved p<.05 er 2.09, og en forsker har beregnet, at t-værdien for hans test også er 2.09, hvad skal han så gøre?
a. Afvise nulhypotesen
b. Ikke afvise nulhypotesen
c. Konkludere, at den alternative hypotese er blevet bevist korrekt
d. Både A og C

8.En statistisk signifikant test viser:
a. At der er en høj sandsynlighed for, at testresultatet forekommer, hvis nulhypotesen er sand
b. At resultatet er vigtigt
c. At nulhypotesen er blevet bevist falsk
d. At der er en lav sandsynlighed for, at testresultatet forekommer, hvis nulhypotesen er sand

9.Hvis vi har beregnet en effektstørrelse og fundet, at det var r=.42, hvilken beskrivelse beskriver så bedst effektstørrelsen?
a. Lille
b. Medium
c. Stor
d. Signifikans

10.Overvej nedenstående SPSS output. Hvad viser det?

  1. At forudsætningen om ’sphericity’ er blevet brudt
  2. At forudsætningen om homogen varians er blevet brudt
  3. At forudsætningen om ’sphericity’ er tilfredsstillet
  4. At forudsætningen om homogen varians er tilfredsstillet

11.Hvis din stikprøve data er fordelt normalt, men din stikprøvestørrelse er lille (<30), bør du:
a. Automatisk bruge en parametrisk test
b. Automatisk bruge en ikke-parametrisk test
c. Inddrage den centrale grænseværdisætning og derefter bruge en parametrisk test
d. Indsamle mere data

12.Hyppighedsfordelingen nedenfor er:

a. Positivt skæv
b. Leptokurtic
c. Platykurtic
d. Negativt skæv

13.Hvad er medianen i følgende fordeling8, 12, 15, 19, 19, 20, 22, 25, 32, 34?

a. 19
b. 20
c. 19.5
d. 5.5

14.I en standard z-fordeling, hvor mange procent af værdier vil falde mellem 2.58 og -2.58?

  1. 95%
  2. 99%

c. 1%

d. 99.9%

15.Hvis gennemsnittet af scoren til en eksamen er 26 og standardafvigelsen er 4, hvad er z-scoren for en score på 18 så?

  1. –2
  2. 11
  3. 2
  4. –1.41

16.Outputtet nedenfor kan fortolkes som følger:

a. Ved øget forsinkelse steg eksamens scoren
b. Ved øget forsinkelse dalede eksamens scoren
c. Ved øget forsinkelse forblev eksamens scoren konstant
d. Der var ingen sammenhæng mellem forsinkelse og eksamens scoren

17.Hvad er procentdelen af fælles varians for en korrelation på r= .90?
a. 81%
b. 18%
c. 9%
d. Ingen af ovenstående

18. Hvilke af følgende udsagn om en Pearsons korrelation er ikke sande?
a. Værdierne af r varierer fra +1 til -1
b. Nulhypotesen er, at r=0
c. Signifikansen af r bestemmes ved hjælp af t distribution
d. Et signifikant r angiver, at en variabel forårsager den anden

19.Brug regressionsligningen i din formelark til at bestemme den forudsagte værdi af y for en regressionslinje med en skæring (intercept) på 13.26, en hældning (slope) på 0.43, og en x-værdi på 24.

a. 328.56

b. 23.58

c. 37.69

d. 318.67

20.Hvilken regressionsforudsætning, om nogen, er blevet overtrådt i scatterplottet nedenfor?

a. Forudsætningen om ’sphericity’
b. Forudsætningen om linearitet
c. Forudsætningen om normalitet
d. Ingen forudsætninger er blevet overtrådt

21.En forsker er interesseret i effekten af gruppeterapi for social angst. Hun tager den samme gruppe af 8 patienter og måler deres sociale angst (ud af en mulig maxscore på 15) før påbegyndelse af gruppeterapi og igen 2 uger efter behandlingen er afsluttet. Data er ikke normalt-fordelt. Hvilken test skal hun bruge til at sammenligne hendes to betingelser?
a. Dependent t-test
b. Mann-Whitney test
c. Wilcoxon signed ranks test
d. Chi i anden(χ2) test

22.En psykolog indsamler data om 20 studerendes på en quiz før og efter de deltog i en spørgstime om emnet. Baseret på nedenstående SPSS output, hvad er værdien af Wilcoxon's T?